{"question_id": 1, "text": " Design a feature for a social media website to recommend articles to users based on how similar the articles are to their previously liked articles.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Summarization\\', \\'api_name\\': \\'facebook/bart-large-cnn\\', \\'api_call\\': \"pipeline(\\'summarization\\', model=\\'facebook/bart-large-cnn\\')\", \\'api_arguments\\': [\\'ARTICLE\\', \\'max_length\\', \\'min_length\\', \\'do_sample\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'from transformers import pipeline\\\\nsummarizer = pipeline(summarization, model=facebook/bart-large-cnn)\\\\nARTICLE = ...\\\\nprint(summarizer(ARTICLE, max_length=130, min_length=30, do_sample=False))\\', \\'performance\\': {\\'dataset\\': \\'cnn_dailymail\\', \\'accuracy\\': {\\'ROUGE-1\\': 42.949, \\'ROUGE-2\\': 20.815, \\'ROUGE-L\\': 30.619, \\'ROUGE-LSUM\\': 40.038}}, \\'description\\': \\'BART (large-sized model), fine-tuned on CNN Daily Mail. BART is a transformer encoder-encoder (seq2seq) model with a bidirectional (BERT-like) encoder and an autoregressive (GPT-like) decoder. BART is pre-trained by (1) corrupting text with an arbitrary noising function, and (2) learning a model to reconstruct the original text. BART is particularly effective when fine-tuned for text generation (e.g. summarization, translation) but also works well for comprehension tasks (e.g. text classification, question answering). This particular checkpoint has been fine-tuned on CNN Daily Mail, a large collection of text-summary pairs.\\'}', metadata={})]", "category": "generic"}
{"question_id": 2, "text": " The user is interested in a tool to find relationships between medical terms.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Token Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Named Entity Recognition\\', \\'api_name\\': \\'d4data/biomedical-ner-all\\', \\'api_call\\': \"AutoModelForTokenClassification.from_pretrained(\\'d4data/biomedical-ner-all\\')\", \\'api_arguments\\': {\\'model\\': \\'AutoModelForTokenClassification.from_pretrained(d4data/biomedical-ner-all)\\', \\'tokenizer\\': \\'AutoTokenizer.from_pretrained(d4data/biomedical-ner-all)\\', \\'aggregation_strategy\\': \\'simple\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'from transformers import pipeline, AutoTokenizer, AutoModelForTokenClassification\\'}, \\'example_code\\': \\'pipe(The patient reported no recurrence of palpitations at follow-up 6 months after the ablation.)\\', \\'performance\\': {\\'dataset\\': \\'Maccrobat\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'An English Named Entity Recognition model, trained on Maccrobat to recognize the bio-medical entities (107 entities) from a given text corpus (case reports etc.). This model was built on top of distilbert-base-uncased.\\'}', metadata={})]", "category": "generic"}
{"question_id": 3, "text": " As a journalist, I am curious about speech sentiment analysis in a group of people in a crowd. I want to extract features from the audio to run sentiment analysis.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'michellejieli/NSFW_text_classifier\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'michellejieli/NSFW_text_classification\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'classifier(I see you\u2019ve set aside this special time to humiliate yourself in public.)\\', \\'performance\\': {\\'dataset\\': \\'Reddit posts\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'DistilBERT is a transformer model that performs sentiment analysis. I fine-tuned the model on Reddit posts with the purpose of classifying not safe for work (NSFW) content, specifically text that is considered inappropriate and unprofessional. The model predicts 2 classes, which are NSFW or safe for work (SFW). The model is a fine-tuned version of DistilBERT. It was fine-tuned on 14317 Reddit posts pulled from the Reddit API.\\'}', metadata={})]", "category": "generic"}
{"question_id": 4, "text": " A chat service needs a way to compare and cluster similar sentences from users in different languages. Find a suitable feature extraction method to achieve this.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Sentence Similarity\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Feature Extraction\\', \\'api_name\\': \\'sentence-transformers/distilbert-base-nli-mean-tokens\\', \\'api_call\\': \"SentenceTransformer(\\'sentence-transformers/distilbert-base-nli-mean-tokens\\')\", \\'api_arguments\\': [\\'sentences\\'], \\'python_environment_requirements\\': \\'pip install -U sentence-transformers\\', \\'example_code\\': \"from sentence_transformers import SentenceTransformer\\\\nsentences = [This is an example sentence, Each sentence is converted]\\\\nmodel = SentenceTransformer(\\'sentence-transformers/distilbert-base-nli-mean-tokens\\')\\\\nembeddings = model.encode(sentences)\\\\nprint(embeddings)\", \\'performance\\': {\\'dataset\\': \\'https://seb.sbert.net\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This is a sentence-transformers model: It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search.\\'}', metadata={})]", "category": "generic"}
{"question_id": 5, "text": " I am an interior designer and want to showcase a modern living room with a fireplace and a large window overlooking a forest. Create an image according to this description.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Unconditional Image Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Denoising Diffusion Probabilistic Models (DDPM)\\', \\'api_name\\': \\'google/ddpm-bedroom-256\\', \\'api_call\\': \"DDPMPipeline.from_pretrained(\\'google/ddpm-bedroom-256\\')\", \\'api_arguments\\': \\'None\\', \\'python_environment_requirements\\': \\'diffusers\\', \\'example_code\\': \\'!pip install diffusers\\\\nfrom diffusers import DDPMPipeline, DDIMPipeline, PNDMPipeline\\\\nmodel_id = google/ddpm-bedroom-256\\\\nddpm = DDPMPipeline.from_pretrained(model_id)\\\\nimage = ddpm().images[0]\\\\nimage.save(ddpm_generated_image.png)\\', \\'performance\\': {\\'dataset\\': \\'CIFAR10\\', \\'accuracy\\': {\\'Inception score\\': 9.46, \\'FID score\\': 3.17}}, \\'description\\': \\'We present high quality image synthesis results using diffusion probabilistic models, a class of latent variable models inspired by considerations from nonequilibrium thermodynamics. Our best results are obtained by training on a weighted variational bound designed according to a novel connection between diffusion probabilistic models and denoising score matching with Langevin dynamics, and our models naturally admit a progressive lossy decompression scheme that can be interpreted as a generalization of autoregressive decoding. On the unconditional CIFAR10 dataset, we obtain an Inception score of 9.46 and a state-of-the-art FID score of 3.17. On 256x256 LSUN, we obtain sample quality similar to ProgressiveGAN.\\'}', metadata={})]", "category": "generic"}
{"question_id": 6, "text": " We need a product description for an image-based online store platform that will help customers understand the specifics of the product.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'Linaqruf/anything-v3.0\\', \\'api_call\\': \"Text2ImagePipeline(model=\\'Linaqruf/anything-v3.0\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A text-to-image model that generates images from text descriptions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 7, "text": " Create a program to generate a description for an image provided as input.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'prompthero/openjourney-v4\\', \\'api_call\\': \"pipeline(\\'text-to-image\\', model=\\'prompthero/openjourney-v4\\')\", \\'api_arguments\\': {\\'text\\': \\'string\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"generate_image(\\'your text here\\')\", \\'performance\\': {\\'dataset\\': \\'Midjourney v4 images\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Openjourney v4 is trained on +124k Midjourney v4 images by PromptHero. It is used for generating images based on text inputs.\\'}', metadata={})]", "category": "generic"}
{"question_id": 8, "text": " I am a financial analyst, and I receive report after report filled with charts helping to explain trends and data in my field. However, I also need to have this information in tabular format. Please help me extract a linearized table from this chart.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'tabular regression\\', \\'api_name\\': \\'farouk97/autotrain-test7-2644pc-linearregr-38619101723\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'farouk97/autotrain-data-test7-2644pc-linearregr\\', \\'accuracy\\': {\\'Loss\\': 0.145, \\'R2\\': 0.0, \\'MSE\\': 0.021, \\'MAE\\': 0.099, \\'RMSLE\\': 0.101}}, \\'description\\': \\'A tabular regression model trained using AutoTrain to predict CO2 emissions (in grams).\\'}', metadata={})]", "category": "generic"}
{"question_id": 9, "text": " We are building an automatic video generation platform based on user-provided text. We need a reliable model to convert text instructions into appropriate videos.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'camenduru/text2-video-zero\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'camenduru/text2-video-zero\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'This model is used for generating videos from text inputs. It is based on the Hugging Face framework and can be used with the transformers library. The model is trained on a variety of text and video datasets, and can be used for tasks such as video summarization, video generation from text prompts, and more.\\'}', metadata={})]", "category": "generic"}
{"question_id": 10, "text": " How can I extract video content from a text file? Provide a code sample to generate the video based on the text.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'chavinlo/TempoFunk\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'chavinlo/TempoFunk\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Text-to-Video model using Hugging Face Transformers library. Model is capable of generating video content based on the input text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 11, "text": " We are developing a mobile app to demonstrate the AI's ability to generate a short video from text. The app focuses on processing written stories into video.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'ImRma/Brucelee\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'ImRma/Brucelee\\')\", \\'api_arguments\\': [\\'your_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Hugging Face model for converting Persian and English text into video.\\'}', metadata={})]", "category": "generic"}
{"question_id": 12, "text": " Hey, I want to analyze images in my phone gallery and answer questions about them.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 13, "text": " My company wants to develop an application that will analyze images in relation to food and answer questions about them. We want it to handle questions like \\\"what is in the dish\\\" and \\\"how many calories does it have\\\".\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 14, "text": " We have received an invoice document, and would like to extract the total amount from it.\\n###Input: {'question': 'What is the total amount?', 'context': 'Invoice information for order ABC_123\\\\nProduct: Widget A, Quantity: 10, Price: $5 each\\\\nProduct: Widget B, Quantity: 5, Price: $3 each\\\\nProduct: Widget C, Quantity: 15, Price: $2 each\\\\nSubtotal: $75, Tax: $6.38, Total Amount Due: $81.38'}\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'layoutlm-invoices\\', \\'api_call\\': \"AutoModelForDocumentQuestionAnswering.from_pretrained(\\'impira/layoutlm-invoices\\')\", \\'api_arguments\\': \\'question, context\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"nlp(question=\\'What is the total amount?\\', context=\\'your_invoice_text\\')\", \\'performance\\': {\\'dataset\\': \\'proprietary dataset of invoices, SQuAD2.0, and DocVQA\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'A fine-tuned version of the multi-modal LayoutLM model for the task of question answering on invoices and other documents. It has been fine-tuned on a proprietary dataset of invoices as well as both SQuAD2.0 and DocVQA for general comprehension. Unlike other QA models, which can only extract consecutive tokens, this model can predict longer-range, non-consecutive sequences with an additional classifier head.\\'}', metadata={})]", "category": "generic"}
{"question_id": 15, "text": " As a clerk in a school, you want to extract information from some student enrollment forms. These forms contain students' details such as Name, age, and address.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Text Classification\\', \\'api_name\\': \\'joeddav/distilbert-base-uncased-go-emotions-student\\', \\'api_call\\': \"pipeline(\\'text-classification\\', model=\\'joeddav/distilbert-base-uncased-go-emotions-student\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': [\\'transformers\\', \\'torch\\', \\'tensorflow\\'], \\'example_code\\': \"from transformers import pipeline\\\\nnlp = pipeline(\\'text-classification\\', model=\\'joeddav/distilbert-base-uncased-go-emotions-student\\')\\\\nresult = nlp(\\'I am so happy today!\\')\", \\'performance\\': {\\'dataset\\': \\'go_emotions\\'}, \\'description\\': \\'This model is distilled from the zero-shot classification pipeline on the unlabeled GoEmotions dataset. It is primarily intended as a demo of how an expensive NLI-based zero-shot model can be distilled to a more efficient student, allowing a classifier to be trained with only unlabeled data.\\'}', metadata={})]", "category": "generic"}
{"question_id": 16, "text": " Find a model that can be used to predict the properties of molecules based on their graph representations.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'kochetkovIT/autotrain-ironhack-49741119788\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'json\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'kochetkovIT/autotrain-data-ironhack\\', \\'accuracy\\': {\\'Loss\\': 2.603, \\'R2\\': 0.013, \\'MSE\\': 6.776, \\'MAE\\': 1.666, \\'RMSLE\\': 0.502}}, \\'description\\': \\'A tabular regression model trained using AutoTrain to predict carbon emissions (in grams) with an R2 score of 0.013.\\'}', metadata={})]", "category": "generic"}
{"question_id": 17, "text": " Estimate the depth of a pool using computational depth estimation, given an underwater photo.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Depth Estimation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Depth Estimation\\', \\'api_name\\': \\'glpn-kitti\\', \\'api_call\\': \"GLPNForDepthEstimation.from_pretrained(\\'vinvino02/glpn-kitti\\')\", \\'api_arguments\\': \\'images, return_tensors\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'from transformers import GLPNFeatureExtractor, GLPNForDepthEstimation\\\\nimport torch\\\\nimport numpy as np\\\\nfrom PIL import Image\\\\nimport requests\\\\nurl = http://images.cocodataset.org/val2017/000000039769.jpg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nfeature_extractor = GLPNFeatureExtractor.from_pretrained(vinvino02/glpn-kitti)\\\\nmodel = GLPNForDepthEstimation.from_pretrained(vinvino02/glpn-kitti)\\\\ninputs = feature_extractor(images=image, return_tensors=pt)\\\\nwith torch.no_grad():\\\\n outputs = model(**inputs)\\\\n predicted_depth = outputs.predicted_depth\\\\nprediction = torch.nn.functional.interpolate(\\\\n predicted_depth.unsqueeze(1),\\\\n size=image.size[::-1],\\\\n mode=bicubic,\\\\n align_corners=False,\\\\n)\\\\noutput = prediction.squeeze().cpu().numpy()\\\\nformatted = (output * 255 / np.max(output)).astype(uint8)\\\\ndepth = Image.fromarray(formatted)\\', \\'performance\\': {\\'dataset\\': \\'KITTI\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Global-Local Path Networks (GLPN) model trained on KITTI for monocular depth estimation. It was introduced in the paper Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth by Kim et al. and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 18, "text": " I need technology that can analyze images and estimate their depth in a single camera.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Depth Estimation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Depth Estimation\\', \\'api_name\\': \\'glpn-kitti\\', \\'api_call\\': \"GLPNForDepthEstimation.from_pretrained(\\'vinvino02/glpn-kitti\\')\", \\'api_arguments\\': \\'images, return_tensors\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'from transformers import GLPNFeatureExtractor, GLPNForDepthEstimation\\\\nimport torch\\\\nimport numpy as np\\\\nfrom PIL import Image\\\\nimport requests\\\\nurl = http://images.cocodataset.org/val2017/000000039769.jpg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nfeature_extractor = GLPNFeatureExtractor.from_pretrained(vinvino02/glpn-kitti)\\\\nmodel = GLPNForDepthEstimation.from_pretrained(vinvino02/glpn-kitti)\\\\ninputs = feature_extractor(images=image, return_tensors=pt)\\\\nwith torch.no_grad():\\\\n outputs = model(**inputs)\\\\n predicted_depth = outputs.predicted_depth\\\\nprediction = torch.nn.functional.interpolate(\\\\n predicted_depth.unsqueeze(1),\\\\n size=image.size[::-1],\\\\n mode=bicubic,\\\\n align_corners=False,\\\\n)\\\\noutput = prediction.squeeze().cpu().numpy()\\\\nformatted = (output * 255 / np.max(output)).astype(uint8)\\\\ndepth = Image.fromarray(formatted)\\', \\'performance\\': {\\'dataset\\': \\'KITTI\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Global-Local Path Networks (GLPN) model trained on KITTI for monocular depth estimation. It was introduced in the paper Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth by Kim et al. and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 19, "text": " The client is a real estate company working on virtual tours. We need to help them estimate depth in images of houses.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Depth Estimation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Depth Estimation\\', \\'api_name\\': \\'glpn-kitti\\', \\'api_call\\': \"GLPNForDepthEstimation.from_pretrained(\\'vinvino02/glpn-kitti\\')\", \\'api_arguments\\': \\'images, return_tensors\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'from transformers import GLPNFeatureExtractor, GLPNForDepthEstimation\\\\nimport torch\\\\nimport numpy as np\\\\nfrom PIL import Image\\\\nimport requests\\\\nurl = http://images.cocodataset.org/val2017/000000039769.jpg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nfeature_extractor = GLPNFeatureExtractor.from_pretrained(vinvino02/glpn-kitti)\\\\nmodel = GLPNForDepthEstimation.from_pretrained(vinvino02/glpn-kitti)\\\\ninputs = feature_extractor(images=image, return_tensors=pt)\\\\nwith torch.no_grad():\\\\n outputs = model(**inputs)\\\\n predicted_depth = outputs.predicted_depth\\\\nprediction = torch.nn.functional.interpolate(\\\\n predicted_depth.unsqueeze(1),\\\\n size=image.size[::-1],\\\\n mode=bicubic,\\\\n align_corners=False,\\\\n)\\\\noutput = prediction.squeeze().cpu().numpy()\\\\nformatted = (output * 255 / np.max(output)).astype(uint8)\\\\ndepth = Image.fromarray(formatted)\\', \\'performance\\': {\\'dataset\\': \\'KITTI\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Global-Local Path Networks (GLPN) model trained on KITTI for monocular depth estimation. It was introduced in the paper Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth by Kim et al. and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 20, "text": " Assist me in setting up an image classifier that can recognize objects within an image.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Zero-Shot Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Zero-Shot Image Classification\\', \\'api_name\\': \\'patrickjohncyh/fashion-clip\\', \\'api_call\\': \"CLIPModel.from_pretrained(\\'patrickjohncyh/fashion-clip\\')\", \\'api_arguments\\': {\\'image\\': \\'File\\', \\'class_names\\': \\'String (comma-separated)\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import CLIPProcessor, CLIPModel; model = CLIPModel.from_pretrained(\\'patrickjohncyh/fashion-clip\\'); processor = CLIPProcessor.from_pretrained(\\'patrickjohncyh/fashion-clip\\'); inputs = processor(text=\\'blue shoes\\', images=image, return_tensors=\\'pt\\', padding=True); logits_per_image = model(**inputs).logits_per_image; probs = logits_per_image.softmax(dim=-1).tolist()[0]\", \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'FMNIST\\', \\'accuracy\\': 0.83}, {\\'name\\': \\'KAGL\\', \\'accuracy\\': 0.73}, {\\'name\\': \\'DEEP\\', \\'accuracy\\': 0.62}]}, \\'description\\': \\'FashionCLIP is a CLIP-based model developed to produce general product representations for fashion concepts. Leveraging the pre-trained checkpoint (ViT-B/32) released by OpenAI, it is trained on a large, high-quality novel fashion dataset to study whether domain specific fine-tuning of CLIP-like models is sufficient to produce product representations that are zero-shot transferable to entirely new datasets and tasks.\\'}', metadata={})]", "category": "generic"}
{"question_id": 21, "text": " Identify an object within an image based on textual description. For example, find a dog in the image.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'zero-shot-object-detection\\', \\'api_name\\': \\'google/owlvit-base-patch16\\', \\'api_call\\': \"OwlViTForObjectDetection.from_pretrained(\\'google/owlvit-base-patch16\\')\", \\'api_arguments\\': [\\'texts\\', \\'images\\'], \\'python_environment_requirements\\': [\\'requests\\', \\'PIL\\', \\'torch\\', \\'transformers\\'], \\'example_code\\': \\'processor = OwlViTProcessor.from_pretrained(google/owlvit-base-patch16)\\\\nmodel = OwlViTForObjectDetection.from_pretrained(google/owlvit-base-patch16)\\\\nurl = http://images.cocodataset.org/val2017/000000039769.jpg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\ntexts = [[a photo of a cat, a photo of a dog]]\\\\ninputs = processor(text=texts, images=image, return_tensors=pt)\\\\noutputs = model(**inputs)\\\\ntarget_sizes = torch.Tensor([image.size[::-1]])\\\\nresults = processor.post_process(outputs=outputs, target_sizes=target_sizes)\\', \\'performance\\': {\\'dataset\\': \\'COCO\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'OWL-ViT is a zero-shot text-conditioned object detection model that can be used to query an image with one or multiple text queries. OWL-ViT uses CLIP as its multi-modal backbone, with a ViT-like Transformer to get visual features and a causal language model to get the text features.\\'}', metadata={})]", "category": "generic"}
{"question_id": 22, "text": " Our client is an AI gaming company and we need to develop a bot for the game Valorant. The bot should detect objects like dropped spike, enemy, planted spike, and teammate within the game.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Object Detection\\', \\'api_name\\': \\'keremberke/yolov8m-valorant-detection\\', \\'api_call\\': \"YOLO(\\'keremberke/yolov8m-valorant-detection\\')\", \\'api_arguments\\': {\\'conf\\': 0.25, \\'iou\\': 0.45, \\'agnostic_nms\\': False, \\'max_det\\': 1000}, \\'python_environment_requirements\\': \\'pip install ultralyticsplus==0.0.23 ultralytics==8.0.21\\', \\'example_code\\': \"from ultralyticsplus import YOLO, render_result\\\\nmodel = YOLO(\\'keremberke/yolov8m-valorant-detection\\')\\\\nmodel.overrides[\\'conf\\'] = 0.25\\\\nmodel.overrides[\\'iou\\'] = 0.45\\\\nmodel.overrides[\\'agnostic_nms\\'] = False\\\\nmodel.overrides[\\'max_det\\'] = 1000\\\\nimage = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\\\\nresults = model.predict(image)\\\\nprint(results[0].boxes)\\\\nrender = render_result(model=model, image=image, result=results[0])\\\\nrender.show()\", \\'performance\\': {\\'dataset\\': \\'valorant-object-detection\\', \\'accuracy\\': 0.965}, \\'description\\': \\'A YOLOv8 model for object detection in Valorant game, trained on a custom dataset. It detects dropped spike, enemy, planted spike, and teammate objects.\\'}', metadata={})]", "category": "generic"}
{"question_id": 23, "text": " A client from real estate agency needs to get a list of objects present in a series of pictures to prepare their property listings.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'dreamlike-art/dreamlike-photoreal-2.0\\', \\'api_call\\': \"StableDiffusionPipeline.from_pretrained(\\'dreamlike-art/dreamlike-photoreal-2.0\\', torch_dtype=torch.float16)(prompt).images[0]\", \\'api_arguments\\': {\\'prompt\\': \\'photo, a church in the middle of a field of crops, bright cinematic lighting, gopro, fisheye lens\\'}, \\'python_environment_requirements\\': {\\'torch\\': \\'torch.float16\\', \\'diffusers\\': \\'StableDiffusionPipeline\\'}, \\'example_code\\': \\'from diffusers import StableDiffusionPipeline\\\\nimport torch\\\\nmodel_id = dreamlike-art/dreamlike-photoreal-2.0\\\\npipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16)\\\\npipe = pipe.to(cuda)\\\\nprompt = photo, a church in the middle of a field of crops, bright cinematic lighting, gopro, fisheye lens\\\\nimage = pipe(prompt).images[0]\\\\nimage.save(./result.jpg)\\', \\'performance\\': {\\'dataset\\': \\'Stable Diffusion 1.5\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'Dreamlike Photoreal 2.0 is a photorealistic model based on Stable Diffusion 1.5, made by dreamlike.art. It can be used to generate photorealistic images from text prompts.\\'}', metadata={})]", "category": "generic"}
{"question_id": 24, "text": " We are developing an application for smartphones which automatically separates elements in a user's photo, and we need to implement this feature.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Image-to-Text\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'kha-white/manga-ocr-base\\', \\'api_call\\': \"pipeline(\\'ocr\\', model=\\'kha-white/manga-ocr-base\\')\", \\'api_arguments\\': \\'image\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'manga109s\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Optical character recognition for Japanese text, with the main focus being Japanese manga. It uses Vision Encoder Decoder framework. Manga OCR can be used as a general purpose printed Japanese OCR, but its main goal was to provide a high quality text recognition, robust against various scenarios specific to manga: both vertical and horizontal text, text with furigana, text overlaid on images, wide variety of fonts and font styles, and low quality images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 25, "text": " I have a picture of a room demonstrating a mixture of objects. The model needs to seperate the objects and label them accordingly.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Joblib\\', \\'functionality\\': \\'Multi-class Classification\\', \\'api_name\\': \\'Alexei1/imdb\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'json\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'IMDB\\', \\'accuracy\\': 0.487}, \\'description\\': \\'A tabular classification model trained using AutoTrain for sentiment analysis on the IMDB dataset. The model has a CO2 emission of 0.0186 grams and an accuracy of 0.487.\\'}', metadata={})]", "category": "generic"}
{"question_id": 26, "text": " We want to randomly generate high-quality images of celebrity faces.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'gsdf/Counterfeit-V2.5\\', \\'api_call\\': \"pipeline(\\'text-to-image\\', model=\\'gsdf/Counterfeit-V2.5\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'((masterpiece,best quality)),1girl, solo, animal ears, rabbit, barefoot, knees up, dress, sitting, rabbit ears, short sleeves, looking at viewer, grass, short hair, smile, white hair, puffy sleeves, outdoors, puffy short sleeves, bangs, on ground, full body, animal, white dress, sunlight, brown eyes, dappled sunlight, day, depth of field\\', \\'performance\\': {\\'dataset\\': \\'EasyNegative\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Counterfeit-V2.5 is a text-to-image model that generates anime-style images based on text prompts. It has been updated for ease of use and can be used with negative prompts to create high-quality images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 27, "text": " Generate a new image based on the online database of bedroom art.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Unconditional Image Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Unconditional Image Generation\\', \\'api_name\\': \\'johnowhitaker/sd-class-wikiart-from-bedrooms\\', \\'api_call\\': \"DDPMPipeline.from_pretrained(\\'johnowhitaker/sd-class-wikiart-from-bedrooms\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'diffusers\\', \\'example_code\\': \"from diffusers import DDPMPipeline\\\\npipeline = DDPMPipeline.from_pretrained(\\'johnowhitaker/sd-class-wikiart-from-bedrooms\\')\\\\nimage = pipeline().images[0]\\\\nimage\", \\'performance\\': {\\'dataset\\': \\'https://huggingface.co/datasets/huggan/wikiart\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This model is a diffusion model initialized from https://huggingface.co/google/ddpm-bedroom-256 and trained for 5000 steps on https://huggingface.co/datasets/huggan/wikiart.\\'}', metadata={})]", "category": "generic"}
{"question_id": 28, "text": " I run an online store that sells butterfly-themed products. Please generate an image of a cute butterfly for our social media page.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Unconditional Image Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Unconditional Image Generation\\', \\'api_name\\': \\'ntrant7/sd-class-butterflies-32\\', \\'api_call\\': \"DDPMPipeline.from_pretrained(\\'ntrant7/sd-class-butterflies-32\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'diffusers\\'], \\'example_code\\': \"from diffusers import DDPMPipeline\\\\npipeline = DDPMPipeline.from_pretrained(\\'ntrant7/sd-class-butterflies-32\\')\\\\nimage = pipeline().images[0]\\\\nimage\", \\'performance\\': {\\'dataset\\': \\'Not specified\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'This model is a diffusion model for unconditional image generation of cute butterflies.\\'}', metadata={})]", "category": "generic"}
{"question_id": 29, "text": " We need a video-based AI model for security purposes. We want the AI to check and categorize footage based on existing security guidelines.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Video Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'videomae-base-finetuned-RealLifeViolenceSituations-subset\\', \\'api_call\\': \"AutoModelForVideoClassification.from_pretrained(\\'dangle124/videomae-base-finetuned-RealLifeViolenceSituations-subset\\')\", \\'api_arguments\\': {\\'model_name\\': \\'dangle124/videomae-base-finetuned-RealLifeViolenceSituations-subset\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'4.27.2\\', \\'pytorch\\': \\'1.13.1\\', \\'datasets\\': \\'2.10.1\\', \\'tokenizers\\': \\'0.13.2\\'}, \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'unknown\\', \\'accuracy\\': 0.9533}, \\'description\\': \\'This model is a fine-tuned version of MCG-NJU/videomae-base on an unknown dataset. It is trained for video classification task, specifically for RealLifeViolenceSituations.\\'}', metadata={})]", "category": "generic"}
{"question_id": 30, "text": " A new project demands to classify videos for a social media platform. Let us create a video classification pipeline.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'chavinlo/TempoFunk\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'chavinlo/TempoFunk\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Text-to-Video model using Hugging Face Transformers library. Model is capable of generating video content based on the input text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 31, "text": " I am an insurance adjustor. I need a zero-shot image classifier that will tell me whether a car has been involved in a major accident or had minor damages.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Zero-Shot Image Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Zero-Shot Image Classification\\', \\'api_name\\': \\'openai/clip-vit-large-patch14-336\\', \\'api_call\\': \"CLIPModel.from_pretrained(\\'openai/clip-vit-large-patch14\\').\", \\'api_arguments\\': \\'image_path, tokenizer, model\\', \\'python_environment_requirements\\': \\'Transformers 4.21.3, TensorFlow 2.8.2, Tokenizers 0.12.1\\', \\'example_code\\': \\'N/A\\', \\'performance\\': {\\'dataset\\': \\'unknown\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'This model was trained from scratch on an unknown dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 32, "text": " I want to analyze a medical image to find out if it's an X-ray, an MRI scan, or a CT scan.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Image-to-Text\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'kha-white/manga-ocr-base\\', \\'api_call\\': \"pipeline(\\'ocr\\', model=\\'kha-white/manga-ocr-base\\')\", \\'api_arguments\\': \\'image\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'manga109s\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Optical character recognition for Japanese text, with the main focus being Japanese manga. It uses Vision Encoder Decoder framework. Manga OCR can be used as a general purpose printed Japanese OCR, but its main goal was to provide a high quality text recognition, robust against various scenarios specific to manga: both vertical and horizontal text, text with furigana, text overlaid on images, wide variety of fonts and font styles, and low quality images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 33, "text": " We are building a quiz application where the image will be shown, and we have to choose a dressings matching that image. Please help in classifying the image.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Zero-Shot Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Zero-Shot Image Classification\\', \\'api_name\\': \\'clip-vit-base-patch32-ko\\', \\'api_call\\': \"pipeline(\\'zero-shot-image-classification\\', model=\\'Bingsu/clip-vit-base-patch32-ko\\')\", \\'api_arguments\\': {\\'images\\': \\'url\\', \\'candidate_labels\\': \\'Array of strings\\', \\'hypothesis_template\\': \\'String\\'}, \\'python_environment_requirements\\': [\\'transformers\\', \\'torch\\', \\'PIL\\', \\'requests\\'], \\'example_code\\': \"from transformers import pipeline\\\\nrepo = \\'Bingsu/clip-vit-base-patch32-ko\\'\\\\npipe = pipeline(\\'zero-shot-image-classification\\', model=repo)\\\\nurl = \\'http://images.cocodataset.org/val2017/000000039769.jpg\\'\\\\nresult = pipe(images=url, candidate_labels=[], hypothesis_template=\\'{}\\')\\\\nresult\", \\'performance\\': {\\'dataset\\': \\'AIHUB\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Korean CLIP model trained by Making Monolingual Sentence Embeddings Multilingual using Knowledge Distillation. It is a zero-shot image classification model that can be used to classify images without any training data.\\'}', metadata={})]", "category": "generic"}
{"question_id": 34, "text": " We're developing a chatbot that can quickly identify and describe images for our Chinese-speaking users.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Generation\\', \\'api_name\\': \\'mywateriswet/ShuanBot\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'mywateriswet/ShuanBot\\')\", \\'api_arguments\\': \\'message\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"response = chatbot(\\'What is your name?\\')\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'ShuanBot is a conversational chatbot model based on the GPT-2 architecture. It can be used for generating human-like responses in a chat context.\\'}', metadata={})]", "category": "generic"}
{"question_id": 35, "text": " We would like to understand the sentiment of user's messages in a customer support chat system.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Generation\\', \\'api_name\\': \\'mywateriswet/ShuanBot\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'mywateriswet/ShuanBot\\')\", \\'api_arguments\\': \\'message\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"response = chatbot(\\'What is your name?\\')\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'ShuanBot is a conversational chatbot model based on the GPT-2 architecture. It can be used for generating human-like responses in a chat context.\\'}', metadata={})]", "category": "generic"}
{"question_id": 36, "text": " As a book store owner, I want to classify customer reviews into positive and negative sentiments.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'michellejieli/NSFW_text_classifier\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'michellejieli/NSFW_text_classification\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'classifier(I see you\u2019ve set aside this special time to humiliate yourself in public.)\\', \\'performance\\': {\\'dataset\\': \\'Reddit posts\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'DistilBERT is a transformer model that performs sentiment analysis. I fine-tuned the model on Reddit posts with the purpose of classifying not safe for work (NSFW) content, specifically text that is considered inappropriate and unprofessional. The model predicts 2 classes, which are NSFW or safe for work (SFW). The model is a fine-tuned version of DistilBERT. It was fine-tuned on 14317 Reddit posts pulled from the Reddit API.\\'}', metadata={})]", "category": "generic"}
{"question_id": 37, "text": " I am the owner of a news website. I have several consumers' comments about our publishing news. I want to analyze the sentiments of these comments.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'michellejieli/NSFW_text_classifier\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'michellejieli/NSFW_text_classification\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'classifier(I see you\u2019ve set aside this special time to humiliate yourself in public.)\\', \\'performance\\': {\\'dataset\\': \\'Reddit posts\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'DistilBERT is a transformer model that performs sentiment analysis. I fine-tuned the model on Reddit posts with the purpose of classifying not safe for work (NSFW) content, specifically text that is considered inappropriate and unprofessional. The model predicts 2 classes, which are NSFW or safe for work (SFW). The model is a fine-tuned version of DistilBERT. It was fine-tuned on 14317 Reddit posts pulled from the Reddit API.\\'}', metadata={})]", "category": "generic"}
{"question_id": 38, "text": " Our business is expanding to international markets. Analyze the sentiment of the following customer review to better understand their satisfaction with our product: \\\"\\u00a1Esto es maravilloso! Me encanta.\\\"\\n###Input: \\\"\\u00a1Esto es maravilloso! Me encanta.\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'finiteautomata/beto-sentiment-analysis\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'finiteautomata/beto-sentiment-analysis\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'Hugging Face Transformers library\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'TASS 2020 corpus\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Model trained with TASS 2020 corpus (around ~5k tweets) of several dialects of Spanish. Base model is BETO, a BERT model trained in Spanish. Uses POS, NEG, NEU labels.\\'}', metadata={})]", "category": "generic"}
{"question_id": 39, "text": " We are a forum moderator team looking for a solution to classify comments into toxic or non-toxic categories.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'martin-ha/toxic-comment-model\\', \\'api_call\\': \"pipeline(model=\\'martin-ha/toxic-comment-model\\')\", \\'api_arguments\\': {\\'model_path\\': \\'martin-ha/toxic-comment-model\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import AutoModelForSequenceClassification, AutoTokenizer, TextClassificationPipeline\\\\nmodel_path = martin-ha/toxic-comment-model\\\\ntokenizer = AutoTokenizer.from_pretrained(model_path)\\\\nmodel = AutoModelForSequenceClassification.from_pretrained(model_path)\\\\npipeline = TextClassificationPipeline(model=model, tokenizer=tokenizer)\\\\nprint(pipeline(\\'This is a test text.\\'))\", \\'performance\\': {\\'dataset\\': \\'held-out test set\\', \\'accuracy\\': 0.94, \\'f1-score\\': 0.59}, \\'description\\': \\'This model is a fine-tuned version of the DistilBERT model to classify toxic comments.\\'}', metadata={})]", "category": "generic"}
{"question_id": 40, "text": " My company is launching a social media campaign. We need an AI-based system that would automatically analyze the sentiment of any user-generated reviews or tweets concerning our product.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'cardiffnlp/twitter-roberta-base-sentiment-latest\\', \\'api_call\\': \"pipeline(sentiment-analysis, model=AutoModel.from_pretrained(\\'cardiffnlp/twitter-roberta-base-sentiment-latest\\'), tokenizer=AutoTokenizer.from_pretrained(\\'cardiffnlp/twitter-roberta-base-sentiment-latest\\'))\", \\'api_arguments\\': {\\'model\\': \\'model_path\\', \\'tokenizer\\': \\'model_path\\'}, \\'python_environment_requirements\\': [\\'transformers\\', \\'numpy\\', \\'scipy\\'], \\'example_code\\': \\'from transformers import pipeline\\\\nsentiment_task = pipeline(sentiment-analysis, model=model_path, tokenizer=model_path)\\\\nsentiment_task(Covid cases are increasing fast!)\\', \\'performance\\': {\\'dataset\\': \\'tweet_eval\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This is a RoBERTa-base model trained on ~124M tweets from January 2018 to December 2021, and finetuned for sentiment analysis with the TweetEval benchmark. The model is suitable for English.\\'}', metadata={})]", "category": "generic"}
{"question_id": 41, "text": " I have jobs descriptions in French for sales manager, please highlight names of organizations or cities within the text.\\n###Input: \\\"La soci\\u00e9t\\u00e9 de Paris est sp\\u00e9cialis\\u00e9e dans la vente de v\\u00e9hicules \\u00e9lectriques. Responsable des ventes, vous travaillerez au sein d'une \\u00e9quipe dynamique dans l'agence de Lyon. Vous \\u00eates charg\\u00e9(e) de d\\u00e9velopper le portefeuille client et d'assurer la satisfaction des clients existants. Dans ce contexte, vous devrez travailler en lien \\u00e9troit avec le directeur commercial et les autres \\u00e9quipes de l'entreprise. Une exp\\u00e9rience pr\\u00e9alable chez Renault est un atout.\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Audio-to-Audio\\', \\'framework\\': \\'Fairseq\\', \\'functionality\\': \\'speech-to-speech-translation\\', \\'api_name\\': \\'facebook/textless_sm_en_fr\\', \\'api_call\\': \"load_model_ensemble_and_task_from_hf_hub(\\'facebook/textless_sm_en_fr\\')\", \\'api_arguments\\': [\\'input_file\\'], \\'python_environment_requirements\\': [\\'huggingface_hub\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'This model is a speech-to-speech translation model trained by Facebook. It is designed for translating English speech to French speech.\\'}', metadata={})]", "category": "generic"}
{"question_id": 42, "text": " In order to have a better understanding of our clients, I'd like to identify the names of people and organizations mentioned in the following customer review.\\n###Input: \\\"I recently purchased a MacBook Pro from Apple Inc. and had a fantastic customer support experience. John from their tech support team was incredibly helpful and professional.\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'Zixtrauce/JohnBot\\', \\'api_call\\': \"AutoModelForCausalLM.from_pretrained(\\'Zixtrauce/JohnBot\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'Input a message to start chatting with Zixtrauce/JohnBot.\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'JohnBot is a conversational model based on the gpt2 architecture and trained using the Hugging Face Transformers library. It can be used for generating text responses in a chat-based interface.\\'}', metadata={})]", "category": "generic"}
{"question_id": 43, "text": " I am building a social media app that requires people to write fascinating stories rather than boring sentences. Detect named entities in a sentence by using an NER model.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Token Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Named Entity Recognition\\', \\'api_name\\': \\'dslim/bert-base-NER-uncased\\', \\'api_call\\': \"pipeline(\\'ner\\', model=\\'dslim/bert-base-NER-uncased\\')\", \\'api_arguments\\': {}, \\'python_environment_requirements\\': {\\'transformers\\': \\'>=4.0.0\\'}, \\'example_code\\': \"nlp(\\'My name is John and I live in New York.\\')\", \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A pretrained BERT model for Named Entity Recognition (NER) on uncased text. It can be used to extract entities such as person names, locations, and organizations from text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 44, "text": " We have a large dataset of customer orders in the form of a table. Help us answer questions about this data.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Table Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Table Question Answering\\', \\'api_name\\': \\'google/tapas-small-finetuned-sqa\\', \\'api_call\\': \"pipeline(\\'table-question-answering\\', model=\\'google/tapas-small-finetuned-sqa\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'msr_sqa\\', \\'accuracy\\': 0.6155}, \\'description\\': \\'TAPAS small model fine-tuned on Sequential Question Answering (SQA). It uses relative position embeddings (i.e. resetting the position index at every cell of the table).\\'}', metadata={})]", "category": "generic"}
{"question_id": 45, "text": " Gather information about annual income and age demographics of employees to predict retirement patterns. Make sure to identify top employees for potential promotions.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Joblib\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'abhishek/autotrain-adult-census-xgboost\\', \\'api_call\\': \"AutoModel.from_pretrained(\\'abhishek/autotrain-adult-census-xgboost\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'json\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'scikit-learn/adult-census-income\\', \\'accuracy\\': 0.8750191923844618}, \\'description\\': \"A binary classification model trained on the Adult Census Income dataset using the XGBoost algorithm. The model predicts whether an individual\\'s income is above or below $50,000 per year.\"}', metadata={})]", "category": "generic"}
{"question_id": 46, "text": " To track our sales data, we need to find total sales of a specific product based on a table containing sales information per week.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Table Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Table Question Answering\\', \\'api_name\\': \\'google/tapas-small-finetuned-sqa\\', \\'api_call\\': \"pipeline(\\'table-question-answering\\', model=\\'google/tapas-small-finetuned-sqa\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'msr_sqa\\', \\'accuracy\\': 0.6155}, \\'description\\': \\'TAPAS small model fine-tuned on Sequential Question Answering (SQA). It uses relative position embeddings (i.e. resetting the position index at every cell of the table).\\'}', metadata={})]", "category": "generic"}
{"question_id": 47, "text": " I have a table containing information about various animals and their important characteristics. I need the system to answer a query to provide information about the tallest animal in the table.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Table Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Table Question Answering\\', \\'api_name\\': \\'google/tapas-small-finetuned-sqa\\', \\'api_call\\': \"pipeline(\\'table-question-answering\\', model=\\'google/tapas-small-finetuned-sqa\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'msr_sqa\\', \\'accuracy\\': 0.6155}, \\'description\\': \\'TAPAS small model fine-tuned on Sequential Question Answering (SQA). It uses relative position embeddings (i.e. resetting the position index at every cell of the table).\\'}', metadata={})]", "category": "generic"}
{"question_id": 48, "text": " You are building an app that allows users to find quick answers to textbook questions. Users will send a message with the question, and the answer should be detected directly from the textbook content.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'distilbert-base-uncased-distilled-squad\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=\\'distilbert-base-uncased-distilled-squad\\')\", \\'api_arguments\\': [\\'question\\', \\'context\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nquestion_answerer = pipeline(question-answering, model=\\'distilbert-base-uncased-distilled-squad\\')\\\\ncontext = r\\\\n... Extractive Question Answering is the task of extracting an answer from a text given a question. An example of a\\\\n... question answering dataset is the SQuAD dataset, which is entirely based on that task. If you would like to fine-tune\\\\n... a model on a SQuAD task, you may leverage the examples/pytorch/question-answering/run_squad.py script.\\\\n... \\\\nresult = question_answerer(question=What is a good example of a question answering dataset?, context=context)\\\\nprint(\\\\n... fAnswer: \\'{result[\\'answer\\']}\\', score: {round(result[\\'score\\'], 4)}, start: {result[\\'start\\']}, end: {result[\\'end\\']}\\\\n...)\", \\'performance\\': {\\'dataset\\': \\'SQuAD v1.1\\', \\'accuracy\\': \\'86.9 F1 score\\'}, \\'description\\': \"DistilBERT base uncased distilled SQuAD is a fine-tuned version of DistilBERT-base-uncased for the task of question answering. It has 40% less parameters than bert-base-uncased, runs 60% faster while preserving over 95% of BERT\\'s performances as measured on the GLUE language understanding benchmark.\"}', metadata={})]", "category": "generic"}
{"question_id": 49, "text": " We want to make sure clarify some questions about the legal implications of a new partnership contract for a real estate development project.\\n###Input: We hereby grant the Licensee the exclusive right to develop, construct, operate and promote the Project, as well as to manage the daily operations of the Licensed Facilities during the Term. In consideration for the grant of the License, the Licensee shall pay to the Licensor the full amount of Ten Million (10,000,000) Dollars within thirty (30) days after the execution hereof.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'impira/layoutlm-invoices\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=\\'impira/layoutlm-invoices\\')\", \\'api_arguments\\': \\'question, context\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"qa_pipeline(question=\\'your question\\', context=\\'your document context\\')\", \\'performance\\': {\\'dataset\\': \\'proprietary dataset of invoices, SQuAD2.0, and DocVQA\\', \\'accuracy\\': \\'not provided\\'}, \\'description\\': \\'This is a fine-tuned version of the multi-modal LayoutLM model for the task of question answering on invoices and other documents. It has been fine-tuned on a proprietary dataset of invoices as well as both SQuAD2.0 and DocVQA for general comprehension. Unlike other QA models, which can only extract consecutive tokens (because they predict the start and end of a sequence), this model can predict longer-range, non-consecutive sequences with an additional classifier head.\\'}', metadata={})]", "category": "generic"}
{"question_id": 50, "text": " Help me setup a tinyroberta model from deepset for Question and Answer. Provide a sample input and output.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'deepset/tinyroberta-squad2\\', \\'api_call\\': \"AutoModelForQuestionAnswering.from_pretrained(\\'deepset/tinyroberta-squad2\\')\", \\'api_arguments\\': {\\'model_name_or_path\\': \\'deepset/tinyroberta-squad2\\', \\'question\\': \\'Why is model conversion important?\\', \\'context\\': \\'The option to convert models between FARM and transformers gives freedom to the user and let people easily switch between frameworks.\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import AutoModelForQuestionAnswering, AutoTokenizer, pipeline\\\\nmodel_name = deepset/tinyroberta-squad2\\\\nnlp = pipeline(\\'question-answering\\', model=model_name, tokenizer=model_name)\\\\nQA_input = {\\\\n \\'question\\': \\'Why is model conversion important?\\',\\\\n \\'context\\': \\'The option to convert models between FARM and transformers gives freedom to the user and let people easily switch between frameworks.\\'\\\\n}\\\\nres = nlp(QA_input)\", \\'performance\\': {\\'dataset\\': \\'squad_v2\\', \\'accuracy\\': {\\'exact\\': 78.69114798281817, \\'f1\\': 81.9198998536977}}, \\'description\\': \\'This is the distilled version of the deepset/roberta-base-squad2 model. This model has a comparable prediction quality and runs at twice the speed of the base model.\\'}', metadata={})]", "category": "generic"}
{"question_id": 51, "text": " I want to build a tool to answer questions automatically from a given document. Which model do you recommend for this task?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'distilbert-base-uncased-distilled-squad\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=\\'distilbert-base-uncased-distilled-squad\\')\", \\'api_arguments\\': [\\'question\\', \\'context\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nquestion_answerer = pipeline(question-answering, model=\\'distilbert-base-uncased-distilled-squad\\')\\\\ncontext = r\\\\n... Extractive Question Answering is the task of extracting an answer from a text given a question. An example of a\\\\n... question answering dataset is the SQuAD dataset, which is entirely based on that task. If you would like to fine-tune\\\\n... a model on a SQuAD task, you may leverage the examples/pytorch/question-answering/run_squad.py script.\\\\n... \\\\nresult = question_answerer(question=What is a good example of a question answering dataset?, context=context)\\\\nprint(\\\\n... fAnswer: \\'{result[\\'answer\\']}\\', score: {round(result[\\'score\\'], 4)}, start: {result[\\'start\\']}, end: {result[\\'end\\']}\\\\n...)\", \\'performance\\': {\\'dataset\\': \\'SQuAD v1.1\\', \\'accuracy\\': \\'86.9 F1 score\\'}, \\'description\\': \"DistilBERT base uncased distilled SQuAD is a fine-tuned version of DistilBERT-base-uncased for the task of question answering. It has 40% less parameters than bert-base-uncased, runs 60% faster while preserving over 95% of BERT\\'s performances as measured on the GLUE language understanding benchmark.\"}', metadata={})]", "category": "generic"}
{"question_id": 52, "text": " We have a French news agency and we want to categorize the news articles based on sports, politics, and science.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Zero-Shot Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Zero-Shot Classification\\', \\'api_name\\': \\'BaptisteDoyen/camembert-base-xnli\\', \\'api_call\\': \"pipeline(\\'zero-shot-classification\\', model=\\'BaptisteDoyen/camembert-base-xnli\\')\", \\'api_arguments\\': {\\'sequence\\': \\'str\\', \\'candidate_labels\\': \\'List[str]\\', \\'hypothesis_template\\': \\'str\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"sequence = L\\'\u00e9quipe de France joue aujourd\\'hui au Parc des Princes\\\\ncandidate_labels = [sport,politique,science]\\\\nhypothesis_template = Ce texte parle de {}.\\\\nclassifier(sequence, candidate_labels, hypothesis_template=hypothesis_template)\", \\'performance\\': {\\'dataset\\': \\'xnli\\', \\'accuracy\\': {\\'validation\\': 81.4, \\'test\\': 81.7}}, \\'description\\': \\'Camembert-base model fine-tuned on french part of XNLI dataset. One of the few Zero-Shot classification models working on French.\\'}', metadata={})]", "category": "generic"}
{"question_id": 53, "text": " I need a solution to detect whether a piece of news is talking about technology, sports, or politics.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'michellejieli/NSFW_text_classifier\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'michellejieli/NSFW_text_classification\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'classifier(I see you\u2019ve set aside this special time to humiliate yourself in public.)\\', \\'performance\\': {\\'dataset\\': \\'Reddit posts\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'DistilBERT is a transformer model that performs sentiment analysis. I fine-tuned the model on Reddit posts with the purpose of classifying not safe for work (NSFW) content, specifically text that is considered inappropriate and unprofessional. The model predicts 2 classes, which are NSFW or safe for work (SFW). The model is a fine-tuned version of DistilBERT. It was fine-tuned on 14317 Reddit posts pulled from the Reddit API.\\'}', metadata={})]", "category": "generic"}
{"question_id": 54, "text": " I want to build a chatbot that is used by language learners who want to communicate in French while they only know English. Generate a response for an English message.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Generation\\', \\'api_name\\': \\'mywateriswet/ShuanBot\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'mywateriswet/ShuanBot\\')\", \\'api_arguments\\': \\'message\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"response = chatbot(\\'What is your name?\\')\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'ShuanBot is a conversational chatbot model based on the GPT-2 architecture. It can be used for generating human-like responses in a chat context.\\'}', metadata={})]", "category": "generic"}
{"question_id": 55, "text": " Translate the following text from French to English: \\u201cLe syst\\u00e8me \\u00e9ducatif fran\\u00e7ais est compos\\u00e9 d'\\u00e9coles maternelles, d'\\u00e9coles \\u00e9l\\u00e9mentaires, de coll\\u00e8ges et de lyc\\u00e9es.\\u201d\\n###Input: Le syst\\u00e8me \\u00e9ducatif fran\\u00e7ais est compos\\u00e9 d'\\u00e9coles maternelles, d'\\u00e9coles \\u00e9l\\u00e9mentaires, de coll\\u00e8ges et de lyc\\u00e9es.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Translation\\', \\'api_name\\': \\'opus-mt-fr-en\\', \\'api_call\\': \"pipeline(\\'translation_fr_to_en\\', model=\\'Helsinki-NLP/opus-mt-fr-en\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\', \\'torch\\'], \\'example_code\\': \"translation_pipeline(\\'Bonjour, comment \u00e7a va?\\')\", \\'performance\\': {\\'dataset\\': \\'opus\\', \\'accuracy\\': {\\'BLEU\\': {\\'newsdiscussdev2015-enfr.fr.en\\': 33.1, \\'newsdiscusstest2015-enfr.fr.en\\': 38.7, \\'newssyscomb2009.fr.en\\': 30.3, \\'news-test2008.fr.en\\': 26.2, \\'newstest2009.fr.en\\': 30.2, \\'newstest2010.fr.en\\': 32.2, \\'newstest2011.fr.en\\': 33.0, \\'newstest2012.fr.en\\': 32.8, \\'newstest2013.fr.en\\': 33.9, \\'newstest2014-fren.fr.en\\': 37.8, \\'Tatoeba.fr.en\\': 57.5}}}, \\'description\\': \\'Helsinki-NLP/opus-mt-fr-en is a machine translation model trained to translate from French to English. It is based on the Marian NMT framework and trained on the OPUS dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 56, "text": " I want to translate a text from one language to another.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Translation\\', \\'api_name\\': \\'Helsinki-NLP/opus-mt-en-fr\\', \\'api_call\\': \"translate(\\'input_text\\', model=\\'Helsinki-NLP/opus-mt-en-fr\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'opus\\', \\'accuracy\\': {\\'BLEU\\': {\\'newsdiscussdev2015-enfr.en.fr\\': 33.8, \\'newsdiscusstest2015-enfr.en.fr\\': 40.0, \\'newssyscomb2009.en.fr\\': 29.8, \\'news-test2008.en.fr\\': 27.5, \\'newstest2009.en.fr\\': 29.4, \\'newstest2010.en.fr\\': 32.7, \\'newstest2011.en.fr\\': 34.3, \\'newstest2012.en.fr\\': 31.8, \\'newstest2013.en.fr\\': 33.2, \\'Tatoeba.en.fr\\': 50.5}}}, \\'description\\': \\'Helsinki-NLP/opus-mt-en-fr is a translation model that translates English text to French using the Hugging Face Transformers library. It is based on the OPUS dataset and uses a transformer-align architecture with normalization and SentencePiece pre-processing.\\'}', metadata={})]", "category": "generic"}
{"question_id": 57, "text": " Our team member has written a long article that needs to be published on a company blog. Can you provide a shorter summary to be used as a snippet on the landing page?\\n###Input: \\\"Apple Inc. reported its quarterly earnings results yesterday. The company posted a record-breaking revenue of $123.9 billion for the first quarter of 2022, up by 11% from the same period last year. The increase was fueled by stronger demand for iPhones, iPads, and Macs, as well as continued growth in its services segment. Apple's operating profit for the quarter came in at $38.3 billion, up 17% from a year earlier. The results surpassed analysts' expectations, who had anticipated revenue of around $118 billion. This strong performance is largely attributed to the successful launch of the iPhone 13, which has enjoyed robust sales since its debut in September. Apple CEO Tim Cook said in a statement, \\\"Our record-breaking quarter reflects the strength of our entire ecosystem, from our innovative products and services to the unmatched dedication of our teams around the world.\\\" Despite the ongoing global supply chain disruptions, Apple has managed to maintain its growth trajectory, thanks in part to its vertically integrated operations and nimble supply chain management. The company is expected to face stiffer competition going forward, particularly in the smartphone market, as rivals introduce new devices and increased pricing pressures.\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'text2text-generation\\', \\'api_name\\': \\'financial-summarization-pegasus\\', \\'api_call\\': \"PegasusForConditionalGeneration.from_pretrained(\\'human-centered-summarization/financial-summarization-pegasus\\')\", \\'api_arguments\\': [\\'model_name\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'from transformers import PegasusTokenizer, PegasusForConditionalGeneration, TFPegasusForConditionalGeneration\\\\nmodel_name = human-centered-summarization/financial-summarization-pegasus\\\\ntokenizer = PegasusTokenizer.from_pretrained(model_name)\\\\nmodel = PegasusForConditionalGeneration.from_pretrained(model_name)\\\\ntext_to_summarize = National Commercial Bank (NCB), Saudi Arabia\u2019s largest lender by assets, agreed to buy rival Samba Financial Group for $15 billion in the biggest banking takeover this year.NCB will pay 28.45 riyals ($7.58) for each Samba share, according to a statement on Sunday, valuing it at about 55.7 billion riyals. NCB will offer 0.739 new shares for each Samba share, at the lower end of the 0.736-0.787 ratio the banks set when they signed an initial framework agreement in June.The offer is a 3.5% premium to Samba\u2019s Oct. 8 closing price of 27.50 riyals and about 24% higher than the level the shares traded at before the talks were made public. Bloomberg News first reported the merger discussions.The new bank will have total assets of more than $220 billion, creating the Gulf region\u2019s third-largest lender. The entity\u2019s $46 billion market capitalization nearly matches that of Qatar National Bank QPSC, which is still the Middle East\u2019s biggest lender with about $268 billion of assets.\\\\ninput_ids = tokenizer(text_to_summarize, return_tensors=pt).input_ids\\\\noutput = model.generate(input_ids, max_length=32, num_beams=5, early_stopping=True)\\\\nprint(tokenizer.decode(output[0], skip_special_tokens=True))\\', \\'performance\\': {\\'dataset\\': \\'xsum\\', \\'accuracy\\': {\\'ROUGE-1\\': 35.206, \\'ROUGE-2\\': 16.569, \\'ROUGE-L\\': 30.128, \\'ROUGE-LSUM\\': 30.171}}, \\'description\\': \\'This model was fine-tuned on a novel financial news dataset, which consists of 2K articles from Bloomberg, on topics such as stock, markets, currencies, rate and cryptocurrencies. It is based on the PEGASUS model and in particular PEGASUS fine-tuned on the Extreme Summarization (XSum) dataset: google/pegasus-xsum model. PEGASUS was originally proposed by Jingqing Zhang, Yao Zhao, Mohammad Saleh and Peter J. Liu in PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization.\\'}', metadata={})]", "category": "generic"}
{"question_id": 58, "text": " Write a summary of a conference held by the World Health Organization discussing the impacts of climate change on human health.\\n###Input: Over the past week, the World Health Organization held a conference discussing the impacts of climate change on human health. The conference brought together leading experts from around the world to examine the current problems affecting people's health due to changing environmental conditions. The topics of discussion included increased occurrence of heat-related illnesses, heightened rates of vector-borne diseases, and the growing problem of air pollution. The conference concluded with a call to action for governments and organizations to invest in mitigating and adapting to the negative consequences of climate change for the sake of public health.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Abstractive Text Summarization\\', \\'api_name\\': \\'plguillou/t5-base-fr-sum-cnndm\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'plguillou/t5-base-fr-sum-cnndm\\')\", \\'api_arguments\\': {\\'input_text\\': \\'summarize: ARTICLE\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'from transformers import T5Tokenizer, T5ForConditionalGeneration\\'}, \\'example_code\\': \\'tokenizer = T5Tokenizer.from_pretrained(plguillou/t5-base-fr-sum-cnndm)\\\\nmodel = T5ForConditionalGeneration.from_pretrained(plguillou/t5-base-fr-sum-cnndm)\\', \\'performance\\': {\\'dataset\\': \\'cnn_dailymail\\', \\'ROUGE-1\\': 44.5252, \\'ROUGE-2\\': 22.652, \\'ROUGE-L\\': 29.8866}, \\'description\\': \\'This model is a T5 Transformers model (JDBN/t5-base-fr-qg-fquad) that was fine-tuned in French for abstractive text summarization.\\'}', metadata={})]", "category": "generic"}
{"question_id": 59, "text": " Please provide a brief overview of a news article.\\n###Input: A new study suggests that eating chocolate at least once a week can lead to better cognition. The study, published in the journal Appetite, analyzed data from over 900 adults and found that individuals who consumed chocolate at least once a week performed better on cognitive tests than those who consumed chocolate less frequently. Researchers believe that the beneficial effects of chocolate on cognition may be due to the presence of flavonoids, which have been shown to be antioxidant-rich and to improve brain blood flow.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Binary Classification\\', \\'api_name\\': \\'desertdev/autotrain-imdb-sentiment-analysis-44994113085\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'desertdev/autotrain-data-imdb-sentiment-analysis\\', \\'accuracy\\': 0.565}, \\'description\\': \\'A binary classification model trained on the IMDb sentiment analysis dataset using AutoTrain. The model is capable of predicting sentiment (positive or negative) for movie reviews.\\'}', metadata={})]", "category": "generic"}
{"question_id": 60, "text": " I developed a document generation app, I need to create a summary of a long article given as input to provide to my users before they read the full article.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'pszemraj/long-t5-tglobal-base-16384-book-summary\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'pszemraj/long-t5-tglobal-base-16384-book-summary\\')\", \\'api_arguments\\': [\\'long_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nimport torch\\\\nsummarizer = pipeline(\\\\n summarization,\\\\n pszemraj/long-t5-tglobal-base-16384-book-summary,\\\\n device=0 if torch.cuda.is_available() else -1,\\\\n)\\\\nlong_text = Here is a lot of text I don\\'t want to read. Replace me\\\\nresult = summarizer(long_text)\\\\nprint(result[0][summary_text])\", \\'performance\\': {\\'dataset\\': \\'kmfoda/booksum\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.408, \\'ROUGE-2\\': 6.065, \\'ROUGE-L\\': 16.721, \\'ROUGE-LSUM\\': 33.34}}, \\'description\\': \\'A fine-tuned version of google/long-t5-tglobal-base on the kmfoda/booksum dataset, which can be used to summarize long text and generate SparkNotes-esque summaries of arbitrary topics. The model generalizes reasonably well to academic and narrative text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 61, "text": " We need a quick summary of a news article we found online. Can you help us with that?\\n###Input: Videos that say approved vaccines are dangerous and cause autism, cancer or infertility are among those that will be taken down, the company said. The policy includes the termination of accounts of anti-vaccine influencers. Tech giants have been criticised for not doing more to counter false health information on their sites. In July, US President Joe Biden said social media platforms were largely responsible for people's scepticism in getting vaccinated by spreading misinformation, and appealed for them to address the issue. YouTube, which is owned by Google, said 130,000 videos were removed from its platform since last year, when it implemented a ban on content spreading misinformation about Covid vaccines. In a blog post, the company said it had seen false claims about Covid jabs spill over into misinformation about vaccines in general. The new policy covers long-approved vaccines, such as those against measles or hepatitis B. We're expanding our medical misinformation policies on YouTube with new guidelines on currently administered vaccines that are approved and confirmed to be safe and effective by local health authorities and the WHO, the post said, referring to the World Health Organization.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'text2text-generation\\', \\'api_name\\': \\'csebuetnlp/mT5_multilingual_XLSum\\', \\'api_call\\': \"AutoModelForSeq2SeqLM.from_pretrained(\\'csebuetnlp/mT5_multilingual_XLSum\\')\", \\'api_arguments\\': [\\'model_name\\'], \\'python_environment_requirements\\': [\\'transformers==4.11.0.dev0\\'], \\'example_code\\': \"import re\\\\nfrom transformers import AutoTokenizer, AutoModelForSeq2SeqLM\\\\nWHITESPACE_HANDLER = lambda k: re.sub(\\'\\\\\\\\s+\\', \\' \\', re.sub(\\'\\\\\\\\n+\\', \\' \\', k.strip()))\\\\narticle_text = Videos that say approved vaccines are dangerous and cause autism, cancer or infertility are among those that will be taken down, the company said. The policy includes the termination of accounts of anti-vaccine influencers. Tech giants have been criticised for not doing more to counter false health information on their sites. In July, US President Joe Biden said social media platforms were largely responsible for people\\'s scepticism in getting vaccinated by spreading misinformation, and appealed for them to address the issue. YouTube, which is owned by Google, said 130,000 videos were removed from its platform since last year, when it implemented a ban on content spreading misinformation about Covid vaccines. In a blog post, the company said it had seen false claims about Covid jabs spill over into misinformation about vaccines in general. The new policy covers long-approved vaccines, such as those against measles or hepatitis B. We\\'re expanding our medical misinformation policies on YouTube with new guidelines on currently administered vaccines that are approved and confirmed to be safe and effective by local health authorities and the WHO, the post said, referring to the World Health Organization.\\\\nmodel_name = csebuetnlp/mT5_multilingual_XLSum\\\\ntokenizer = AutoTokenizer.from_pretrained(model_name)\\\\nmodel = AutoModelForSeq2SeqLM.from_pretrained(model_name)\\\\ninput_ids = tokenizer(\\\\n [WHITESPACE_HANDLER(article_text)],\\\\n return_tensors=pt,\\\\n padding=max_length,\\\\n truncation=True,\\\\n max_length=512\\\\n)[input_ids]\\\\noutput_ids = model.generate(\\\\n input_ids=input_ids,\\\\n max_length=84,\\\\n no_repeat_ngram_size=2,\\\\n num_beams=4\\\\n)[0]\\\\nsummary = tokenizer.decode(\\\\n output_ids,\\\\n skip_special_tokens=True,\\\\n clean_up_tokenization_spaces=False\\\\n)\\\\nprint(summary)\", \\'performance\\': {\\'dataset\\': \\'xsum\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.5, \\'ROUGE-2\\': 13.934, \\'ROUGE-L\\': 28.988, \\'ROUGE-LSUM\\': 28.996, \\'loss\\': 2.067, \\'gen_len\\': 26.973}}, \\'description\\': \\'This repository contains the mT5 checkpoint finetuned on the 45 languages of XL-Sum dataset. It is a multilingual abstractive summarization model that supports text-to-text generation for 43 languages.\\'}', metadata={})]", "category": "generic"}
{"question_id": 62, "text": " We'd like our chatbot to act as a fictional character for engaging with our users.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Generation\\', \\'api_name\\': \\'mywateriswet/ShuanBot\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'mywateriswet/ShuanBot\\')\", \\'api_arguments\\': \\'message\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"response = chatbot(\\'What is your name?\\')\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'ShuanBot is a conversational chatbot model based on the GPT-2 architecture. It can be used for generating human-like responses in a chat context.\\'}', metadata={})]", "category": "generic"}
{"question_id": 63, "text": " Write a story about a spaceship journey to a distant planet in search of a new home for humanity.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'prompthero/openjourney\\', \\'api_call\\': \"StableDiffusionPipeline.from_pretrained(\\'prompthero/openjourney\\', torch_dtype=torch.float16)\", \\'api_arguments\\': {\\'prompt\\': \\'string\\'}, \\'python_environment_requirements\\': [\\'diffusers\\', \\'torch\\'], \\'example_code\\': \\'from diffusers import StableDiffusionPipeline\\\\nimport torch\\\\nmodel_id = prompthero/openjourney\\\\npipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16)\\\\npipe = pipe.to(cuda)\\\\nprompt = retro serie of different cars with different colors and shapes, mdjrny-v4 style\\\\nimage = pipe(prompt).images[0]\\\\nimage.save(./retro_cars.png)\\', \\'performance\\': {\\'dataset\\': \\'Midjourney images\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'Openjourney is an open source Stable Diffusion fine-tuned model on Midjourney images, by PromptHero. It can be used for generating AI art based on text prompts.\\'}', metadata={})]", "category": "generic"}
{"question_id": 64, "text": " I want to write a story about a brave knight and a dragon but I'm unable to come up with a good start. Help me with that.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'text2text-generation\\', \\'api_name\\': \\'tuner007/pegasus_summarizer\\', \\'api_call\\': \"PegasusForConditionalGeneration.from_pretrained(\\'tuner007/pegasus_summarizer\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'pip install sentencepiece\\'], \\'example_code\\': \"context = \\\\nIndia wicket-keeper batsman Rishabh Pant has said someone from the crowd threw a ball on pacer Mohammed Siraj while he was fielding in the ongoing third Test against England on Wednesday. Pant revealed the incident made India skipper Virat Kohli upset. I think, somebody threw a ball inside, at Siraj, so he [Kohli] was upset, said Pant in a virtual press conference after the close of the first day\\'s play.You can say whatever you want to chant, but don\\'t throw things at the fielders and all those things. It is not good for cricket, I guess, he added.In the third session of the opening day of the third Test, a section of spectators seemed to have asked Siraj the score of the match to tease the pacer. The India pacer however came with a brilliant reply as he gestured 1-0 (India leading the Test series) towards the crowd.Earlier this month, during the second Test match, there was some bad crowd behaviour on a show as some unruly fans threw champagne corks at India batsman KL Rahul.Kohli also intervened and he was seen gesturing towards the opening batsman to know more about the incident. An over later, the TV visuals showed that many champagne corks were thrown inside the playing field, and the Indian players were visibly left frustrated.Coming back to the game, after bundling out India for 78, openers Rory Burns and Haseeb Hameed ensured that England took the honours on the opening day of the ongoing third Test.At stumps, England\\'s score reads 120/0 and the hosts have extended their lead to 42 runs. For the Three Lions, Burns (52) and Hameed (60) are currently unbeaten at the crease.Talking about the pitch on opening day, Pant said, They took the heavy roller, the wicket was much more settled down, and they batted nicely also, he said. But when we batted, the wicket was slightly soft, and they bowled in good areas, but we could have applied [ourselves] much better.Both England batsmen managed to see off the final session and the hosts concluded the opening day with all ten wickets intact, extending the lead to 42.(ANI)\\\\n\\\\nget_response(context)\", \\'performance\\': {\\'dataset\\': \\'cnn_dailymail\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.604, \\'ROUGE-2\\': 14.64, \\'ROUGE-L\\': 23.884, \\'ROUGE-LSUM\\': 32.902, \\'loss\\': 2.576, \\'gen_len\\': 76.398}}, \\'description\\': \\'PEGASUS fine-tuned for summarization\\'}', metadata={})]", "category": "generic"}
{"question_id": 65, "text": " I need a text analysis tool that can automatically predict the most plausible missing text in a given sentence.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentence Correction\\', \\'api_name\\': \\'flexudy/t5-base-multi-sentence-doctor\\', \\'api_call\\': \"AutoModelWithLMHead.from_pretrained(\\'flexudy/t5-base-multi-sentence-doctor\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'from transformers import AutoTokenizer, AutoModelWithLMHead\\\\ntokenizer = AutoTokenizer.from_pretrained(flexudy/t5-base-multi-sentence-doctor)\\\\nmodel = AutoModelWithLMHead.from_pretrained(flexudy/t5-base-multi-sentence-doctor)\\\\ninput_text = repair_sentence: m a medical doct context: {That is my job I a}{or I save lives} </s>\\\\ninput_ids = tokenizer.encode(input_text, return_tensors=pt)\\\\noutputs = model.generate(input_ids, max_length=32, num_beams=1)\\\\nsentence = tokenizer.decode(outputs[0], skip_special_tokens=True, clean_up_tokenization_spaces=True)\\\\nassert sentence == I am a medical doctor.\\', \\'performance\\': {\\'dataset\\': \\'tatoeba\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'Sentence doctor is a T5 model that attempts to correct the errors or mistakes found in sentences. Model works on English, German and French text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 66, "text": " Help me fill in the blanks in the following Chinese sentence: \\\"\\u4e0a\\u6d77\\u662f\\u4e2d\\u56fd\\u7684[MASK]\\u5927\\u57ce\\u5e02\\u3002\\\"\\n###Input: \\u4e0a\\u6d77\\u662f\\u4e2d\\u56fd\\u7684[MASK]\\u5927\\u57ce\\u5e02\\u3002\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Fill-Mask\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Fill-Mask\\', \\'api_name\\': \\'uer/albert-base-chinese-cluecorpussmall\\', \\'api_call\\': \"AlbertForMaskedLM.from_pretrained(\\'uer/albert-base-chinese-cluecorpussmall\\')\", \\'api_arguments\\': [\\'model\\', \\'tokenizer\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'from transformers import BertTokenizer, AlbertForMaskedLM, FillMaskPipeline\\\\ntokenizer = BertTokenizer.from_pretrained(uer/albert-base-chinese-cluecorpussmall)\\\\nmodel = AlbertForMaskedLM.from_pretrained(uer/albert-base-chinese-cluecorpussmall)\\\\nunmasker = FillMaskPipeline(model, tokenizer)\\\\nunmasker(\u4e2d\u56fd\u7684\u9996\u90fd\u662f[MASK]\u4eac\u3002)\\', \\'performance\\': {\\'dataset\\': \\'CLUECorpusSmall\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This is the set of Chinese ALBERT models pre-trained by UER-py on the CLUECorpusSmall dataset. The model can be used for tasks like text generation and feature extraction.\\'}', metadata={})]", "category": "generic"}
{"question_id": 67, "text": " We are building a source code autocompletion tool which will complete the code snippet containing a masked token.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Fill-Mask\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Fill-Mask\\', \\'api_name\\': \\'camembert-base\\', \\'api_call\\': \"pipeline(\\'fill-mask\\', model=\\'camembert-base\\', tokenizer=\\'camembert-base\\')\", \\'api_arguments\\': [\\'model\\', \\'tokenizer\\'], \\'python_environment_requirements\\': [\\'transformers\\', \\'torch\\'], \\'example_code\\': \"from transformers import pipeline; camembert_fill_mask = pipeline(\\'fill-mask\\', model=\\'camembert-base\\', tokenizer=\\'camembert-base\\'); results = camembert_fill_mask(\\'Le camembert est <mask> :)\\')\", \\'performance\\': {\\'dataset\\': \\'oscar\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'CamemBERT is a state-of-the-art language model for French based on the RoBERTa model. It is available on Hugging Face in 6 different versions with varying number of parameters, amount of pretraining data, and pretraining data source domains. It can be used for Fill-Mask tasks.\\'}', metadata={})]", "category": "generic"}
{"question_id": 68, "text": " I work for a Japanese company, and my manager needs me to take a look at a request from a client. I can understand fluent Japanese, but I need a little help filling in missing words from the text.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Fill-Mask\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Fill-Mask\\', \\'api_name\\': \\'cl-tohoku/bert-base-japanese\\', \\'api_call\\': \"AutoModelForMaskedLM.from_pretrained(\\'cl-tohoku/bert-base-japanese\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"fill_mask(\\'[MASK]\\')\", \\'performance\\': {\\'dataset\\': \\'wikipedia\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'This is a BERT model pretrained on texts in the Japanese language. This version of the model processes input texts with word-level tokenization based on the IPA dictionary, followed by the WordPiece subword tokenization.\\'}', metadata={})]", "category": "generic"}
{"question_id": 69, "text": " We are building a platform to compare and contrast user input sentences with existing sentences in our database. It should return similar results.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Sentence Similarity\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Sentence Transformers\\', \\'api_name\\': \\'sentence-transformers/distiluse-base-multilingual-cased-v1\\', \\'api_call\\': \"SentenceTransformer(\\'sentence-transformers/distiluse-base-multilingual-cased-v1\\')\", \\'api_arguments\\': [\\'sentences\\'], \\'python_environment_requirements\\': \\'pip install -U sentence-transformers\\', \\'example_code\\': \"from sentence_transformers import SentenceTransformer\\\\nsentences = [This is an example sentence, Each sentence is converted]\\\\nmodel = SentenceTransformer(\\'sentence-transformers/distiluse-base-multilingual-cased-v1\\')\\\\nembeddings = model.encode(sentences)\\\\nprint(embeddings)\", \\'performance\\': {\\'dataset\\': \\'https://seb.sbert.net\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'This is a sentence-transformers model: It maps sentences & paragraphs to a 512 dimensional dense vector space and can be used for tasks like clustering or semantic search.\\'}', metadata={})]", "category": "generic"}
{"question_id": 70, "text": " I need a method to compare the similarity between two sentences to be used within a meme generator, so we can produce a meme with a similar caption.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Sentence Similarity\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Sentence Transformers\\', \\'api_name\\': \\'sentence-transformers/distiluse-base-multilingual-cased-v2\\', \\'api_call\\': \"SentenceTransformer(\\'sentence-transformers/distiluse-base-multilingual-cased-v2\\')\", \\'api_arguments\\': [\\'sentences\\'], \\'python_environment_requirements\\': \\'pip install -U sentence-transformers\\', \\'example_code\\': \"from sentence_transformers import SentenceTransformer\\\\nsentences = [This is an example sentence, Each sentence is converted]\\\\nmodel = SentenceTransformer(\\'sentence-transformers/distiluse-base-multilingual-cased-v2\\')\\\\nembeddings = model.encode(sentences)\\\\nprint(embeddings)\", \\'performance\\': {\\'dataset\\': \\'https://seb.sbert.net\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This is a sentence-transformers model: It maps sentences & paragraphs to a 512 dimensional dense vector space and can be used for tasks like clustering or semantic search.\\'}', metadata={})]", "category": "generic"}
{"question_id": 71, "text": " A student is writing a research paper and needs help with finding similar articles in order to include them in the literature review section.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Sentence Similarity\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Sentence Embeddings\\', \\'api_name\\': \\'sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2\\', \\'api_call\\': \"SentenceTransformer(\\'sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2\\')\", \\'api_arguments\\': [\\'sentences\\'], \\'python_environment_requirements\\': \\'pip install -U sentence-transformers\\', \\'example_code\\': \"from sentence_transformers import SentenceTransformer\\\\nsentences = [This is an example sentence, Each sentence is converted]\\\\nmodel = SentenceTransformer(\\'sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2\\')\\\\nembeddings = model.encode(sentences)\\\\nprint(embeddings)\", \\'performance\\': {\\'dataset\\': \\'https://seb.sbert.net\\', \\'accuracy\\': \\'Automated evaluation\\'}, \\'description\\': \\'This is a sentence-transformers model: It maps sentences & paragraphs to a 384 dimensional dense vector space and can be used for tasks like clustering or semantic search.\\'}', metadata={})]", "category": "generic"}
{"question_id": 72, "text": " Create a solution to convert a given Japanese sentence into a speech audio file.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Speech Recognition\\', \\'api_name\\': \\'jonatasgrosman/wav2vec2-large-xlsr-53-japanese\\', \\'api_call\\': \"SpeechRecognitionModel(\\'jonatasgrosman/wav2vec2-large-xlsr-53-japanese\\')\", \\'api_arguments\\': [\\'audio_paths\\'], \\'python_environment_requirements\\': [\\'huggingsound\\', \\'torch\\', \\'librosa\\', \\'datasets\\', \\'transformers\\'], \\'example_code\\': \\'from huggingsound import SpeechRecognitionModel\\\\nmodel = SpeechRecognitionModel(jonatasgrosman/wav2vec2-large-xlsr-53-japanese)\\\\naudio_paths = [/path/to/file.mp3, /path/to/another_file.wav]\\\\ntranscriptions = model.transcribe(audio_paths)\\', \\'performance\\': {\\'dataset\\': \\'common_voice\\', \\'accuracy\\': {\\'WER\\': 81.8, \\'CER\\': 20.16}}, \\'description\\': \\'Fine-tuned XLSR-53 large model for speech recognition in Japanese. Trained on Common Voice 6.1, CSS10, and JSUT datasets. Make sure your speech input is sampled at 16kHz.\\'}', metadata={})]", "category": "generic"}
{"question_id": 73, "text": " We are working on a transcription service for our customers. We need a way to convert audio files into text.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'CTranslate2\\', \\'functionality\\': \\'Automatic Speech Recognition\\', \\'api_name\\': \\'guillaumekln/faster-whisper-large-v2\\', \\'api_call\\': \"WhisperModel(\\'large-v2\\')\", \\'api_arguments\\': [\\'audio.mp3\\'], \\'python_environment_requirements\\': [\\'faster_whisper\\'], \\'example_code\\': \\'from faster_whisper import WhisperModel\\\\nmodel = WhisperModel(large-v2)\\\\nsegments, info = model.transcribe(audio.mp3)\\\\nfor segment in segments:\\\\n print([%.2fs -&gt; %.2fs] %s % (segment.start, segment.end, segment.text))\\', \\'performance\\': {\\'dataset\\': \\'99 languages\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Whisper large-v2 model for CTranslate2. This model can be used in CTranslate2 or projets based on CTranslate2 such as faster-whisper.\\'}', metadata={})]", "category": "generic"}
{"question_id": 74, "text": " We are creating an online video conference service, and we need to detect when two or more speakers are speaking at the same time in the audio.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'pyannote.audio\\', \\'functionality\\': \\'Speaker diarization\\', \\'api_name\\': \\'johnislarry/cloned-pyannote-speaker-diarization-endpoint\\', \\'api_call\\': \"Pipeline.from_pretrained(\\'pyannote/speaker-diarization@2.1\\',use_auth_token=\\'ACCESS_TOKEN_GOES_HERE\\')\", \\'api_arguments\\': [\\'num_speakers\\', \\'min_speakers\\', \\'max_speakers\\', \\'segmentation_onset\\'], \\'python_environment_requirements\\': \\'pyannote.audio 2.0\\', \\'example_code\\': {\\'load_pipeline\\': \\'from pyannote.audio import Pipeline\\\\npipeline = Pipeline.from_pretrained(pyannote/speaker-diarization@2022.07)\\', \\'apply_pipeline\\': \\'diarization = pipeline(audio.wav)\\', \\'save_output\\': \\'with open(audio.rttm, w) as rttm:\\\\n  diarization.write_rttm(rttm)\\'}, \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'AISHELL-4\\', \\'accuracy\\': {\\'DER%\\': 14.61, \\'FA%\\': 3.31, \\'Miss%\\': 4.35, \\'Conf%\\': 6.95}}, {\\'name\\': \\'AMI Mix-Headset only_words\\', \\'accuracy\\': {\\'DER%\\': 18.21, \\'FA%\\': 3.28, \\'Miss%\\': 11.07, \\'Conf%\\': 3.87}}, {\\'name\\': \\'AMI Array1-01 only_words\\', \\'accuracy\\': {\\'DER%\\': 29.0, \\'FA%\\': 2.71, \\'Miss%\\': 21.61, \\'Conf%\\': 4.68}}, {\\'name\\': \\'CALLHOME Part2\\', \\'accuracy\\': {\\'DER%\\': 30.24, \\'FA%\\': 3.71, \\'Miss%\\': 16.86, \\'Conf%\\': 9.66}}, {\\'name\\': \\'DIHARD 3 Full\\', \\'accuracy\\': {\\'DER%\\': 20.99, \\'FA%\\': 4.25, \\'Miss%\\': 10.74, \\'Conf%\\': 6.0}}, {\\'name\\': \\'REPERE Phase 2\\', \\'accuracy\\': {\\'DER%\\': 12.62, \\'FA%\\': 1.55, \\'Miss%\\': 3.3, \\'Conf%\\': 7.76}}, {\\'name\\': \\'VoxConverse v0.0.2\\', \\'accuracy\\': {\\'DER%\\': 12.76, \\'FA%\\': 3.45, \\'Miss%\\': 3.85, \\'Conf%\\': 5.46}}]}, \\'description\\': \\'This API provides speaker diarization functionality using the pyannote.audio framework. It is capable of processing audio files and outputting speaker diarization results in RTTM format. The API supports providing the number of speakers, minimum and maximum number of speakers, and adjusting the segmentation onset threshold.\\'}', metadata={})]", "category": "generic"}
{"question_id": 75, "text": " Our company develops smart speaker devices that involve interaction with the user. We need to transcribe the input from the users with the maintained accent or language.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'SYSPIN/Telugu_Male_TTS\\', \\'api_call\\': \"pipeline(\\'text-to-speech\\', model=\\'SYSPIN/Telugu_Male_TTS\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Telugu Male Text-to-Speech model using the ESPnet framework, provided by Hugging Face.\\'}', metadata={})]", "category": "generic"}
{"question_id": 76, "text": " One of our clients is facing noise issues on their audio recordings. Can you help them to remove the noise from the audio?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Voice Activity Detection, Speech-to-Noise Ratio, and C50 Room Acoustics Estimation\\', \\'api_name\\': \\'pyannote/brouhaha\\', \\'api_call\\': \"Model.from_pretrained(\\'pyannote/brouhaha\\', use_auth_token=\\'ACCESS_TOKEN_GOES_HERE\\')\", \\'api_arguments\\': [\\'audio.wav\\'], \\'python_environment_requirements\\': [\\'pyannote-audio\\', \\'brouhaha-vad\\'], \\'example_code\\': [\\'from pyannote.audio import Model\\', \\'model = Model.from_pretrained(pyannote/brouhaha, use_auth_token=ACCESS_TOKEN_GOES_HERE)\\', \\'from pyannote.audio import Inference\\', \\'inference = Inference(model)\\', \\'output = inference(audio.wav)\\', \\'for frame, (vad, snr, c50) in output:\\', \\'  t = frame.middle\\', \\'  print(f{t:8.3f} vad={100*vad:.0f}% snr={snr:.0f} c50={c50:.0f})\\'], \\'performance\\': {\\'dataset\\': \\'LibriSpeech, AudioSet, EchoThief, MIT-Acoustical-Reverberation-Scene\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Brouhaha is a joint voice activity detection, speech-to-noise ratio, and C50 room acoustics estimation model. It is based on the PyTorch framework and uses the pyannote.audio library.\\'}', metadata={})]", "category": "generic"}
{"question_id": 77, "text": " We are a media company and we have a large volume of Chinese language audio files. We want to transcribe the audios into chinese text.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'CTranslate2\\', \\'functionality\\': \\'Automatic Speech Recognition\\', \\'api_name\\': \\'guillaumekln/faster-whisper-large-v2\\', \\'api_call\\': \"WhisperModel(\\'large-v2\\')\", \\'api_arguments\\': [\\'audio.mp3\\'], \\'python_environment_requirements\\': [\\'faster_whisper\\'], \\'example_code\\': \\'from faster_whisper import WhisperModel\\\\nmodel = WhisperModel(large-v2)\\\\nsegments, info = model.transcribe(audio.mp3)\\\\nfor segment in segments:\\\\n print([%.2fs -&gt; %.2fs] %s % (segment.start, segment.end, segment.text))\\', \\'performance\\': {\\'dataset\\': \\'99 languages\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Whisper large-v2 model for CTranslate2. This model can be used in CTranslate2 or projets based on CTranslate2 such as faster-whisper.\\'}', metadata={})]", "category": "generic"}
{"question_id": 78, "text": " Help us improve the listener experience from our customers by enhancing the audio of noisy recordings.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Voice Activity Detection, Speech-to-Noise Ratio, and C50 Room Acoustics Estimation\\', \\'api_name\\': \\'pyannote/brouhaha\\', \\'api_call\\': \"Model.from_pretrained(\\'pyannote/brouhaha\\', use_auth_token=\\'ACCESS_TOKEN_GOES_HERE\\')\", \\'api_arguments\\': [\\'audio.wav\\'], \\'python_environment_requirements\\': [\\'pyannote-audio\\', \\'brouhaha-vad\\'], \\'example_code\\': [\\'from pyannote.audio import Model\\', \\'model = Model.from_pretrained(pyannote/brouhaha, use_auth_token=ACCESS_TOKEN_GOES_HERE)\\', \\'from pyannote.audio import Inference\\', \\'inference = Inference(model)\\', \\'output = inference(audio.wav)\\', \\'for frame, (vad, snr, c50) in output:\\', \\'  t = frame.middle\\', \\'  print(f{t:8.3f} vad={100*vad:.0f}% snr={snr:.0f} c50={c50:.0f})\\'], \\'performance\\': {\\'dataset\\': \\'LibriSpeech, AudioSet, EchoThief, MIT-Acoustical-Reverberation-Scene\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Brouhaha is a joint voice activity detection, speech-to-noise ratio, and C50 room acoustics estimation model. It is based on the PyTorch framework and uses the pyannote.audio library.\\'}', metadata={})]", "category": "generic"}
{"question_id": 79, "text": " Our company is working on a project to automatically translate spoken English audio to spoken Hokkien audio. We need a speech-to-speech translation model.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'Fairseq\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'unit_hifigan_HK_layer12.km2500_frame_TAT-TTS\\', \\'api_call\\': \"load_model_ensemble_and_task_from_hf_hub(\\'facebook/unit_hifigan_HK_layer12.km2500_frame_TAT-TT\\')\", \\'api_arguments\\': {\\'unit\\': \\'Text input for the TTS model\\'}, \\'python_environment_requirements\\': [\\'fairseq\\', \\'huggingface_hub\\', \\'torchaudio\\'], \\'example_code\\': \"import json\\\\nimport os\\\\nfrom pathlib import Path\\\\nimport IPython.display as ipd\\\\nfrom fairseq import hub_utils\\\\nfrom fairseq.checkpoint_utils import load_model_ensemble_and_task_from_hf_hub\\\\nfrom fairseq.models.speech_to_text.hub_interface import S2THubInterface\\\\nfrom fairseq.models.text_to_speech import CodeHiFiGANVocoder\\\\nfrom fairseq.models.text_to_speech.hub_interface import VocoderHubInterface\\\\nfrom huggingface_hub import snapshot_download\\\\nimport torchaudio\\\\ncache_dir = os.getenv(HUGGINGFACE_HUB_CACHE)\\\\nlibrary_name = fairseq\\\\ncache_dir = (\\\\n cache_dir or (Path.home() / .cache / library_name).as_posix()\\\\n)\\\\ncache_dir = snapshot_download(\\\\n ffacebook/unit_hifigan_HK_layer12.km2500_frame_TAT-TTS, cache_dir=cache_dir, library_name=library_name\\\\n)\\\\nx = hub_utils.from_pretrained(\\\\n cache_dir,\\\\n model.pt,\\\\n .,\\\\n archive_map=CodeHiFiGANVocoder.hub_models(),\\\\n config_yaml=config.json,\\\\n fp16=False,\\\\n is_vocoder=True,\\\\n)\\\\nwith open(f{x[\\'args\\'][\\'data\\']}/config.json) as f:\\\\n vocoder_cfg = json.load(f)\\\\nassert (\\\\n len(x[args][model_path]) == 1\\\\n), Too many vocoder models in the input\\\\nvocoder = CodeHiFiGANVocoder(x[args][model_path][0], vocoder_cfg)\\\\ntts_model = VocoderHubInterface(vocoder_cfg, vocoder)\\\\ntts_sample = tts_model.get_model_input(unit)\\\\nwav, sr = tts_model.get_prediction(tts_sample)\\\\nipd.Audio(wav, rate=sr)\", \\'performance\\': {\\'dataset\\': \\'TAT-TTS\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Hokkien unit HiFiGAN based vocoder from fairseq. Trained with TAT-TTS data with 4 speakers in Taiwanese Hokkien accent.\\'}', metadata={})]", "category": "generic"}
{"question_id": 80, "text": " We are a startup developing voice assistants. We need a keyword spotting system that can recognize user commands.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Voice Activity Detection\\', \\'api_name\\': \\'d4data/Indian-voice-cloning\\', \\'api_call\\': \"pipeline(\\'voice-activity-detection\\', model=\\'d4data/Indian-voice-cloning\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A model for detecting voice activity in Indian languages.\\'}', metadata={})]", "category": "generic"}
{"question_id": 81, "text": " The model needs to have speech recognition capability to identify languages in a given audio file.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Speech Recognition\\', \\'api_name\\': \\'jonatasgrosman/wav2vec2-large-xlsr-53-portuguese\\', \\'api_call\\': \"SpeechRecognitionModel(\\'jonatasgrosman/wav2vec2-large-xlsr-53-portuguese\\')\", \\'api_arguments\\': [\\'audio_paths\\'], \\'python_environment_requirements\\': [\\'huggingsound\\', \\'torch\\', \\'librosa\\', \\'datasets\\', \\'transformers\\'], \\'example_code\\': \\'from huggingsound import SpeechRecognitionModel\\\\nmodel = SpeechRecognitionModel(jonatasgrosman/wav2vec2-large-xlsr-53-portuguese)\\\\naudio_paths = [/path/to/file.mp3, /path/to/another_file.wav]\\\\ntranscriptions = model.transcribe(audio_paths)\\', \\'performance\\': {\\'dataset\\': \\'mozilla-foundation/common_voice_6_0\\', \\'accuracy\\': {\\'Test WER\\': 11.31, \\'Test CER\\': 3.74, \\'Test WER (+LM)\\': 9.01, \\'Test CER (+LM)\\': 3.21}}, \\'description\\': \\'Fine-tuned facebook/wav2vec2-large-xlsr-53 on Portuguese using the train and validation splits of Common Voice 6.1. When using this model, make sure that your speech input is sampled at 16kHz.\\'}', metadata={})]", "category": "generic"}
{"question_id": 82, "text": " I have just recorded a meeting, I want to find the best segments from the audio where people are speaking, and construct a summary.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'pyannote.audio\\', \\'functionality\\': \\'Speaker diarization\\', \\'api_name\\': \\'johnislarry/cloned-pyannote-speaker-diarization-endpoint\\', \\'api_call\\': \"Pipeline.from_pretrained(\\'pyannote/speaker-diarization@2.1\\',use_auth_token=\\'ACCESS_TOKEN_GOES_HERE\\')\", \\'api_arguments\\': [\\'num_speakers\\', \\'min_speakers\\', \\'max_speakers\\', \\'segmentation_onset\\'], \\'python_environment_requirements\\': \\'pyannote.audio 2.0\\', \\'example_code\\': {\\'load_pipeline\\': \\'from pyannote.audio import Pipeline\\\\npipeline = Pipeline.from_pretrained(pyannote/speaker-diarization@2022.07)\\', \\'apply_pipeline\\': \\'diarization = pipeline(audio.wav)\\', \\'save_output\\': \\'with open(audio.rttm, w) as rttm:\\\\n  diarization.write_rttm(rttm)\\'}, \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'AISHELL-4\\', \\'accuracy\\': {\\'DER%\\': 14.61, \\'FA%\\': 3.31, \\'Miss%\\': 4.35, \\'Conf%\\': 6.95}}, {\\'name\\': \\'AMI Mix-Headset only_words\\', \\'accuracy\\': {\\'DER%\\': 18.21, \\'FA%\\': 3.28, \\'Miss%\\': 11.07, \\'Conf%\\': 3.87}}, {\\'name\\': \\'AMI Array1-01 only_words\\', \\'accuracy\\': {\\'DER%\\': 29.0, \\'FA%\\': 2.71, \\'Miss%\\': 21.61, \\'Conf%\\': 4.68}}, {\\'name\\': \\'CALLHOME Part2\\', \\'accuracy\\': {\\'DER%\\': 30.24, \\'FA%\\': 3.71, \\'Miss%\\': 16.86, \\'Conf%\\': 9.66}}, {\\'name\\': \\'DIHARD 3 Full\\', \\'accuracy\\': {\\'DER%\\': 20.99, \\'FA%\\': 4.25, \\'Miss%\\': 10.74, \\'Conf%\\': 6.0}}, {\\'name\\': \\'REPERE Phase 2\\', \\'accuracy\\': {\\'DER%\\': 12.62, \\'FA%\\': 1.55, \\'Miss%\\': 3.3, \\'Conf%\\': 7.76}}, {\\'name\\': \\'VoxConverse v0.0.2\\', \\'accuracy\\': {\\'DER%\\': 12.76, \\'FA%\\': 3.45, \\'Miss%\\': 3.85, \\'Conf%\\': 5.46}}]}, \\'description\\': \\'This API provides speaker diarization functionality using the pyannote.audio framework. It is capable of processing audio files and outputting speaker diarization results in RTTM format. The API supports providing the number of speakers, minimum and maximum number of speakers, and adjusting the segmentation onset threshold.\\'}', metadata={})]", "category": "generic"}
{"question_id": 83, "text": " I am running a wine store, and I am looking for a machine learning model that can help me classify the quality of wine based on some given features.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Scikit-learn\\', \\'functionality\\': \\'Wine Quality classification\\', \\'api_name\\': \\'julien-c/wine-quality\\', \\'api_call\\': \"joblib.load(cached_download(hf_hub_url(\\'julien-c/wine-quality\\', \\'winequality-red.csv\\')))\", \\'api_arguments\\': [\\'X\\'], \\'python_environment_requirements\\': [\\'huggingface_hub\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \\'from huggingface_hub import hf_hub_url, cached_download\\\\nimport joblib\\\\nimport pandas as pd\\\\nREPO_ID = julien-c/wine-quality\\\\nFILENAME = sklearn_model.joblib\\\\nmodel = joblib.load(cached_download(\\\\n hf_hub_url(REPO_ID, FILENAME)\\\\n))\\\\ndata_file = cached_download(\\\\n hf_hub_url(REPO_ID, winequality-red.csv)\\\\n)\\\\nwinedf = pd.read_csv(data_file, sep=;)\\\\nX = winedf.drop([quality], axis=1)\\\\nY = winedf[quality]\\\\nprint(X[:3])\\\\nlabels = model.predict(X[:3])\\\\nmodel.score(X, Y)\\', \\'performance\\': {\\'dataset\\': \\'julien-c/wine-quality\\', \\'accuracy\\': 0.6616635397123202}, \\'description\\': \\'A Simple Example of Scikit-learn Pipeline for Wine Quality classification. Inspired by https://towardsdatascience.com/a-simple-example-of-pipeline-in-machine-learning-with-scikit-learn-e726ffbb6976 by Saptashwa Bhattacharyya.\\'}', metadata={})]", "category": "generic"}
{"question_id": 84, "text": " Build a simple application to predict the survival status of passengers on the Titanic based on their age, gender, and passenger class.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Scikit-learn\\', \\'functionality\\': \\'Binary Classification\\', \\'api_name\\': \\'danupurnomo/dummy-titanic\\', \\'api_call\\': \"load_model(cached_download(hf_hub_url(\\'danupurnomo/dummy-titanic\\', \\'titanic_model.h5\\')))\", \\'api_arguments\\': [\\'new_data\\'], \\'python_environment_requirements\\': [\\'huggingface_hub\\', \\'joblib\\', \\'pandas\\', \\'numpy\\', \\'tensorflow\\'], \\'example_code\\': \"from huggingface_hub import hf_hub_url, cached_download\\\\nimport joblib\\\\nimport pandas as pd\\\\nimport numpy as np\\\\nfrom tensorflow.keras.models import load_model\\\\nREPO_ID = \\'danupurnomo/dummy-titanic\\'\\\\nPIPELINE_FILENAME = \\'final_pipeline.pkl\\'\\\\nTF_FILENAME = \\'titanic_model.h5\\'\\\\nmodel_pipeline = joblib.load(cached_download(\\\\n hf_hub_url(REPO_ID, PIPELINE_FILENAME)\\\\n))\\\\nmodel_seq = load_model(cached_download(\\\\n hf_hub_url(REPO_ID, TF_FILENAME)\\\\n))\", \\'performance\\': {\\'dataset\\': \\'Titanic\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This model is a binary classifier for predicting whether a passenger on the Titanic survived or not, based on features such as passenger class, age, sex, fare, and more.\\'}', metadata={})]", "category": "generic"}
{"question_id": 85, "text": " I need to estimate CO2 emissions from vehicles based on their characteristics, such as engine size, transmission type, and miles traveled.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'kochetkovIT/autotrain-ironhack-49741119788\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'json\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'kochetkovIT/autotrain-data-ironhack\\', \\'accuracy\\': {\\'Loss\\': 2.603, \\'R2\\': 0.013, \\'MSE\\': 6.776, \\'MAE\\': 1.666, \\'RMSLE\\': 0.502}}, \\'description\\': \\'A tabular regression model trained using AutoTrain to predict carbon emissions (in grams) with an R2 score of 0.013.\\'}', metadata={})]", "category": "generic"}
{"question_id": 86, "text": " We have been asked to predict future criminal re-offense from a given dataset. What model should we adopt and how do we proceed?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Scikit-learn\\', \\'functionality\\': \\'Classification\\', \\'api_name\\': \\'imodels/figs-compas-recidivism\\', \\'api_call\\': \"joblib.load(cached_download(hf_hub_url(\\'imodels/figs-compas-recidivism\\', \\'sklearn_model.joblib\\')))\", \\'api_arguments\\': [\\'REPO_ID\\', \\'FILENAME\\'], \\'python_environment_requirements\\': [\\'joblib\\', \\'huggingface_hub\\', \\'pandas\\', \\'numpy\\', \\'datasets\\', \\'imodels\\', \\'sklearn.model_selection\\'], \\'example_code\\': \"from huggingface_hub import hf_hub_url, cached_download\\\\nimport joblib\\\\nimport pandas as pd\\\\nREPO_ID = imodels/figs-compas-recidivism\\\\nFILENAME = sklearn_model.joblib\\\\nmodel = joblib.load(cached_download(\\\\n hf_hub_url(REPO_ID, FILENAME)\\\\n))\\\\npreds = model.predict(X_test)\\\\nprint(\\'accuracy\\', np.mean(preds==y_test))\", \\'performance\\': {\\'dataset\\': \\'imodels/compas-recidivism\\', \\'accuracy\\': 0.6759165485112416}, \\'description\\': \\'A tabular classification model for predicting recidivism using the COMPAS dataset. The model is an imodels.FIGSClassifier trained with Scikit-learn and can be used with the Hugging Face Inference API.\\'}', metadata={})]", "category": "generic"}
{"question_id": 87, "text": " Our company's goal is to predict carbon emissions based on the given features of the compound.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'kochetkovIT/autotrain-ironhack-49741119788\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'json\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'kochetkovIT/autotrain-data-ironhack\\', \\'accuracy\\': {\\'Loss\\': 2.603, \\'R2\\': 0.013, \\'MSE\\': 6.776, \\'MAE\\': 1.666, \\'RMSLE\\': 0.502}}, \\'description\\': \\'A tabular regression model trained using AutoTrain to predict carbon emissions (in grams) with an R2 score of 0.013.\\'}', metadata={})]", "category": "generic"}
{"question_id": 88, "text": " The factory wants to make its production process more eco-friendly. Calculate the carbon emissions for given data.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'45473113800\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'samvelkoch/autotrain-data-prknsn-2\\', \\'accuracy\\': {\\'Loss\\': 5.079, \\'R2\\': 0.109, \\'MSE\\': 25.795, \\'MAE\\': 3.78, \\'RMSLE\\': 0.849}}, \\'description\\': \\'A tabular regression model trained with AutoTrain for predicting carbon emissions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 89, "text": " We want to predict the carbon emissions of a new line of electric vehicles for an annual report. Automate the process of loading a regression model, then calculate the forecast of emissions for this year.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'kochetkovIT/autotrain-ironhack-49741119788\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'json\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'kochetkovIT/autotrain-data-ironhack\\', \\'accuracy\\': {\\'Loss\\': 2.603, \\'R2\\': 0.013, \\'MSE\\': 6.776, \\'MAE\\': 1.666, \\'RMSLE\\': 0.502}}, \\'description\\': \\'A tabular regression model trained using AutoTrain to predict carbon emissions (in grams) with an R2 score of 0.013.\\'}', metadata={})]", "category": "generic"}
{"question_id": 90, "text": " We are planning to launch a website which provides tips to people for their daily lives. Can you please build a model to predict the appropriate amount of tips?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Scikit-learn\\', \\'functionality\\': \\'baseline-trainer\\', \\'api_name\\': \\'merve/tips9y0jvt5q-tip-regression\\', \\'api_call\\': \"pipeline(\\'tabular-regression\\', model=\\'merve/tips9y0jvt5q-tip-regression\\')\", \\'api_arguments\\': \\'N/A\\', \\'python_environment_requirements\\': \\'dabl\\', \\'example_code\\': \\'N/A\\', \\'performance\\': {\\'dataset\\': \\'tips9y0jvt5q\\', \\'accuracy\\': {\\'r2\\': 0.41524, \\'neg_mean_squared_error\\': -1.098792}}, \\'description\\': \\'Baseline Model trained on tips9y0jvt5q to apply regression on tip. The model uses Ridge(alpha=10) and is trained with dabl library as a baseline. For better results, use AutoTrain.\\'}', metadata={})]", "category": "generic"}
{"question_id": 91, "text": " We have a robotic arm in our warehouse that needs to be trained to optimize loading and unloading tasks. The robotic arm is based on the CartPole environment.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Reinforcement Learning\\', \\'framework\\': \\'Stable-Baselines3\\', \\'functionality\\': \\'CartPole-v1\\', \\'api_name\\': \\'dqn-CartPole-v1\\', \\'api_call\\': \"load_from_hub(repo_id=\\'sb3/dqn-CartPole-v1\\',filename=\\'{MODEL FILENAME}.zip\\',)\", \\'api_arguments\\': [\\'algo\\', \\'env\\', \\'logs\\'], \\'python_environment_requirements\\': [\\'rl_zoo3\\', \\'stable-baselines3\\', \\'stable-baselines3-contrib\\'], \\'example_code\\': \\'python train.py --algo dqn --env CartPole-v1 -f logs/\\', \\'performance\\': {\\'dataset\\': \\'CartPole-v1\\', \\'accuracy\\': \\'500.00 +/- 0.00\\'}, \\'description\\': \\'This is a trained model of a DQN agent playing CartPole-v1 using the stable-baselines3 library and the RL Zoo. The RL Zoo is a training framework for Stable Baselines3 reinforcement learning agents, with hyperparameter optimization and pre-trained agents included.\\'}', metadata={})]", "category": "generic"}
{"question_id": 92, "text": " There is an upcoming event called \\\"Space Party\\\" and we need a representative image for the event. Can you assist us in creating an image containing a party in space with astronauts and aliens having fun together?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'prompthero/openjourney-v4\\', \\'api_call\\': \"pipeline(\\'text-to-image\\', model=\\'prompthero/openjourney-v4\\')\", \\'api_arguments\\': {\\'text\\': \\'string\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"generate_image(\\'your text here\\')\", \\'performance\\': {\\'dataset\\': \\'Midjourney v4 images\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Openjourney v4 is trained on +124k Midjourney v4 images by PromptHero. It is used for generating images based on text inputs.\\'}', metadata={})]", "category": "generic"}
{"question_id": 93, "text": " We're creating a promotional image for a wildlife-themed event. We need to display two tigers in a natural setting.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'text2text-generation\\', \\'api_name\\': \\'tuner007/pegasus_summarizer\\', \\'api_call\\': \"PegasusForConditionalGeneration.from_pretrained(\\'tuner007/pegasus_summarizer\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'pip install sentencepiece\\'], \\'example_code\\': \"context = \\\\nIndia wicket-keeper batsman Rishabh Pant has said someone from the crowd threw a ball on pacer Mohammed Siraj while he was fielding in the ongoing third Test against England on Wednesday. Pant revealed the incident made India skipper Virat Kohli upset. I think, somebody threw a ball inside, at Siraj, so he [Kohli] was upset, said Pant in a virtual press conference after the close of the first day\\'s play.You can say whatever you want to chant, but don\\'t throw things at the fielders and all those things. It is not good for cricket, I guess, he added.In the third session of the opening day of the third Test, a section of spectators seemed to have asked Siraj the score of the match to tease the pacer. The India pacer however came with a brilliant reply as he gestured 1-0 (India leading the Test series) towards the crowd.Earlier this month, during the second Test match, there was some bad crowd behaviour on a show as some unruly fans threw champagne corks at India batsman KL Rahul.Kohli also intervened and he was seen gesturing towards the opening batsman to know more about the incident. An over later, the TV visuals showed that many champagne corks were thrown inside the playing field, and the Indian players were visibly left frustrated.Coming back to the game, after bundling out India for 78, openers Rory Burns and Haseeb Hameed ensured that England took the honours on the opening day of the ongoing third Test.At stumps, England\\'s score reads 120/0 and the hosts have extended their lead to 42 runs. For the Three Lions, Burns (52) and Hameed (60) are currently unbeaten at the crease.Talking about the pitch on opening day, Pant said, They took the heavy roller, the wicket was much more settled down, and they batted nicely also, he said. But when we batted, the wicket was slightly soft, and they bowled in good areas, but we could have applied [ourselves] much better.Both England batsmen managed to see off the final session and the hosts concluded the opening day with all ten wickets intact, extending the lead to 42.(ANI)\\\\n\\\\nget_response(context)\", \\'performance\\': {\\'dataset\\': \\'cnn_dailymail\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.604, \\'ROUGE-2\\': 14.64, \\'ROUGE-L\\': 23.884, \\'ROUGE-LSUM\\': 32.902, \\'loss\\': 2.576, \\'gen_len\\': 76.398}}, \\'description\\': \\'PEGASUS fine-tuned for summarization\\'}', metadata={})]", "category": "generic"}
{"question_id": 94, "text": " We have a collection of low-resolution images of movie characters, and we need to upscale those images to get a more detailed high-resolution image.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Image Upscaling\\', \\'api_name\\': \\'stabilityai/sd-x2-latent-upscaler\\', \\'api_call\\': \\'StableDiffusionLatentUpscalePipeline.from_pretrained(stabilityai/sd-x2-latent-upscaler, torch_dtype=torch.float16)\\', \\'api_arguments\\': {\\'prompt\\': \\'text prompt\\', \\'image\\': \\'low resolution latents\\', \\'num_inference_steps\\': 20, \\'guidance_scale\\': 0, \\'generator\\': \\'torch generator\\'}, \\'python_environment_requirements\\': [\\'git+https://github.com/huggingface/diffusers.git\\', \\'transformers\\', \\'accelerate\\', \\'scipy\\', \\'safetensors\\'], \\'example_code\\': \\'from diffusers import StableDiffusionLatentUpscalePipeline, StableDiffusionPipeline\\\\nimport torch\\\\npipeline = StableDiffusionPipeline.from_pretrained(CompVis/stable-diffusion-v1-4, torch_dtype=torch.float16)\\\\npipeline.to(cuda)\\\\nupscaler = StableDiffusionLatentUpscalePipeline.from_pretrained(stabilityai/sd-x2-latent-upscaler, torch_dtype=torch.float16)\\\\nupscaler.to(cuda)\\\\nprompt = a photo of an astronaut high resolution, unreal engine, ultra realistic\\\\ngenerator = torch.manual_seed(33)\\\\nlow_res_latents = pipeline(prompt, generator=generator, output_type=latent).images\\\\nupscaled_image = upscaler(prompt=prompt, image=low_res_latents, num_inference_steps=20, guidance_scale=0, generator=generator).images[0]\\\\nupscaled_image.save(astronaut_1024.png)\\', \\'performance\\': {\\'dataset\\': \\'LAION-2B\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \"Stable Diffusion x2 latent upscaler is a diffusion-based upscaler model developed by Katherine Crowson in collaboration with Stability AI. It is designed to upscale Stable Diffusion\\'s latent denoised image embeddings, allowing for fast text-to-image and upscaling pipelines. The model was trained on a high-resolution subset of the LAION-2B dataset and works with all Stable Diffusion checkpoints.\"}', metadata={})]", "category": "generic"}
{"question_id": 95, "text": " I want you to create a function that generates captions for a list of images.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Image-to-Text\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'promptcap-coco-vqa\\', \\'api_call\\': \"PromptCap(\\'vqascore/promptcap-coco-vqa\\')\", \\'api_arguments\\': {\\'prompt\\': \\'string\\', \\'image\\': \\'string\\'}, \\'python_environment_requirements\\': \\'pip install promptcap\\', \\'example_code\\': [\\'import torch\\', \\'from promptcap import PromptCap\\', \\'model = PromptCap(vqascore/promptcap-coco-vqa)\\', \\'if torch.cuda.is_available():\\', \\'  model.cuda()\\', \\'prompt = please describe this image according to the given question: what piece of clothing is this boy putting on?\\', \\'image = glove_boy.jpeg\\', \\'print(model.caption(prompt, image))\\'], \\'performance\\': {\\'dataset\\': {\\'coco\\': {\\'accuracy\\': \\'150 CIDEr\\'}, \\'OK-VQA\\': {\\'accuracy\\': \\'60.4%\\'}, \\'A-OKVQA\\': {\\'accuracy\\': \\'59.6%\\'}}}, \\'description\\': \\'PromptCap is a captioning model that can be controlled by natural language instruction. The instruction may contain a question that the user is interested in. It achieves SOTA performance on COCO captioning (150 CIDEr) and knowledge-based VQA tasks when paired with GPT-3 (60.4% on OK-VQA and 59.6% on A-OKVQA).\\'}', metadata={})]", "category": "generic"}
{"question_id": 96, "text": " We need a tool to help us generate textual descriptions for images and videos related to our product.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'camenduru/text2-video-zero\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'camenduru/text2-video-zero\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'This model is used for generating videos from text inputs. It is based on the Hugging Face framework and can be used with the transformers library. The model is trained on a variety of text and video datasets, and can be used for tasks such as video summarization, video generation from text prompts, and more.\\'}', metadata={})]", "category": "generic"}
{"question_id": 97, "text": " We need to build an AI-powered tool to assist visually impaired users in understanding their surroundings by answering questions about images.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Visual Question Answering\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Visual Question Answering\\', \\'api_name\\': \\'temp_vilt_vqa\\', \\'api_call\\': \"pipeline(\\'visual-question-answering\\', model=\\'Bingsu/temp_vilt_vqa\\', tokenizer=\\'Bingsu/temp_vilt_vqa\\')\", \\'api_arguments\\': {\\'model\\': \\'Bingsu/temp_vilt_vqa\\', \\'tokenizer\\': \\'Bingsu/temp_vilt_vqa\\'}, \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A visual question answering model for answering questions related to images using the Hugging Face Transformers library.\\'}', metadata={})]", "category": "generic"}
{"question_id": 98, "text": " I'm a nutritionist and want to help my clients by answering questions about their meals. They will send me an image of their food and ask me a question about it, like \\\"Is this vegan?\\\" or \\\"How many calories do you think it contains?\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'vision-encoder-decoder\\', \\'api_name\\': \\'naver-clova-ix/donut-base-finetuned-docvqa\\', \\'api_call\\': \"pipeline(\\'document-question-answering\\', model=\\'donut-base-finetuned-docvqa\\')\", \\'api_arguments\\': {\\'image\\': \\'path_to_image\\', \\'question\\': \\'your_question\\'}, \\'python_environment_requirements\\': \\'Transformers\\', \\'example_code\\': \"from transformers import pipeline\\\\n\\\\n# Initialize the pipeline\\\\ndoc_qa = pipeline(\\'document-question-answering\\', model=\\'naver-clova-ix/donut-base-finetuned-docvqa\\')\\\\n\\\\n# Load an image and ask a question\\\\nimage_path = \\'path_to_image\\'\\\\nquestion = \\'your_question\\'\\\\n\\\\n# Get the answer\\\\nanswer = doc_qa({\\'image\\': image_path, \\'question\\': question})\\\\nprint(answer)\", \\'performance\\': {\\'dataset\\': \\'DocVQA\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Donut model fine-tuned on DocVQA. It was introduced in the paper OCR-free Document Understanding Transformer by Geewok et al. and first released in this repository. Donut consists of a vision encoder (Swin Transformer) and a text decoder (BART). Given an image, the encoder first encodes the image into a tensor of embeddings (of shape batch_size, seq_len, hidden_size), after which the decoder autoregressively generates text, conditioned on the encoding of the encoder.\\'}', metadata={})]", "category": "generic"}
{"question_id": 99, "text": " Our client is a legal firm that needs assistance in extracting specific information from a large number of legal documents. Automate the process of answering questions related to these documents.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 100, "text": " In a healthcare company, we are trying to create an automated system for answering patient-related questions based on their medical documents. We need a solution using NLP.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'deepset/roberta-base-squad2-covid\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=RobertaForQuestionAnswering.from_pretrained(\\'deepset/roberta-base-squad2-covid\\'), tokenizer=RobertaTokenizer.from_pretrained(\\'deepset/roberta-base-squad2-covid\\'))\", \\'api_arguments\\': {\\'model_name\\': \\'deepset/roberta-base-squad2-covid\\', \\'tokenizer\\': \\'deepset/roberta-base-squad2-covid\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': {\\'QA_input\\': {\\'question\\': \\'Why is model conversion important?\\', \\'context\\': \\'The option to convert models between FARM and transformers gives freedom to the user and let people easily switch between frameworks.\\'}, \\'res\\': \\'nlp(QA_input)\\'}, \\'performance\\': {\\'dataset\\': \\'squad_v2\\', \\'accuracy\\': {\\'XVAL_EM\\': 0.17890995260663506, \\'XVAL_f1\\': 0.49925444207319924, \\'XVAL_top_3_recall\\': 0.8021327014218009}}, \\'description\\': \\'This model is a Roberta-based model fine-tuned on SQuAD-style CORD-19 annotations for the task of extractive question answering in the context of COVID-19. It can be used with the Hugging Face Transformers library for question answering tasks.\\'}', metadata={})]", "category": "generic"}
{"question_id": 101, "text": " Develop a program which can answer questions related to a scanned document.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 102, "text": " I have received a PDF document and a question. My task is to find the answer part in the document.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 103, "text": " An interior design firm builds a software to understand the depth of rooms captured in photographs for remodeling activities.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Unconditional Image Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Denoising Diffusion Probabilistic Models (DDPM)\\', \\'api_name\\': \\'google/ddpm-bedroom-256\\', \\'api_call\\': \"DDPMPipeline.from_pretrained(\\'google/ddpm-bedroom-256\\')\", \\'api_arguments\\': \\'None\\', \\'python_environment_requirements\\': \\'diffusers\\', \\'example_code\\': \\'!pip install diffusers\\\\nfrom diffusers import DDPMPipeline, DDIMPipeline, PNDMPipeline\\\\nmodel_id = google/ddpm-bedroom-256\\\\nddpm = DDPMPipeline.from_pretrained(model_id)\\\\nimage = ddpm().images[0]\\\\nimage.save(ddpm_generated_image.png)\\', \\'performance\\': {\\'dataset\\': \\'CIFAR10\\', \\'accuracy\\': {\\'Inception score\\': 9.46, \\'FID score\\': 3.17}}, \\'description\\': \\'We present high quality image synthesis results using diffusion probabilistic models, a class of latent variable models inspired by considerations from nonequilibrium thermodynamics. Our best results are obtained by training on a weighted variational bound designed according to a novel connection between diffusion probabilistic models and denoising score matching with Langevin dynamics, and our models naturally admit a progressive lossy decompression scheme that can be interpreted as a generalization of autoregressive decoding. On the unconditional CIFAR10 dataset, we obtain an Inception score of 9.46 and a state-of-the-art FID score of 3.17. On 256x256 LSUN, we obtain sample quality similar to ProgressiveGAN.\\'}', metadata={})]", "category": "generic"}
{"question_id": 104, "text": " We are running an autonomous vehicle company and want to implement a depth estimation module for the real-time video feed captured by our camera.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Depth Estimation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Depth Estimation\\', \\'api_name\\': \\'glpn-kitti\\', \\'api_call\\': \"GLPNForDepthEstimation.from_pretrained(\\'vinvino02/glpn-kitti\\')\", \\'api_arguments\\': \\'images, return_tensors\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'from transformers import GLPNFeatureExtractor, GLPNForDepthEstimation\\\\nimport torch\\\\nimport numpy as np\\\\nfrom PIL import Image\\\\nimport requests\\\\nurl = http://images.cocodataset.org/val2017/000000039769.jpg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nfeature_extractor = GLPNFeatureExtractor.from_pretrained(vinvino02/glpn-kitti)\\\\nmodel = GLPNForDepthEstimation.from_pretrained(vinvino02/glpn-kitti)\\\\ninputs = feature_extractor(images=image, return_tensors=pt)\\\\nwith torch.no_grad():\\\\n outputs = model(**inputs)\\\\n predicted_depth = outputs.predicted_depth\\\\nprediction = torch.nn.functional.interpolate(\\\\n predicted_depth.unsqueeze(1),\\\\n size=image.size[::-1],\\\\n mode=bicubic,\\\\n align_corners=False,\\\\n)\\\\noutput = prediction.squeeze().cpu().numpy()\\\\nformatted = (output * 255 / np.max(output)).astype(uint8)\\\\ndepth = Image.fromarray(formatted)\\', \\'performance\\': {\\'dataset\\': \\'KITTI\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Global-Local Path Networks (GLPN) model trained on KITTI for monocular depth estimation. It was introduced in the paper Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth by Kim et al. and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 105, "text": " Our team wants to create a new app for autonomous vehicles. For that, we need to estimate the depth of the field from images.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Depth Estimation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Depth Estimation\\', \\'api_name\\': \\'glpn-kitti\\', \\'api_call\\': \"GLPNForDepthEstimation.from_pretrained(\\'vinvino02/glpn-kitti\\')\", \\'api_arguments\\': \\'images, return_tensors\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'from transformers import GLPNFeatureExtractor, GLPNForDepthEstimation\\\\nimport torch\\\\nimport numpy as np\\\\nfrom PIL import Image\\\\nimport requests\\\\nurl = http://images.cocodataset.org/val2017/000000039769.jpg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nfeature_extractor = GLPNFeatureExtractor.from_pretrained(vinvino02/glpn-kitti)\\\\nmodel = GLPNForDepthEstimation.from_pretrained(vinvino02/glpn-kitti)\\\\ninputs = feature_extractor(images=image, return_tensors=pt)\\\\nwith torch.no_grad():\\\\n outputs = model(**inputs)\\\\n predicted_depth = outputs.predicted_depth\\\\nprediction = torch.nn.functional.interpolate(\\\\n predicted_depth.unsqueeze(1),\\\\n size=image.size[::-1],\\\\n mode=bicubic,\\\\n align_corners=False,\\\\n)\\\\noutput = prediction.squeeze().cpu().numpy()\\\\nformatted = (output * 255 / np.max(output)).astype(uint8)\\\\ndepth = Image.fromarray(formatted)\\', \\'performance\\': {\\'dataset\\': \\'KITTI\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Global-Local Path Networks (GLPN) model trained on KITTI for monocular depth estimation. It was introduced in the paper Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth by Kim et al. and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 106, "text": " As a city planner, I need to measure the depth of spaces in a series of images taken from streets.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Depth Estimation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Depth Estimation\\', \\'api_name\\': \\'glpn-nyu-finetuned-diode\\', \\'api_call\\': \"pipeline(\\'depth-estimation\\', model=\\'sayakpaul/glpn-nyu-finetuned-diode\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'diode-subset\\', \\'accuracy\\': {\\'Loss\\': 0.4359, \\'Rmse\\': 0.4276}}, \\'description\\': \\'This model is a fine-tuned version of vinvino02/glpn-nyu on the diode-subset dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 107, "text": " In our online ecommerce platform, we want to build an AI app to automatically recognize the type of products. It should be able to identify common items like clothing, electronics, furniture, and more.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'abhishek/autotrain-dog-vs-food\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'abhishek/autotrain-dog-vs-food\\')\", \\'api_arguments\\': \\'image_path\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'sasha/dog-food\\', \\'accuracy\\': 0.998}, \\'description\\': \"A pre-trained model for classifying images as either dog or food using Hugging Face\\'s AutoTrain framework.\"}', metadata={})]", "category": "generic"}
{"question_id": 108, "text": " We need to recognize the breed of dog in the given image.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'julien-c/hotdog-not-hotdog\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'julien-c/hotdog-not-hotdog\\')\", \\'api_arguments\\': \\'image\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': 0.825}, \\'description\\': \\'A model that classifies images as hotdog or not hotdog.\\'}', metadata={})]", "category": "generic"}
{"question_id": 109, "text": " Develop a solution that can categorize an image of a cell phone, laptop, or smartwatch as one of these respective device types.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'julien-c/hotdog-not-hotdog\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'julien-c/hotdog-not-hotdog\\')\", \\'api_arguments\\': \\'image\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': 0.825}, \\'description\\': \\'A model that classifies images as hotdog or not hotdog.\\'}', metadata={})]", "category": "generic"}
{"question_id": 110, "text": " Build a system to help companies identify logos from a collection of images.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Zero-Shot Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Zero-Shot Image Classification\\', \\'api_name\\': \\'patrickjohncyh/fashion-clip\\', \\'api_call\\': \"CLIPModel.from_pretrained(\\'patrickjohncyh/fashion-clip\\')\", \\'api_arguments\\': {\\'image\\': \\'File\\', \\'class_names\\': \\'String (comma-separated)\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import CLIPProcessor, CLIPModel; model = CLIPModel.from_pretrained(\\'patrickjohncyh/fashion-clip\\'); processor = CLIPProcessor.from_pretrained(\\'patrickjohncyh/fashion-clip\\'); inputs = processor(text=\\'blue shoes\\', images=image, return_tensors=\\'pt\\', padding=True); logits_per_image = model(**inputs).logits_per_image; probs = logits_per_image.softmax(dim=-1).tolist()[0]\", \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'FMNIST\\', \\'accuracy\\': 0.83}, {\\'name\\': \\'KAGL\\', \\'accuracy\\': 0.73}, {\\'name\\': \\'DEEP\\', \\'accuracy\\': 0.62}]}, \\'description\\': \\'FashionCLIP is a CLIP-based model developed to produce general product representations for fashion concepts. Leveraging the pre-trained checkpoint (ViT-B/32) released by OpenAI, it is trained on a large, high-quality novel fashion dataset to study whether domain specific fine-tuning of CLIP-like models is sufficient to produce product representations that are zero-shot transferable to entirely new datasets and tasks.\\'}', metadata={})]", "category": "generic"}
{"question_id": 111, "text": " Develop a pipeline that detects objects present in an image using computer vision.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'microsoft/table-transformer-structure-recognition\\', \\'api_call\\': \"pipeline(\\'object-detection\\', model=\\'microsoft/table-transformer-structure-recognition\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'PubTables1M\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Table Transformer (DETR) model trained on PubTables1M for detecting the structure (like rows, columns) in tables.\\'}', metadata={})]", "category": "generic"}
{"question_id": 112, "text": " Assit me to process and segment an image for further analysis.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Segmentation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Segmentation\\', \\'api_name\\': \\'keremberke/yolov8m-pcb-defect-segmentation\\', \\'api_call\\': \"YOLO(\\'keremberke/yolov8m-pcb-defect-segmentation\\')\", \\'api_arguments\\': {\\'image\\': \\'URL or local path to the image\\'}, \\'python_environment_requirements\\': [\\'ultralyticsplus==0.0.24\\', \\'ultralytics==8.0.23\\'], \\'example_code\\': [\\'from ultralyticsplus import YOLO, render_result\\', \"model = YOLO(\\'keremberke/yolov8m-pcb-defect-segmentation\\')\", \"model.overrides[\\'conf\\'] = 0.25\", \"model.overrides[\\'iou\\'] = 0.45\", \"model.overrides[\\'agnostic_nms\\'] = False\", \"model.overrides[\\'max_det\\'] = 1000\", \"image = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\", \\'results = model.predict(image)\\', \\'print(results[0].boxes)\\', \\'print(results[0].masks)\\', \\'render = render_result(model=model, image=image, result=results[0])\\', \\'render.show()\\'], \\'performance\\': {\\'dataset\\': \\'pcb-defect-segmentation\\', \\'accuracy\\': {\\'mAP@0.5(box)\\': 0.568, \\'mAP@0.5(mask)\\': 0.557}}, \\'description\\': \\'A YOLOv8 model for PCB defect segmentation trained on the pcb-defect-segmentation dataset. The model can detect and segment defects in PCB images, such as Dry_joint, Incorrect_installation, PCB_damage, and Short_circuit.\\'}', metadata={})]", "category": "generic"}
{"question_id": 113, "text": " We need to analyze satellite images to categorize the types of land use. For this purpose, I need to segment the images and identify different objects.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Segmentation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Segmentation\\', \\'api_name\\': \\'keremberke/yolov8s-building-segmentation\\', \\'api_call\\': \"YOLO(\\'keremberke/yolov8s-building-segmentation\\')\", \\'api_arguments\\': [\\'conf\\', \\'iou\\', \\'agnostic_nms\\', \\'max_det\\', \\'image\\'], \\'python_environment_requirements\\': [\\'ultralyticsplus==0.0.21\\'], \\'example_code\\': [\\'from ultralyticsplus import YOLO, render_result\\', \"model = YOLO(\\'keremberke/yolov8s-building-segmentation\\')\", \"model.overrides[\\'conf\\'] = 0.25\", \"model.overrides[\\'iou\\'] = 0.45\", \"model.overrides[\\'agnostic_nms\\'] = False\", \"model.overrides[\\'max_det\\'] = 1000\", \"image = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\", \\'results = model.predict(image)\\', \\'print(results[0].boxes)\\', \\'print(results[0].masks)\\', \\'render = render_result(model=model, image=image, result=results[0])\\', \\'render.show()\\'], \\'performance\\': {\\'dataset\\': \\'satellite-building-segmentation\\', \\'accuracy\\': {\\'mAP@0.5(box)\\': 0.661, \\'mAP@0.5(mask)\\': 0.651}}, \\'description\\': \\'A YOLOv8 model for building segmentation in satellite images. Trained on the satellite-building-segmentation dataset, it can detect and segment buildings with high accuracy.\\'}', metadata={})]", "category": "generic"}
{"question_id": 114, "text": " We are a city planning department and want to evaluate the city layout. Analyze the image we provide to segment and understand the various urban elements.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Segmentation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Semantic Segmentation\\', \\'api_name\\': \\'nvidia/segformer-b2-finetuned-cityscapes-1024-1024\\', \\'api_call\\': \"SegformerForSemanticSegmentation.from_pretrained(\\'nvidia/segformer-b2-finetuned-cityscapes-1024-1024\\')\", \\'api_arguments\\': {\\'images\\': \\'image\\', \\'return_tensors\\': \\'pt\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'latest\\', \\'PIL\\': \\'latest\\', \\'requests\\': \\'latest\\'}, \\'example_code\\': \"from transformers import SegformerFeatureExtractor, SegformerForSemanticSegmentation\\\\nfrom PIL import Image\\\\nimport requests\\\\nfeature_extractor = SegformerFeatureExtractor.from_pretrained(\\'nvidia/segformer-b2-finetuned-cityscapes-1024-1024\\')\\\\nmodel = SegformerForSemanticSegmentation.from_pretrained(\\'nvidia/segformer-b2-finetuned-cityscapes-1024-1024\\')\\\\nurl = \\'http://images.cocodataset.org/val2017/000000039769.jpg\\'\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\ninputs = feature_extractor(images=image, return_tensors=\\'pt\\')\\\\noutputs = model(**inputs)\\\\nlogits = outputs.logits\", \\'performance\\': {\\'dataset\\': \\'Cityscapes\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'SegFormer model fine-tuned on CityScapes at resolution 1024x1024. It was introduced in the paper SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers by Xie et al. and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 115, "text": " My company develops drones for agriculture purposes, and we need a model to segment aerial images accurately.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Segmentation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'clipseg-rd64-refined\\', \\'api_call\\': \"pipeline(\\'image-segmentation\\', model=\\'CIDAS/clipseg-rd64-refined\\')\", \\'api_arguments\\': {\\'model\\': \\'CIDAS/clipseg-rd64-refined\\'}, \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'CLIPSeg model with reduce dimension 64, refined (using a more complex convolution). It was introduced in the paper Image Segmentation Using Text and Image Prompts by L\u00fcddecke et al. and first released in this repository. This model is intended for zero-shot and one-shot image segmentation.\\'}', metadata={})]", "category": "generic"}
{"question_id": 116, "text": " I want to generate images from text descriptions and use the scribble images as control inputs for my project.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image-to-Image\\', \\'framework\\': \\'Diffusers\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'lllyasviel/control_v11p_sd15_scribble\\', \\'api_call\\': \"ControlNetModel.from_pretrained(\\'lllyasviel/control_v11p_sd15_scribble\\')\", \\'api_arguments\\': {\\'checkpoint\\': \\'lllyasviel/control_v11p_sd15_scribble\\', \\'torch_dtype\\': \\'torch.float16\\'}, \\'python_environment_requirements\\': [\\'diffusers\\', \\'transformers\\', \\'accelerate\\', \\'controlnet_aux==0.3.0\\'], \\'example_code\\': \"import torch\\\\nimport os\\\\nfrom huggingface_hub import HfApi\\\\nfrom pathlib import Path\\\\nfrom diffusers.utils import load_image\\\\nfrom PIL import Image\\\\nimport numpy as np\\\\nfrom controlnet_aux import PidiNetDetector, HEDdetector\\\\nfrom diffusers import (\\\\n ControlNetModel,\\\\n StableDiffusionControlNetPipeline,\\\\n UniPCMultistepScheduler,\\\\n)\\\\ncheckpoint = lllyasviel/control_v11p_sd15_scribble\\\\nimage = load_image(\\\\n https://huggingface.co/lllyasviel/control_v11p_sd15_scribble/resolve/main/images/input.png\\\\n)\\\\nprompt = royal chamber with fancy bed\\\\nprocessor = HEDdetector.from_pretrained(\\'lllyasviel/Annotators\\')\\\\ncontrol_image = processor(image, scribble=True)\\\\ncontrol_image.save(./images/control.png)\\\\ncontrolnet = ControlNetModel.from_pretrained(checkpoint, torch_dtype=torch.float16)\\\\npipe = StableDiffusionControlNetPipeline.from_pretrained(\\\\n runwayml/stable-diffusion-v1-5, controlnet=controlnet, torch_dtype=torch.float16\\\\n)\\\\npipe.scheduler = UniPCMultistepScheduler.from_config(pipe.scheduler.config)\\\\npipe.enable_model_cpu_offload()\\\\ngenerator = torch.manual_seed(0)\\\\nimage = pipe(prompt, num_inference_steps=30, generator=generator, image=control_image).images[0]\\\\nimage.save(\\'images/image_out.png\\')\", \\'performance\\': {\\'dataset\\': \\'Stable Diffusion v1-5\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'Controlnet v1.1 is a neural network structure to control diffusion models by adding extra conditions. This checkpoint corresponds to the ControlNet conditioned on Scribble images. It can be used in combination with Stable Diffusion, such as runwayml/stable-diffusion-v1-5.\\'}', metadata={})]", "category": "generic"}
{"question_id": 118, "text": " We want to recommend workouts to our users, based on the type of sports they enjoy. Help us classify sports videos.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Video Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Video Classification\\', \\'api_name\\': \\'MCG-NJU/videomae-base-short-finetuned-kinetics\\', \\'api_call\\': \"VideoMAEForVideoClassification.from_pretrained(\\'MCG-NJU/videomae-base-short-finetuned-kinetics\\')\", \\'api_arguments\\': [\\'video\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import VideoMAEImageProcessor, VideoMAEForVideoClassification\\\\nimport numpy as np\\\\nimport torch\\\\nvideo = list(np.random.randn(16, 3, 224, 224))\\\\nprocessor = VideoMAEImageProcessor.from_pretrained(\\'MCG-NJU/videomae-base-short-finetuned-kinetics\\')\\\\nmodel = VideoMAEForVideoClassification.from_pretrained(\\'MCG-NJU/videomae-base-short-finetuned-kinetics\\')\\\\ninputs = processor(video, return_tensors=\\'pt\\')\\\\nwith torch.no_grad():\\\\n outputs = model(**inputs)\\\\n logits = outputs.logits\\\\npredicted_class_idx = logits.argmax(-1).item()\\\\nprint(\\'Predicted class:\\', model.config.id2label[predicted_class_idx])\", \\'performance\\': {\\'dataset\\': \\'Kinetics-400\\', \\'accuracy\\': {\\'top-1\\': 79.4, \\'top-5\\': 94.1}}, \\'description\\': \\'VideoMAE model pre-trained for 800 epochs in a self-supervised way and fine-tuned in a supervised way on Kinetics-400. It was introduced in the paper VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training by Tong et al. and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 119, "text": " We need to classify videos showing different actions for our new video moderation system.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Video Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Video Classification\\', \\'api_name\\': \\'MCG-NJU/videomae-base-short-finetuned-kinetics\\', \\'api_call\\': \"VideoMAEForVideoClassification.from_pretrained(\\'MCG-NJU/videomae-base-short-finetuned-kinetics\\')\", \\'api_arguments\\': [\\'video\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import VideoMAEImageProcessor, VideoMAEForVideoClassification\\\\nimport numpy as np\\\\nimport torch\\\\nvideo = list(np.random.randn(16, 3, 224, 224))\\\\nprocessor = VideoMAEImageProcessor.from_pretrained(\\'MCG-NJU/videomae-base-short-finetuned-kinetics\\')\\\\nmodel = VideoMAEForVideoClassification.from_pretrained(\\'MCG-NJU/videomae-base-short-finetuned-kinetics\\')\\\\ninputs = processor(video, return_tensors=\\'pt\\')\\\\nwith torch.no_grad():\\\\n outputs = model(**inputs)\\\\n logits = outputs.logits\\\\npredicted_class_idx = logits.argmax(-1).item()\\\\nprint(\\'Predicted class:\\', model.config.id2label[predicted_class_idx])\", \\'performance\\': {\\'dataset\\': \\'Kinetics-400\\', \\'accuracy\\': {\\'top-1\\': 79.4, \\'top-5\\': 94.1}}, \\'description\\': \\'VideoMAE model pre-trained for 800 epochs in a self-supervised way and fine-tuned in a supervised way on Kinetics-400. It was introduced in the paper VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training by Tong et al. and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 120, "text": " I need the AI to tell if an image is of a cat or a dog, regardless of its breed or species.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'abhishek/autotrain-dog-vs-food\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'abhishek/autotrain-dog-vs-food\\')\", \\'api_arguments\\': \\'image_path\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'sasha/dog-food\\', \\'accuracy\\': 0.998}, \\'description\\': \"A pre-trained model for classifying images as either dog or food using Hugging Face\\'s AutoTrain framework.\"}', metadata={})]", "category": "generic"}
{"question_id": 121, "text": " We have a set of pictures for pets (dogs and cats). We need to offer an AI-based solution to classify the pictures given the pet name.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'abhishek/autotrain-dog-vs-food\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'abhishek/autotrain-dog-vs-food\\')\", \\'api_arguments\\': \\'image_path\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'sasha/dog-food\\', \\'accuracy\\': 0.998}, \\'description\\': \"A pre-trained model for classifying images as either dog or food using Hugging Face\\'s AutoTrain framework.\"}', metadata={})]", "category": "generic"}
{"question_id": 122, "text": " Our startup team is now building an app for diagnosing plant diseases based on images. We need to get the diagnosis for different types of plant issues.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'martinezomg/vit-base-patch16-224-diabetic-retinopathy\\', \\'api_call\\': \"pipeline(\\'image-classification\\', \\'martinezomg/vit-base-patch16-224-diabetic-retinopathy\\')\", \\'api_arguments\\': {\\'model_name\\': \\'martinezomg/vit-base-patch16-224-diabetic-retinopathy\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'4.28.1\\', \\'pytorch\\': \\'2.0.0+cu118\\', \\'datasets\\': \\'2.11.0\\', \\'tokenizers\\': \\'0.13.3\\'}, \\'example_code\\': \"from transformers import pipeline\\\\nimage_classifier = pipeline(\\'image-classification\\', \\'martinezomg/vit-base-patch16-224-diabetic-retinopathy\\')\\\\nresult = image_classifier(\\'path/to/image.jpg\\')\", \\'performance\\': {\\'dataset\\': \\'None\\', \\'accuracy\\': 0.7744}, \\'description\\': \\'This model is a fine-tuned version of google/vit-base-patch16-224 on the None dataset. It is designed for image classification tasks, specifically for diabetic retinopathy detection.\\'}', metadata={})]", "category": "generic"}
{"question_id": 123, "text": " We need to analyze customer reviews and find out how well our new product is doing in the market.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'michellejieli/NSFW_text_classifier\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'michellejieli/NSFW_text_classification\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'classifier(I see you\u2019ve set aside this special time to humiliate yourself in public.)\\', \\'performance\\': {\\'dataset\\': \\'Reddit posts\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'DistilBERT is a transformer model that performs sentiment analysis. I fine-tuned the model on Reddit posts with the purpose of classifying not safe for work (NSFW) content, specifically text that is considered inappropriate and unprofessional. The model predicts 2 classes, which are NSFW or safe for work (SFW). The model is a fine-tuned version of DistilBERT. It was fine-tuned on 14317 Reddit posts pulled from the Reddit API.\\'}', metadata={})]", "category": "generic"}
{"question_id": 124, "text": " A new tutoring company is founded, and they want a tutoring AI. To do so, they need help in creating better explanations for a chemistry concept.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Generation\\', \\'api_name\\': \\'mywateriswet/ShuanBot\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'mywateriswet/ShuanBot\\')\", \\'api_arguments\\': \\'message\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"response = chatbot(\\'What is your name?\\')\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'ShuanBot is a conversational chatbot model based on the GPT-2 architecture. It can be used for generating human-like responses in a chat context.\\'}', metadata={})]", "category": "generic"}
{"question_id": 125, "text": " Create a function that can determine if a given text is a question or a statement.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Classification\\', \\'api_name\\': \\'shahrukhx01/question-vs-statement-classifier\\', \\'api_call\\': \"AutoModelForSequenceClassification.from_pretrained(\\'shahrukhx01/question-vs-statement-classifier\\')\", \\'api_arguments\\': {\\'tokenizer\\': \\'AutoTokenizer.from_pretrained(shahrukhx01/question-vs-statement-classifier)\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'from transformers import AutoTokenizer, AutoModelForSequenceClassification\\'}, \\'example_code\\': \\'tokenizer = AutoTokenizer.from_pretrained(shahrukhx01/question-vs-statement-classifier)\\\\nmodel = AutoModelForSequenceClassification.from_pretrained(shahrukhx01/question-vs-statement-classifier)\\', \\'performance\\': {\\'dataset\\': \\'Haystack\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Trained to add the feature for classifying queries between Question Query vs Statement Query using classification in Haystack\\'}', metadata={})]", "category": "generic"}
{"question_id": 126, "text": " I want to create a system that can answer questions by sorting out possible answers to a question.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'distilbert-base-uncased-distilled-squad\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=\\'distilbert-base-uncased-distilled-squad\\')\", \\'api_arguments\\': [\\'question\\', \\'context\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nquestion_answerer = pipeline(question-answering, model=\\'distilbert-base-uncased-distilled-squad\\')\\\\ncontext = r\\\\n... Extractive Question Answering is the task of extracting an answer from a text given a question. An example of a\\\\n... question answering dataset is the SQuAD dataset, which is entirely based on that task. If you would like to fine-tune\\\\n... a model on a SQuAD task, you may leverage the examples/pytorch/question-answering/run_squad.py script.\\\\n... \\\\nresult = question_answerer(question=What is a good example of a question answering dataset?, context=context)\\\\nprint(\\\\n... fAnswer: \\'{result[\\'answer\\']}\\', score: {round(result[\\'score\\'], 4)}, start: {result[\\'start\\']}, end: {result[\\'end\\']}\\\\n...)\", \\'performance\\': {\\'dataset\\': \\'SQuAD v1.1\\', \\'accuracy\\': \\'86.9 F1 score\\'}, \\'description\\': \"DistilBERT base uncased distilled SQuAD is a fine-tuned version of DistilBERT-base-uncased for the task of question answering. It has 40% less parameters than bert-base-uncased, runs 60% faster while preserving over 95% of BERT\\'s performances as measured on the GLUE language understanding benchmark.\"}', metadata={})]", "category": "generic"}
{"question_id": 127, "text": " We have a news article and we need to extract all the entities like the names of people, organizations, and locations.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Token Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Named Entity Recognition\\', \\'api_name\\': \\'dslim/bert-base-NER-uncased\\', \\'api_call\\': \"pipeline(\\'ner\\', model=\\'dslim/bert-base-NER-uncased\\')\", \\'api_arguments\\': {}, \\'python_environment_requirements\\': {\\'transformers\\': \\'>=4.0.0\\'}, \\'example_code\\': \"nlp(\\'My name is John and I live in New York.\\')\", \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A pretrained BERT model for Named Entity Recognition (NER) on uncased text. It can be used to extract entities such as person names, locations, and organizations from text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 128, "text": " We are purchasing a CRM system to keep track of our customers and their organizations. We want to extract useful entities from customer emails automatically.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Generation\\', \\'api_name\\': \\'mywateriswet/ShuanBot\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'mywateriswet/ShuanBot\\')\", \\'api_arguments\\': \\'message\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"response = chatbot(\\'What is your name?\\')\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'ShuanBot is a conversational chatbot model based on the GPT-2 architecture. It can be used for generating human-like responses in a chat context.\\'}', metadata={})]", "category": "generic"}
{"question_id": 129, "text": " As a researcher, I am trying to find an answer to my question in a table containing information about animals and their characteristics.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Table Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Table Question Answering\\', \\'api_name\\': \\'google/tapas-small-finetuned-sqa\\', \\'api_call\\': \"pipeline(\\'table-question-answering\\', model=\\'google/tapas-small-finetuned-sqa\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'msr_sqa\\', \\'accuracy\\': 0.6155}, \\'description\\': \\'TAPAS small model fine-tuned on Sequential Question Answering (SQA). It uses relative position embeddings (i.e. resetting the position index at every cell of the table).\\'}', metadata={})]", "category": "generic"}
{"question_id": 130, "text": " A teacher wants to create a quiz for her students. We are now working on the questions and answers for the quiz that be arranged in a table format.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Table Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Table Question Answering\\', \\'api_name\\': \\'google/tapas-small-finetuned-sqa\\', \\'api_call\\': \"pipeline(\\'table-question-answering\\', model=\\'google/tapas-small-finetuned-sqa\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'msr_sqa\\', \\'accuracy\\': 0.6155}, \\'description\\': \\'TAPAS small model fine-tuned on Sequential Question Answering (SQA). It uses relative position embeddings (i.e. resetting the position index at every cell of the table).\\'}', metadata={})]", "category": "generic"}
{"question_id": 131, "text": " I work for a financial company that stores all of its data in tables. We need a way to extract key information efficiently by asking natural language questions.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Table Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Table Question Answering\\', \\'api_name\\': \\'google/tapas-small-finetuned-sqa\\', \\'api_call\\': \"pipeline(\\'table-question-answering\\', model=\\'google/tapas-small-finetuned-sqa\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'msr_sqa\\', \\'accuracy\\': 0.6155}, \\'description\\': \\'TAPAS small model fine-tuned on Sequential Question Answering (SQA). It uses relative position embeddings (i.e. resetting the position index at every cell of the table).\\'}', metadata={})]", "category": "generic"}
{"question_id": 132, "text": " We have a dataset related to coffee and tea prices. We need to answer a question on who sells hot chocolate and their prices.\\n###Input: {\\\"table\\\": [[\\\"Shop\\\", \\\"Drink\\\", \\\"Price\\\"], [\\\"Cafe A\\\", \\\"Coffee\\\", \\\"3.00\\\"], [\\\"Cafe B\\\", \\\"Tea\\\", \\\"2.50\\\"], [\\\"Cafe C\\\", \\\"Hot Chocolate\\\", \\\"4.50\\\"], [\\\"Cafe D\\\", \\\"Hot Chocolate\\\", \\\"3.75\\\"]], \\\"queries\\\": [\\\"Which shops sell hot chocolate and what are their prices?\\\"]}\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Table Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Table Question Answering\\', \\'api_name\\': \\'google/tapas-small-finetuned-sqa\\', \\'api_call\\': \"pipeline(\\'table-question-answering\\', model=\\'google/tapas-small-finetuned-sqa\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'msr_sqa\\', \\'accuracy\\': 0.6155}, \\'description\\': \\'TAPAS small model fine-tuned on Sequential Question Answering (SQA). It uses relative position embeddings (i.e. resetting the position index at every cell of the table).\\'}', metadata={})]", "category": "generic"}
{"question_id": 133, "text": " A company is running a survey and they want to know how many respondents have given a specific answer for each question of the survey.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 134, "text": " Extract information about a non-compete clause from a legal document with a context related to data protection.\\n###Input: \\\"The data protection provisions set forth in this agreement shall be in effect for a period of 2 years after the termination of services. The non-compete clause states that the service provider is prohibited from providing similar services to any competitor within a 50-mile radius and during the 1-year period following termination of services.\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'PyTorch Transformers\\', \\'functionality\\': \\'text2text-generation\\', \\'api_name\\': \\'facebook/nllb-200-distilled-600M\\', \\'api_call\\': \"pipeline(\\'translation_xx_to_yy\\', model=\\'facebook/nllb-200-distilled-600M\\')\", \\'api_arguments\\': [\\'model\\', \\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\', \\'torch\\'], \\'example_code\\': \"from transformers import pipeline; translator = pipeline(\\'translation_xx_to_yy\\', model=\\'facebook/nllb-200-distilled-600M\\'); translator(\\'Hello World\\')\", \\'performance\\': {\\'dataset\\': \\'Flores-200\\', \\'accuracy\\': \\'BLEU, spBLEU, chrF++\\'}, \\'description\\': \\'NLLB-200 is a machine translation model primarily intended for research in machine translation, especially for low-resource languages. It allows for single sentence translation among 200 languages. The model was trained on general domain text data and is not intended to be used with domain specific texts, such as medical domain or legal domain. The model is not intended to be used for document translation.\\'}', metadata={})]", "category": "generic"}
{"question_id": 135, "text": " Tell me the day of the game when it was played given the following context: \\\"The game was played on February 7, 2016 at Levi's Stadium in the San Francisco Bay Area at Santa Clara, California.\\\"\\n###Input: {'context': \\\"The game was played on February 7, 2016 at Levi's Stadium in the San Francisco Bay Area at Santa Clara, California.\\\", 'question': \\\"What day was the game played on?\\\"}\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Question Answering\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'csarron/bert-base-uncased-squad-v1\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=\\'csarron/bert-base-uncased-squad-v1\\', tokenizer=\\'csarron/bert-base-uncased-squad-v1\\')\", \\'api_arguments\\': {\\'model\\': \\'csarron/bert-base-uncased-squad-v1\\', \\'tokenizer\\': \\'csarron/bert-base-uncased-squad-v1\\'}, \\'python_environment_requirements\\': \\'Python 3.7.5\\', \\'example_code\\': \"from transformers import pipeline\\\\nqa_pipeline = pipeline(\\\\n question-answering,\\\\n model=csarron/bert-base-uncased-squad-v1,\\\\n tokenizer=csarron/bert-base-uncased-squad-v1\\\\n)\\\\npredictions = qa_pipeline({\\\\n \\'context\\': The game was played on February 7, 2016 at Levi\\'s Stadium in the San Francisco Bay Area at Santa Clara, California.,\\\\n \\'question\\': What day was the game played on?\\\\n})\\\\nprint(predictions)\", \\'performance\\': {\\'dataset\\': \\'SQuAD1.1\\', \\'accuracy\\': {\\'EM\\': 80.9, \\'F1\\': 88.2}}, \\'description\\': \\'BERT-base uncased model fine-tuned on SQuAD v1. This model is case-insensitive and does not make a difference between english and English.\\'}', metadata={})]", "category": "generic"}
{"question_id": 136, "text": " We need to identify the relationship between two sentences whether they are contradictory, entail each other, or neutral.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Sentence Similarity\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Sentence Embeddings\\', \\'api_name\\': \\'sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2\\', \\'api_call\\': \"SentenceTransformer(\\'sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2\\')\", \\'api_arguments\\': [\\'sentences\\'], \\'python_environment_requirements\\': \\'pip install -U sentence-transformers\\', \\'example_code\\': \"from sentence_transformers import SentenceTransformer\\\\nsentences = [This is an example sentence, Each sentence is converted]\\\\nmodel = SentenceTransformer(\\'sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2\\')\\\\nembeddings = model.encode(sentences)\\\\nprint(embeddings)\", \\'performance\\': {\\'dataset\\': \\'https://seb.sbert.net\\', \\'accuracy\\': \\'Automated evaluation\\'}, \\'description\\': \\'This is a sentence-transformers model: It maps sentences & paragraphs to a 384 dimensional dense vector space and can be used for tasks like clustering or semantic search.\\'}', metadata={})]", "category": "generic"}
{"question_id": 137, "text": " I am a climate change agency, looking to have my research summaries translated into Chinese for international audiences.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Summarization\\', \\'api_name\\': \\'Randeng-Pegasus-238M-Summary-Chinese\\', \\'api_call\\': \"PegasusForConditionalGeneration.from_pretrained(\\'IDEA-CCNL/Randeng-Pegasus-238M-Summary-Chinese\\')\", \\'api_arguments\\': {\\'text\\': \\'string\\', \\'max_length\\': \\'integer\\'}, \\'python_environment_requirements\\': [\\'transformers\\', \\'tokenizers_pegasus.py\\', \\'data_utils.py\\'], \\'example_code\\': \"from transformers import PegasusForConditionalGeneration\\\\nfrom tokenizers_pegasus import PegasusTokenizer\\\\nmodel = PegasusForConditionalGeneration.from_pretrained(\\'IDEA-CCNL/Randeng-Pegasus-238M-Summary-Chinese\\')\\\\ntokenizer = PegasusTokenizer.from_pretrained(\\'IDEA-CCNL/Randeng-Pegasus-238M-Summary-Chinese\\')\\\\ntext = \\'1\\'\\\\ninputs = tokenizer(text, max_length=1024, return_tensors=\\'pt\\')\\\\nsummary_ids = model.generate(inputs[\\'input_ids\\'])\\\\ntokenizer.batch_decode(summary_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]\", \\'performance\\': {\\'dataset\\': \\'LCSTS\\', \\'accuracy\\': {\\'rouge-1\\': 43.46, \\'rouge-2\\': 29.59, \\'rouge-L\\': 39.76}}, \\'description\\': \\'Randeng-Pegasus-238M-Summary-Chinese is a Chinese text summarization model based on Pegasus. It is fine-tuned on 7 Chinese text summarization datasets including education, new2016zh, nlpcc, shence, sohu, thucnews, and weibo. The model can be used to generate summaries for Chinese text inputs.\\'}', metadata={})]", "category": "generic"}
{"question_id": 138, "text": " We are building an app to summarize long articles for users. We need a solution to create a condensed summary of the given text.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'pszemraj/long-t5-tglobal-base-16384-book-summary\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'pszemraj/long-t5-tglobal-base-16384-book-summary\\')\", \\'api_arguments\\': [\\'long_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nimport torch\\\\nsummarizer = pipeline(\\\\n summarization,\\\\n pszemraj/long-t5-tglobal-base-16384-book-summary,\\\\n device=0 if torch.cuda.is_available() else -1,\\\\n)\\\\nlong_text = Here is a lot of text I don\\'t want to read. Replace me\\\\nresult = summarizer(long_text)\\\\nprint(result[0][summary_text])\", \\'performance\\': {\\'dataset\\': \\'kmfoda/booksum\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.408, \\'ROUGE-2\\': 6.065, \\'ROUGE-L\\': 16.721, \\'ROUGE-LSUM\\': 33.34}}, \\'description\\': \\'A fine-tuned version of google/long-t5-tglobal-base on the kmfoda/booksum dataset, which can be used to summarize long text and generate SparkNotes-esque summaries of arbitrary topics. The model generalizes reasonably well to academic and narrative text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 139, "text": " A news agency wants to summaries their international news articles daily.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'it5-base-news-summarization\\', \\'api_call\\': \"pipeline(\\'summarization\\', model=\\'it5/it5-base-news-summarization\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'newsum(Dal 31 maggio \u00e8 infine partita la piattaforma ITsART, a pi\u00f9 di un anno da quando \u2013 durante il primo lockdown \u2013 il ministro della Cultura Dario Franceschini ne aveva parlato come di \u00abuna sorta di Netflix della cultura\u00bb, pensata per \u00aboffrire a tutto il mondo la cultura italiana a pagamento\u00bb. \u00c8 presto per dare giudizi definitivi sulla piattaforma, e di certo sar\u00e0 difficile farlo anche pi\u00f9 avanti senza numeri precisi. Al momento, l\u2019unica cosa che si pu\u00f2 fare \u00e8 guardare com\u2019\u00e8 fatto il sito, contare quanti contenuti ci sono (circa 700 \u201ctitoli\u201d, tra film, documentari, spettacoli teatrali e musicali e altri eventi) e provare a dare un giudizio sul loro valore e sulla loro variet\u00e0. Intanto, una cosa notata da pi\u00f9 parti \u00e8 che diversi contenuti di ITsART sono a pagamento sulla piattaforma sebbene altrove, per esempio su RaiPlay, siano invece disponibili gratuitamente.)\\', \\'performance\\': {\\'dataset\\': \\'NewsSum-IT\\', \\'accuracy\\': {\\'Rouge1\\': 0.339, \\'Rouge2\\': 0.16, \\'RougeL\\': 0.263}}, \\'description\\': \\'IT5 Base model fine-tuned on news summarization on the Fanpage and Il Post corpora for Italian Language Understanding and Generation.\\'}', metadata={})]", "category": "generic"}
{"question_id": 140, "text": " Design a smart home system that can have conversations with the user for controlling the home appliances.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'hyunwoongko/blenderbot-9B\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'hyunwoongko/blenderbot-9B\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'Input a message to start chatting with hyunwoongko/blenderbot-9B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high-performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi-turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 141, "text": " An educational publishing company is developing a language comprehension program for elementary school students. They want a system that can generate a short story based on a given prompt.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Paraphrase-based utterance augmentation\\', \\'api_name\\': \\'prithivida/parrot_fluency_model\\', \\'api_call\\': \"pipeline(\\'text-classification\\', model=\\'prithivida/parrot_fluency_model\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"parrot(\\'your input text\\')\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'Parrot is a paraphrase-based utterance augmentation framework purpose-built to accelerate training NLU models. A paraphrase framework is more than just a paraphrasing model.\\'}', metadata={})]", "category": "generic"}
{"question_id": 142, "text": " Can you initiate a conversation with an AI model that plays the role of a friend who just got back from a conference?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Generation\\', \\'api_name\\': \\'mywateriswet/ShuanBot\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'mywateriswet/ShuanBot\\')\", \\'api_arguments\\': \\'message\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"response = chatbot(\\'What is your name?\\')\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'ShuanBot is a conversational chatbot model based on the GPT-2 architecture. It can be used for generating human-like responses in a chat context.\\'}', metadata={})]", "category": "generic"}
{"question_id": 143, "text": " As a software developer, I want a program that can automatically generate code snippets for me based on natural language descriptions.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Program Synthesis\\', \\'api_name\\': \\'Salesforce/codegen-350M-multi\\', \\'api_call\\': \"AutoTokenizer.from_pretrained(\\'Salesforce/codegen-350M-multi\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'from transformers import AutoTokenizer, AutoModelForCausalLM\\\\ntokenizer = AutoTokenizer.from_pretrained(Salesforce/codegen-350M-multi)\\\\nmodel = AutoModelForCausalLM.from_pretrained(Salesforce/codegen-350M-multi)\\\\ntext = def hello_world():\\\\ninput_ids = tokenizer(text, return_tensors=pt).input_ids\\\\ngenerated_ids = model.generate(input_ids, max_length=128)\\\\nprint(tokenizer.decode(generated_ids[0], skip_special_tokens=True))\\', \\'performance\\': {\\'dataset\\': \\'HumanEval and MTPB\\', \\'accuracy\\': \\'Refer to the paper for accuracy details\\'}, \\'description\\': \\'CodeGen is a family of autoregressive language models for program synthesis. The checkpoint included in this repository is denoted as CodeGen-Multi 350M, where Multi means the model is initialized with CodeGen-NL 350M and further pre-trained on a dataset of multiple programming languages, and 350M refers to the number of trainable parameters. The model is capable of extracting features from given natural language and programming language texts, and calculating the likelihood of them. It is best at program synthesis, generating executable code given English prompts, and can complete partially-generated code as well.\\'}', metadata={})]", "category": "generic"}
{"question_id": 144, "text": " Provide a short summary of an article about cryptocurrency investment risks.\\n###Input: Cryptocurrencies have become exceedingly popular among investors seeking higher returns and diversification in their portfolios. However, investing in these digital currencies carries several inherent risks. Market volatility is a major factor \\u2013 cryptocurrencies can experience wild price swings, sometimes even within hours or minutes. This high volatility makes it difficult to predict the future value of the investments and can result in significant losses. Furthermore, the lack of regulatory oversight and security concerns may also lead to potential frauds and hacks, exposing investors to additional risk. Lastly, the environmental impact of mining digital currencies like Bitcoin has come under scrutiny, questioning the long-term sustainability of the cryptocurrency market.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Abstractive Russian Summarization\\', \\'api_name\\': \\'cointegrated/rut5-base-absum\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'cointegrated/rut5-base-absum\\')\", \\'api_arguments\\': {\\'n_words\\': \\'int\\', \\'compression\\': \\'float\\', \\'max_length\\': \\'int\\', \\'num_beams\\': \\'int\\', \\'do_sample\\': \\'bool\\', \\'repetition_penalty\\': \\'float\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'latest\\', \\'torch\\': \\'latest\\'}, \\'example_code\\': \"import torch\\\\nfrom transformers import T5ForConditionalGeneration, T5Tokenizer\\\\nMODEL_NAME = \\'cointegrated/rut5-base-absum\\'\\\\nmodel = T5ForConditionalGeneration.from_pretrained(MODEL_NAME)\\\\ntokenizer = T5Tokenizer.from_pretrained(MODEL_NAME)\\\\nmodel.cuda();\\\\nmodel.eval();\\\\ndef summarize(\\\\n text, n_words=None, compression=None,\\\\n max_length=1000, num_beams=3, do_sample=False, repetition_penalty=10.0, \\\\n <strong>kwargs\\\\n):\\\\n \\\\n Summarize the text\\\\n The following parameters are mutually exclusive:\\\\n - n_words (int) is an approximate number of words to generate.\\\\n - compression (float) is an approximate length ratio of summary and original text.\\\\n \\\\n if n_words:\\\\n text = \\'[{}] \\'.format(n_words) + text\\\\n elif compression:\\\\n text = \\'[{0:.1g}] \\'.format(compression) + text\\\\n x = tokenizer(text, return_tensors=\\'pt\\', padding=True).to(model.device)\\\\n with torch.inference_mode():\\\\n out = model.generate(\\\\n </strong>x, \\\\n max_length=max_length, num_beams=num_beams, \\\\n do_sample=do_sample, repetition_penalty=repetition_penalty, \\\\n **kwargs\\\\n )\\\\n return tokenizer.decode(out[0], skip_special_tokens=True)\", \\'performance\\': {\\'dataset\\': [\\'csebuetnlp/xlsum\\', \\'IlyaGusev/gazeta\\', \\'mlsum\\'], \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This is a model for abstractive Russian summarization, based on cointegrated/rut5-base-multitask and fine-tuned on 4 datasets.\\'}', metadata={})]", "category": "generic"}
{"question_id": 145, "text": " You have just met a person that speaks French. As a hotel manager, you need to tell them, \\\"Welcome to our hotel, we hope you enjoy your stay.\\\" in French.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Translation\\', \\'api_name\\': \\'Helsinki-NLP/opus-mt-en-fr\\', \\'api_call\\': \"translate(\\'input_text\\', model=\\'Helsinki-NLP/opus-mt-en-fr\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'opus\\', \\'accuracy\\': {\\'BLEU\\': {\\'newsdiscussdev2015-enfr.en.fr\\': 33.8, \\'newsdiscusstest2015-enfr.en.fr\\': 40.0, \\'newssyscomb2009.en.fr\\': 29.8, \\'news-test2008.en.fr\\': 27.5, \\'newstest2009.en.fr\\': 29.4, \\'newstest2010.en.fr\\': 32.7, \\'newstest2011.en.fr\\': 34.3, \\'newstest2012.en.fr\\': 31.8, \\'newstest2013.en.fr\\': 33.2, \\'Tatoeba.en.fr\\': 50.5}}}, \\'description\\': \\'Helsinki-NLP/opus-mt-en-fr is a translation model that translates English text to French using the Hugging Face Transformers library. It is based on the OPUS dataset and uses a transformer-align architecture with normalization and SentencePiece pre-processing.\\'}', metadata={})]", "category": "generic"}
{"question_id": 146, "text": " They are planning a trip to Germany and want to spend some leisure time in the parks of Munich, find out how to ask a question about the location of parks in Munich in German.\\n \n Use this API documentation for reference: [Document(page_content=\"{'domain': 'Natural Language Processing Token Classification', 'framework': 'Transformers', 'functionality': 'punctuation prediction', 'api_name': 'oliverguhr/fullstop-punctuation-multilang-large', 'api_call': 'PunctuationModel()', 'api_arguments': ['text'], 'python_environment_requirements': ['pip install deepmultilingualpunctuation'], 'example_code': 'from deepmultilingualpunctuation import PunctuationModel\\\\nmodel = PunctuationModel()\\\\ntext = My name is Clara and I live in Berkeley California Ist das eine Frage Frau M\u00fcller\\\\nresult = model.restore_punctuation(text)\\\\nprint(result)', 'performance': {'dataset': 'wmt/europarl', 'EN_accuracy': 0.775, 'DE_accuracy': 0.814, 'FR_accuracy': 0.782, 'IT_accuracy': 0.762}, 'description': 'This model predicts the punctuation of English, Italian, French and German texts. It was developed to restore the punctuation of transcribed spoken language and trained on the Europarl Dataset provided by the SEPP-NLG Shared Task. The model restores the following punctuation markers: ., ,, ?, -, :.'}\", metadata={})]", "category": "generic"}
{"question_id": 147, "text": " We are a company offering speech to text services. We need to summarize the conversion and make it open-ended question.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'pszemraj/long-t5-tglobal-base-16384-book-summary\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'pszemraj/long-t5-tglobal-base-16384-book-summary\\')\", \\'api_arguments\\': [\\'long_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nimport torch\\\\nsummarizer = pipeline(\\\\n summarization,\\\\n pszemraj/long-t5-tglobal-base-16384-book-summary,\\\\n device=0 if torch.cuda.is_available() else -1,\\\\n)\\\\nlong_text = Here is a lot of text I don\\'t want to read. Replace me\\\\nresult = summarizer(long_text)\\\\nprint(result[0][summary_text])\", \\'performance\\': {\\'dataset\\': \\'kmfoda/booksum\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.408, \\'ROUGE-2\\': 6.065, \\'ROUGE-L\\': 16.721, \\'ROUGE-LSUM\\': 33.34}}, \\'description\\': \\'A fine-tuned version of google/long-t5-tglobal-base on the kmfoda/booksum dataset, which can be used to summarize long text and generate SparkNotes-esque summaries of arbitrary topics. The model generalizes reasonably well to academic and narrative text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 148, "text": " To help me with my writing, I need an AI tool that can fill in the gaps for me when I get stuck. It should be able to complete phrases or sentences.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Paraphrasing\\', \\'api_name\\': \\'prithivida/parrot_paraphraser_on_T5\\', \\'api_call\\': \"Parrot(model_tag=\\'prithivida/parrot_paraphraser_on_T5\\', use_gpu=False)\", \\'api_arguments\\': [\\'input_phrase\\', \\'diversity_ranker\\', \\'do_diverse\\', \\'max_return_phrases\\', \\'max_length\\', \\'adequacy_threshold\\', \\'fluency_threshold\\'], \\'python_environment_requirements\\': [\\'torch\\', \\'transformers\\'], \\'example_code\\': \\'from parrot import Parrot\\\\nimport torch\\\\nimport warnings\\\\nwarnings.filterwarnings(ignore)\\\\n\\\\nparrot = Parrot(model_tag=prithivida/parrot_paraphraser_on_T5, use_gpu=False)\\\\nphrases = [Can you recommed some upscale restaurants in Newyork?,\\\\n What are the famous places we should not miss in Russia?\\\\n]\\\\nfor phrase in phrases:\\\\n print(-*100)\\\\n print(Input_phrase: , phrase)\\\\n print(-*100)\\\\n para_phrases = parrot.augment(input_phrase=phrase)\\\\n for para_phrase in para_phrases:\\\\n  print(para_phrase)\\', \\'performance\\': {\\'dataset\\': \\'Not mentioned\\', \\'accuracy\\': \\'Not mentioned\\'}, \\'description\\': \\'Parrot is a paraphrase based utterance augmentation framework purpose built to accelerate training NLU models. It offers knobs to control Adequacy, Fluency, and Diversity as per your needs. It mainly focuses on augmenting texts typed-into or spoken-to conversational interfaces for building robust NLU models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 149, "text": " A writer needs help with generating the next word in the phrase \\\"The dog jumped over the\\\" __.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Paraphrasing\\', \\'api_name\\': \\'prithivida/parrot_paraphraser_on_T5\\', \\'api_call\\': \"Parrot(model_tag=\\'prithivida/parrot_paraphraser_on_T5\\', use_gpu=False)\", \\'api_arguments\\': [\\'input_phrase\\', \\'diversity_ranker\\', \\'do_diverse\\', \\'max_return_phrases\\', \\'max_length\\', \\'adequacy_threshold\\', \\'fluency_threshold\\'], \\'python_environment_requirements\\': [\\'torch\\', \\'transformers\\'], \\'example_code\\': \\'from parrot import Parrot\\\\nimport torch\\\\nimport warnings\\\\nwarnings.filterwarnings(ignore)\\\\n\\\\nparrot = Parrot(model_tag=prithivida/parrot_paraphraser_on_T5, use_gpu=False)\\\\nphrases = [Can you recommed some upscale restaurants in Newyork?,\\\\n What are the famous places we should not miss in Russia?\\\\n]\\\\nfor phrase in phrases:\\\\n print(-*100)\\\\n print(Input_phrase: , phrase)\\\\n print(-*100)\\\\n para_phrases = parrot.augment(input_phrase=phrase)\\\\n for para_phrase in para_phrases:\\\\n  print(para_phrase)\\', \\'performance\\': {\\'dataset\\': \\'Not mentioned\\', \\'accuracy\\': \\'Not mentioned\\'}, \\'description\\': \\'Parrot is a paraphrase based utterance augmentation framework purpose built to accelerate training NLU models. It offers knobs to control Adequacy, Fluency, and Diversity as per your needs. It mainly focuses on augmenting texts typed-into or spoken-to conversational interfaces for building robust NLU models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 151, "text": " Help me find similarity scores for different restaurant reviews.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Sentence Similarity\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Sentence Transformers\\', \\'api_name\\': \\'nikcheerla/nooks-amd-detection-realtime\\', \\'api_call\\': \"SentenceTransformer(\\'nikcheerla/nooks-amd-detection-realtime\\')\", \\'api_arguments\\': [\\'sentences\\'], \\'python_environment_requirements\\': \\'pip install -U sentence-transformers\\', \\'example_code\\': \"from sentence_transformers import SentenceTransformer\\\\nsentences = [This is an example sentence, Each sentence is converted]\\\\nmodel = SentenceTransformer(\\'{MODEL_NAME}\\')\\\\nembeddings = model.encode(sentences)\\\\nprint(embeddings)\", \\'performance\\': {\\'dataset\\': \\'https://seb.sbert.net\\', \\'accuracy\\': \\'Automated evaluation\\'}, \\'description\\': \\'This is a sentence-transformers model that maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search.\\'}', metadata={})]", "category": "generic"}
{"question_id": 152, "text": " Our company is building an automated assistance system for the visually impaired. We need a tool to convert text into spoken instructions.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'kan-bayashi_ljspeech_vits\\', \\'api_call\\': \"pipeline(\\'text-to-speech\\', model=\\'espnet/kan-bayashi_ljspeech_vits\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"from transformers import pipeline; tts = pipeline(\\'text-to-speech\\', model=\\'espnet/kan-bayashi_ljspeech_vits\\'); tts(\\'Hello World\\')\", \\'performance\\': {\\'dataset\\': \\'ljspeech\\', \\'accuracy\\': \\'Not mentioned\\'}, \\'description\\': \\'A Text-to-Speech model trained on the ljspeech dataset using the ESPnet toolkit. This model can be used to convert text input into synthesized speech.\\'}', metadata={})]", "category": "generic"}
{"question_id": 153, "text": " I would like to create an application where users may listen to translations of English sentences. I need a Text-to-Speech model to support this functionality.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'mio/Artoria\\', \\'api_call\\': \"pipeline(\\'text-to-speech\\', model=\\'mio/Artoria\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline; tts = pipeline(\\'text-to-speech\\', model=\\'mio/Artoria\\'); tts(\\'s\\')\", \\'performance\\': {\\'dataset\\': \\'fate\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This model was trained by mio using fate recipe in espnet. It is a text-to-speech model that can convert text input into speech output.\\'}', metadata={})]", "category": "generic"}
{"question_id": 154, "text": " We are a language learning app platform. We need a text-to-speech tool to read a sentence in multiple languages for our users.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'SYSPIN/Marathi_Male_TTS\\', \\'api_call\\': \"api.load(\\'ESPnet/espnet_model_zoo:SYSPIN/Marathi_Male_TTS\\').\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'huggingface_hub\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Marathi Male Text-to-Speech model using ESPnet framework.\\'}', metadata={})]", "category": "generic"}
{"question_id": 155, "text": " I need to create an audio output that translates the given text to speech for a French audiobook assistant.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Audio-to-Audio\\', \\'framework\\': \\'Fairseq\\', \\'functionality\\': \\'speech-to-speech-translation\\', \\'api_name\\': \\'facebook/textless_sm_en_fr\\', \\'api_call\\': \"load_model_ensemble_and_task_from_hf_hub(\\'facebook/textless_sm_en_fr\\')\", \\'api_arguments\\': [\\'input_file\\'], \\'python_environment_requirements\\': [\\'huggingface_hub\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'This model is a speech-to-speech translation model trained by Facebook. It is designed for translating English speech to French speech.\\'}', metadata={})]", "category": "generic"}
{"question_id": 156, "text": " We are opening a platform where users can record their own podcast and host it on our platform, can you help us to convert the audio into text automatically?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'lakahaga/novel_reading_tts\\', \\'api_call\\': \"AutoModelForTTS.from_pretrained(\\'lakahaga/novel_reading_tts\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"inputs = processor(text, return_tensors=\\'pt\\'); generated_audio = model.generate(**inputs);\", \\'performance\\': {\\'dataset\\': \\'novelspeech\\', \\'accuracy\\': None}, \\'description\\': \\'This model was trained by lakahaga using novelspeech recipe in espnet. It is designed for Korean text-to-speech tasks.\\'}', metadata={})]", "category": "generic"}
{"question_id": 157, "text": " We are currently working on the development of a speech-to-text application for transcription purposes. Can you help generate the required transcription code?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'CTranslate2\\', \\'functionality\\': \\'Automatic Speech Recognition\\', \\'api_name\\': \\'guillaumekln/faster-whisper-large-v2\\', \\'api_call\\': \"WhisperModel(\\'large-v2\\')\", \\'api_arguments\\': [\\'audio.mp3\\'], \\'python_environment_requirements\\': [\\'faster_whisper\\'], \\'example_code\\': \\'from faster_whisper import WhisperModel\\\\nmodel = WhisperModel(large-v2)\\\\nsegments, info = model.transcribe(audio.mp3)\\\\nfor segment in segments:\\\\n print([%.2fs -&gt; %.2fs] %s % (segment.start, segment.end, segment.text))\\', \\'performance\\': {\\'dataset\\': \\'99 languages\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Whisper large-v2 model for CTranslate2. This model can be used in CTranslate2 or projets based on CTranslate2 such as faster-whisper.\\'}', metadata={})]", "category": "generic"}
{"question_id": 158, "text": " We are building an app for tour guides to transcribe what they say during the tour. Then the transcript will be translated into sign language.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'CTranslate2\\', \\'functionality\\': \\'Automatic Speech Recognition\\', \\'api_name\\': \\'guillaumekln/faster-whisper-large-v2\\', \\'api_call\\': \"WhisperModel(\\'large-v2\\')\", \\'api_arguments\\': [\\'audio.mp3\\'], \\'python_environment_requirements\\': [\\'faster_whisper\\'], \\'example_code\\': \\'from faster_whisper import WhisperModel\\\\nmodel = WhisperModel(large-v2)\\\\nsegments, info = model.transcribe(audio.mp3)\\\\nfor segment in segments:\\\\n print([%.2fs -&gt; %.2fs] %s % (segment.start, segment.end, segment.text))\\', \\'performance\\': {\\'dataset\\': \\'99 languages\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Whisper large-v2 model for CTranslate2. This model can be used in CTranslate2 or projets based on CTranslate2 such as faster-whisper.\\'}', metadata={})]", "category": "generic"}
{"question_id": 159, "text": " I am an English teacher. I have recently met a Hokkien speaker. I am looking to translate a conversational sentence from English to Hokkien to facilitate communication.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Translation\\', \\'api_name\\': \\'Helsinki-NLP/opus-mt-en-zh\\', \\'api_call\\': \"pipeline(\\'translation_en_to_zh\\', model=\\'Helsinki-NLP/opus-mt-en-zh\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\ntranslation = pipeline(\\'translation_en_to_zh\\', model=\\'Helsinki-NLP/opus-mt-en-zh\\')\\\\ntranslated_text = translation(\\'Hello, world!\\')\", \\'performance\\': {\\'dataset\\': \\'Tatoeba-test.eng.zho\\', \\'accuracy\\': {\\'BLEU\\': 31.4, \\'chr-F\\': 0.268}}, \\'description\\': \"A translation model for English to Chinese using the Hugging Face Transformers library. It is based on the Marian NMT model and trained on the OPUS dataset. The model requires a sentence initial language token in the form of \\'>>id<<\\' (id = valid target language ID).\"}', metadata={})]", "category": "generic"}
{"question_id": 160, "text": " Determine the keyword spoken in a recorded audio file.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'CTranslate2\\', \\'functionality\\': \\'Automatic Speech Recognition\\', \\'api_name\\': \\'guillaumekln/faster-whisper-large-v2\\', \\'api_call\\': \"WhisperModel(\\'large-v2\\')\", \\'api_arguments\\': [\\'audio.mp3\\'], \\'python_environment_requirements\\': [\\'faster_whisper\\'], \\'example_code\\': \\'from faster_whisper import WhisperModel\\\\nmodel = WhisperModel(large-v2)\\\\nsegments, info = model.transcribe(audio.mp3)\\\\nfor segment in segments:\\\\n print([%.2fs -&gt; %.2fs] %s % (segment.start, segment.end, segment.text))\\', \\'performance\\': {\\'dataset\\': \\'99 languages\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Whisper large-v2 model for CTranslate2. This model can be used in CTranslate2 or projets based on CTranslate2 such as faster-whisper.\\'}', metadata={})]", "category": "generic"}
{"question_id": 161, "text": " Determine which speaker an audio segment belongs to using the provided audio file.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'pyannote.audio\\', \\'functionality\\': \\'Speaker diarization\\', \\'api_name\\': \\'johnislarry/cloned-pyannote-speaker-diarization-endpoint\\', \\'api_call\\': \"Pipeline.from_pretrained(\\'pyannote/speaker-diarization@2.1\\',use_auth_token=\\'ACCESS_TOKEN_GOES_HERE\\')\", \\'api_arguments\\': [\\'num_speakers\\', \\'min_speakers\\', \\'max_speakers\\', \\'segmentation_onset\\'], \\'python_environment_requirements\\': \\'pyannote.audio 2.0\\', \\'example_code\\': {\\'load_pipeline\\': \\'from pyannote.audio import Pipeline\\\\npipeline = Pipeline.from_pretrained(pyannote/speaker-diarization@2022.07)\\', \\'apply_pipeline\\': \\'diarization = pipeline(audio.wav)\\', \\'save_output\\': \\'with open(audio.rttm, w) as rttm:\\\\n  diarization.write_rttm(rttm)\\'}, \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'AISHELL-4\\', \\'accuracy\\': {\\'DER%\\': 14.61, \\'FA%\\': 3.31, \\'Miss%\\': 4.35, \\'Conf%\\': 6.95}}, {\\'name\\': \\'AMI Mix-Headset only_words\\', \\'accuracy\\': {\\'DER%\\': 18.21, \\'FA%\\': 3.28, \\'Miss%\\': 11.07, \\'Conf%\\': 3.87}}, {\\'name\\': \\'AMI Array1-01 only_words\\', \\'accuracy\\': {\\'DER%\\': 29.0, \\'FA%\\': 2.71, \\'Miss%\\': 21.61, \\'Conf%\\': 4.68}}, {\\'name\\': \\'CALLHOME Part2\\', \\'accuracy\\': {\\'DER%\\': 30.24, \\'FA%\\': 3.71, \\'Miss%\\': 16.86, \\'Conf%\\': 9.66}}, {\\'name\\': \\'DIHARD 3 Full\\', \\'accuracy\\': {\\'DER%\\': 20.99, \\'FA%\\': 4.25, \\'Miss%\\': 10.74, \\'Conf%\\': 6.0}}, {\\'name\\': \\'REPERE Phase 2\\', \\'accuracy\\': {\\'DER%\\': 12.62, \\'FA%\\': 1.55, \\'Miss%\\': 3.3, \\'Conf%\\': 7.76}}, {\\'name\\': \\'VoxConverse v0.0.2\\', \\'accuracy\\': {\\'DER%\\': 12.76, \\'FA%\\': 3.45, \\'Miss%\\': 3.85, \\'Conf%\\': 5.46}}]}, \\'description\\': \\'This API provides speaker diarization functionality using the pyannote.audio framework. It is capable of processing audio files and outputting speaker diarization results in RTTM format. The API supports providing the number of speakers, minimum and maximum number of speakers, and adjusting the segmentation onset threshold.\\'}', metadata={})]", "category": "generic"}
{"question_id": 162, "text": " We have a database of customer voices and are trying to build a voice recognition product so we can recognize customer voices when they call. How should we process and classify?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Voice Activity Detection\\', \\'api_name\\': \\'d4data/Indian-voice-cloning\\', \\'api_call\\': \"pipeline(\\'voice-activity-detection\\', model=\\'d4data/Indian-voice-cloning\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A model for detecting voice activity in Indian languages.\\'}', metadata={})]", "category": "generic"}
{"question_id": 163, "text": " We are developing a voice-controlled drone. Please identify the spoken command in the audio clip provided.\\n###Input: \\\"audio_clip.wav\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'CTranslate2\\', \\'functionality\\': \\'Automatic Speech Recognition\\', \\'api_name\\': \\'guillaumekln/faster-whisper-large-v2\\', \\'api_call\\': \"WhisperModel(\\'large-v2\\')\", \\'api_arguments\\': [\\'audio.mp3\\'], \\'python_environment_requirements\\': [\\'faster_whisper\\'], \\'example_code\\': \\'from faster_whisper import WhisperModel\\\\nmodel = WhisperModel(large-v2)\\\\nsegments, info = model.transcribe(audio.mp3)\\\\nfor segment in segments:\\\\n print([%.2fs -&gt; %.2fs] %s % (segment.start, segment.end, segment.text))\\', \\'performance\\': {\\'dataset\\': \\'99 languages\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Whisper large-v2 model for CTranslate2. This model can be used in CTranslate2 or projets based on CTranslate2 such as faster-whisper.\\'}', metadata={})]", "category": "generic"}
{"question_id": 164, "text": " Make a summary video for our last team meeting. The audio from the video must identify who is speaking and when.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'chavinlo/TempoFunk\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'chavinlo/TempoFunk\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Text-to-Video model using Hugging Face Transformers library. Model is capable of generating video content based on the input text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 165, "text": " I want to estimate the price of a house based on its features using this API. Please provide the code.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Joblib\\', \\'functionality\\': \\'Single Column Regression\\', \\'api_name\\': \\'jwan2021/autotrain-us-housing-prices-1771761512\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'joblib\\', \\'pandas\\', \\'json\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'jwan2021/autotrain-data-us-housing-prices\\', \\'accuracy\\': {\\'Loss\\': 122809.223, \\'R2\\': 0.884, \\'MSE\\': 15082105200.447, \\'MAE\\': 95586.887, \\'RMSLE\\': 0.13}}, \\'description\\': \\'A model trained using AutoTrain for predicting US housing prices with single column regression. The model is based on the jwan2021/autotrain-data-us-housing-prices dataset and has a CO2 Emissions of 50.5369 grams.\\'}', metadata={})]", "category": "generic"}
{"question_id": 166, "text": " Our company wants to predict housing prices in the US based on given features. Help us use the trained model to predict the prices.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Joblib\\', \\'functionality\\': \\'Single Column Regression\\', \\'api_name\\': \\'jwan2021/autotrain-us-housing-prices-1771761512\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'joblib\\', \\'pandas\\', \\'json\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'jwan2021/autotrain-data-us-housing-prices\\', \\'accuracy\\': {\\'Loss\\': 122809.223, \\'R2\\': 0.884, \\'MSE\\': 15082105200.447, \\'MAE\\': 95586.887, \\'RMSLE\\': 0.13}}, \\'description\\': \\'A model trained using AutoTrain for predicting US housing prices with single column regression. The model is based on the jwan2021/autotrain-data-us-housing-prices dataset and has a CO2 Emissions of 50.5369 grams.\\'}', metadata={})]", "category": "generic"}
{"question_id": 167, "text": " An environmental organization would like to use our Carbon Emissions prediction model to estimate CO2 emissions of different configurations of vehicles.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'45473113800\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'samvelkoch/autotrain-data-prknsn-2\\', \\'accuracy\\': {\\'Loss\\': 5.079, \\'R2\\': 0.109, \\'MSE\\': 25.795, \\'MAE\\': 3.78, \\'RMSLE\\': 0.849}}, \\'description\\': \\'A tabular regression model trained with AutoTrain for predicting carbon emissions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 168, "text": " I am a data analyst working in pollution detection, find a model and develop a piece of code for me for environment monitoring.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Joblib\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'tejas23/autotrain-amx2-1702259725\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': {\\'accuracy\\': 0.827}}, \\'description\\': \\'Multi-class Classification Model for Carbon Emissions\\'}', metadata={})]", "category": "generic"}
{"question_id": 169, "text": " We want to develop an intelligent prosthetic leg that can improve walking. Use a decision transformer to predict actions to be taken.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Reinforcement Learning\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'edbeeching/decision-transformer-gym-walker2d-expert\\', \\'api_call\\': \"AutoModel.from_pretrained(\\'edbeeching/decision-transformer-gym-walker2d-expert\\')\", \\'api_arguments\\': {\\'mean\\': [1.2384834, 0.19578537, -0.10475016, -0.18579608, 0.23003316, 0.022800924, -0.37383768, 0.337791, 3.925096, -0.0047428459, 0.025267061, -0.0039287535, -0.01736751, -0.48212224, 0.00035432147, -0.0037124525, 0.0026285544], \\'std\\': [0.06664903, 0.16980624, 0.17309439, 0.21843709, 0.74599105, 0.02410989, 0.3729872, 0.6226182, 0.9708009, 0.72936815, 1.504065, 2.495893, 3.511518, 5.3656907, 0.79503316, 4.317483, 6.1784487]}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'See our Blog Post, Colab notebook or Example Script for usage.\\', \\'performance\\': {\\'dataset\\': \\'Gym Walker2d environment\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Decision Transformer model trained on expert trajectories sampled from the Gym Walker2d environment.\\'}', metadata={})]", "category": "generic"}
{"question_id": 170, "text": " You want to create a bot that can play the Pong No Frameskip-v4 game with exceptional skill.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Reinforcement Learning\\', \\'framework\\': \\'Stable-Baselines3\\', \\'functionality\\': \\'deep-reinforcement-learning\\', \\'api_name\\': \\'ppo-PongNoFrameskip-v4\\', \\'api_call\\': \"load_from_hub(repo_id=\\'sb3/ppo-PongNoFrameskip-v4\\',filename=\\'{MODEL FILENAME}.zip\\',)\", \\'api_arguments\\': [\\'algo\\', \\'env\\', \\'f\\'], \\'python_environment_requirements\\': [\\'RL Zoo\\', \\'SB3\\', \\'SB3 Contrib\\'], \\'example_code\\': \\'python -m rl_zoo3.load_from_hub --algo ppo --env PongNoFrameskip-v4 -orga sb3 -f logs/\\', \\'performance\\': {\\'dataset\\': \\'PongNoFrameskip-v4\\', \\'accuracy\\': \\'21.00 +/- 0.00\\'}, \\'description\\': \\'This is a trained model of a PPO agent playing PongNoFrameskip-v4 using the stable-baselines3 library and the RL Zoo. The RL Zoo is a training framework for Stable Baselines3 reinforcement learning agents, with hyperparameter optimization and pre-trained agents included.\\'}', metadata={})]", "category": "generic"}
{"question_id": 171, "text": " I am a game developer working on a game project involving moving carts. I need to use reinforcement learning to improve the game experience.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Reinforcement Learning\\', \\'framework\\': \\'Stable-Baselines3\\', \\'functionality\\': \\'CartPole-v1\\', \\'api_name\\': \\'dqn-CartPole-v1\\', \\'api_call\\': \"load_from_hub(repo_id=\\'sb3/dqn-CartPole-v1\\',filename=\\'{MODEL FILENAME}.zip\\',)\", \\'api_arguments\\': [\\'algo\\', \\'env\\', \\'logs\\'], \\'python_environment_requirements\\': [\\'rl_zoo3\\', \\'stable-baselines3\\', \\'stable-baselines3-contrib\\'], \\'example_code\\': \\'python train.py --algo dqn --env CartPole-v1 -f logs/\\', \\'performance\\': {\\'dataset\\': \\'CartPole-v1\\', \\'accuracy\\': \\'500.00 +/- 0.00\\'}, \\'description\\': \\'This is a trained model of a DQN agent playing CartPole-v1 using the stable-baselines3 library and the RL Zoo. The RL Zoo is a training framework for Stable Baselines3 reinforcement learning agents, with hyperparameter optimization and pre-trained agents included.\\'}', metadata={})]", "category": "generic"}
{"question_id": 172, "text": " Develop a soccer playing agent that can outperform its opponents in a 2v2 environment.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Reinforcement Learning\\', \\'framework\\': \\'Unity ML-Agents Library\\', \\'functionality\\': \\'Train and play SoccerTwos\\', \\'api_name\\': \\'poca-SoccerTwosv2\\', \\'api_call\\': \"mlagents-load-from-hf --repo-id=\\'Raiden-1001/poca-SoccerTwosv2\\' --local-dir=\\'./downloads\\'\", \\'api_arguments\\': [\\'your_configuration_file_path.yaml\\', \\'run_id\\'], \\'python_environment_requirements\\': [\\'ml-agents\\'], \\'example_code\\': \\'mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume\\', \\'performance\\': {\\'dataset\\': \\'SoccerTwos\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'A trained model of a poca agent playing SoccerTwos using the Unity ML-Agents Library.\\'}', metadata={})]", "category": "generic"}
{"question_id": 173, "text": " We are tasked to analyze text for a Russian newspaper to help understand general sentiment and trends in the text.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'finiteautomata/beto-sentiment-analysis\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'finiteautomata/beto-sentiment-analysis\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'Hugging Face Transformers library\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'TASS 2020 corpus\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Model trained with TASS 2020 corpus (around ~5k tweets) of several dialects of Spanish. Base model is BETO, a BERT model trained in Spanish. Uses POS, NEG, NEU labels.\\'}', metadata={})]", "category": "generic"}
{"question_id": 174, "text": " We want to generate an image from a textual description for our PowerPoint presentation.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'Linaqruf/anything-v3.0\\', \\'api_call\\': \"Text2ImagePipeline(model=\\'Linaqruf/anything-v3.0\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A text-to-image model that generates images from text descriptions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 175, "text": " A new manga has been released and we would like to provide a manga reader app with translations. Get the text from the manga image.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Image-to-Text\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'kha-white/manga-ocr-base\\', \\'api_call\\': \"pipeline(\\'ocr\\', model=\\'kha-white/manga-ocr-base\\')\", \\'api_arguments\\': \\'image\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'manga109s\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Optical character recognition for Japanese text, with the main focus being Japanese manga. It uses Vision Encoder Decoder framework. Manga OCR can be used as a general purpose printed Japanese OCR, but its main goal was to provide a high quality text recognition, robust against various scenarios specific to manga: both vertical and horizontal text, text with furigana, text overlaid on images, wide variety of fonts and font styles, and low quality images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 176, "text": " We are building a social media site which creates automatic captions for users when they post a picture\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Image-to-Text\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Image Captioning\\', \\'api_name\\': \\'blip-image-captioning-large\\', \\'api_call\\': \\'BlipForConditionalGeneration.from_pretrained(Salesforce/blip-image-captioning-large)\\', \\'api_arguments\\': {\\'raw_image\\': \\'Image\\', \\'text\\': \\'Optional Text\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'BlipProcessor, BlipForConditionalGeneration\\', \\'PIL\\': \\'Image\\', \\'requests\\': \\'requests\\'}, \\'example_code\\': {\\'import_requests\\': \\'import requests\\', \\'import_PIL\\': \\'from PIL import Image\\', \\'import_transformers\\': \\'from transformers import BlipProcessor, BlipForConditionalGeneration\\', \\'load_processor\\': \\'processor = BlipProcessor.from_pretrained(Salesforce/blip-image-captioning-large)\\', \\'load_model\\': \\'model = BlipForConditionalGeneration.from_pretrained(Salesforce/blip-image-captioning-large)\\', \\'load_image\\': \"img_url = \\'https://storage.googleapis.com/sfr-vision-language-research/BLIP/demo.jpg\\'\\\\nraw_image = Image.open(requests.get(img_url, stream=True).raw).convert(\\'RGB\\')\", \\'conditional_captioning\\': \\'text = a photography of\\\\ninputs = processor(raw_image, text, return_tensors=pt)\\\\nout = model.generate(**inputs)\\\\nprint(processor.decode(out[0], skip_special_tokens=True))\\', \\'unconditional_captioning\\': \\'inputs = processor(raw_image, return_tensors=pt)\\\\nout = model.generate(**inputs)\\\\nprint(processor.decode(out[0], skip_special_tokens=True))\\'}, \\'performance\\': {\\'dataset\\': \\'COCO\\', \\'accuracy\\': {\\'image-text retrieval\\': \\'+2.7% recall@1\\', \\'image captioning\\': \\'+2.8% CIDEr\\', \\'VQA\\': \\'+1.6% VQA score\\'}}, \\'description\\': \\'BLIP is a Vision-Language Pre-training (VLP) framework that achieves state-of-the-art results on a wide range of vision-language tasks, such as image-text retrieval, image captioning, and VQA. It effectively utilizes noisy web data by bootstrapping the captions, where a captioner generates synthetic captions and a filter removes the noisy ones.\\'}', metadata={})]", "category": "generic"}
{"question_id": 177, "text": " There is robot in our factory which reads the image from the production line and then generate a text output based on the image.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'gsdf/Counterfeit-V2.5\\', \\'api_call\\': \"pipeline(\\'text-to-image\\', model=\\'gsdf/Counterfeit-V2.5\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'((masterpiece,best quality)),1girl, solo, animal ears, rabbit, barefoot, knees up, dress, sitting, rabbit ears, short sleeves, looking at viewer, grass, short hair, smile, white hair, puffy sleeves, outdoors, puffy short sleeves, bangs, on ground, full body, animal, white dress, sunlight, brown eyes, dappled sunlight, day, depth of field\\', \\'performance\\': {\\'dataset\\': \\'EasyNegative\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Counterfeit-V2.5 is a text-to-image model that generates anime-style images based on text prompts. It has been updated for ease of use and can be used with negative prompts to create high-quality images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 178, "text": " I am a filmmaker, and I need to make a short video based on a scene description from a script.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'chavinlo/TempoFunk\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'chavinlo/TempoFunk\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Text-to-Video model using Hugging Face Transformers library. Model is capable of generating video content based on the input text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 179, "text": " I'm an author and want to create a short video based on a brief passage from my book. Can you generate a video based on this text?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'chavinlo/TempoFunk\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'chavinlo/TempoFunk\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Text-to-Video model using Hugging Face Transformers library. Model is capable of generating video content based on the input text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 180, "text": " I want to build an AI model that can analyze images and answer questions about the content of the image.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 181, "text": " We are trying to develop an application that helps tourists get information about attractions by analyzing images they take and responding to questions.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'prompthero/openjourney-v4\\', \\'api_call\\': \"pipeline(\\'text-to-image\\', model=\\'prompthero/openjourney-v4\\')\", \\'api_arguments\\': {\\'text\\': \\'string\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"generate_image(\\'your text here\\')\", \\'performance\\': {\\'dataset\\': \\'Midjourney v4 images\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Openjourney v4 is trained on +124k Midjourney v4 images by PromptHero. It is used for generating images based on text inputs.\\'}', metadata={})]", "category": "generic"}
{"question_id": 182, "text": " We have the financial documents of a company and we want to extract information about the cash flow. Modify the model so it can answer the questions related to the cash flow.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 183, "text": " Please generate a correct building plan leveraging the data given.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Joblib\\', \\'functionality\\': \\'Single Column Regression\\', \\'api_name\\': \\'jwan2021/autotrain-us-housing-prices-1771761512\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'joblib\\', \\'pandas\\', \\'json\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'jwan2021/autotrain-data-us-housing-prices\\', \\'accuracy\\': {\\'Loss\\': 122809.223, \\'R2\\': 0.884, \\'MSE\\': 15082105200.447, \\'MAE\\': 95586.887, \\'RMSLE\\': 0.13}}, \\'description\\': \\'A model trained using AutoTrain for predicting US housing prices with single column regression. The model is based on the jwan2021/autotrain-data-us-housing-prices dataset and has a CO2 Emissions of 50.5369 grams.\\'}', metadata={})]", "category": "generic"}
{"question_id": 184, "text": " Help our drone video analyzing app estimate the depth in drone footage.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Depth Estimation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Depth Estimation\\', \\'api_name\\': \\'glpn-kitti\\', \\'api_call\\': \"GLPNForDepthEstimation.from_pretrained(\\'vinvino02/glpn-kitti\\')\", \\'api_arguments\\': \\'images, return_tensors\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'from transformers import GLPNFeatureExtractor, GLPNForDepthEstimation\\\\nimport torch\\\\nimport numpy as np\\\\nfrom PIL import Image\\\\nimport requests\\\\nurl = http://images.cocodataset.org/val2017/000000039769.jpg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nfeature_extractor = GLPNFeatureExtractor.from_pretrained(vinvino02/glpn-kitti)\\\\nmodel = GLPNForDepthEstimation.from_pretrained(vinvino02/glpn-kitti)\\\\ninputs = feature_extractor(images=image, return_tensors=pt)\\\\nwith torch.no_grad():\\\\n outputs = model(**inputs)\\\\n predicted_depth = outputs.predicted_depth\\\\nprediction = torch.nn.functional.interpolate(\\\\n predicted_depth.unsqueeze(1),\\\\n size=image.size[::-1],\\\\n mode=bicubic,\\\\n align_corners=False,\\\\n)\\\\noutput = prediction.squeeze().cpu().numpy()\\\\nformatted = (output * 255 / np.max(output)).astype(uint8)\\\\ndepth = Image.fromarray(formatted)\\', \\'performance\\': {\\'dataset\\': \\'KITTI\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Global-Local Path Networks (GLPN) model trained on KITTI for monocular depth estimation. It was introduced in the paper Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth by Kim et al. and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 185, "text": " As a salesperson, I need to analyze customer invoices to answer questions about the total amount, tax, and due date from an image file.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'layoutlm-invoices\\', \\'api_call\\': \"AutoModelForDocumentQuestionAnswering.from_pretrained(\\'impira/layoutlm-invoices\\')\", \\'api_arguments\\': \\'question, context\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"nlp(question=\\'What is the total amount?\\', context=\\'your_invoice_text\\')\", \\'performance\\': {\\'dataset\\': \\'proprietary dataset of invoices, SQuAD2.0, and DocVQA\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'A fine-tuned version of the multi-modal LayoutLM model for the task of question answering on invoices and other documents. It has been fine-tuned on a proprietary dataset of invoices as well as both SQuAD2.0 and DocVQA for general comprehension. Unlike other QA models, which can only extract consecutive tokens, this model can predict longer-range, non-consecutive sequences with an additional classifier head.\\'}', metadata={})]", "category": "generic"}
{"question_id": 186, "text": " We need to find out the depth information of a room for monitoring purposes.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Depth Estimation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Depth Estimation\\', \\'api_name\\': \\'glpn-nyu-finetuned-diode\\', \\'api_call\\': \"pipeline(\\'depth-estimation\\', model=\\'sayakpaul/glpn-nyu-finetuned-diode\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'diode-subset\\', \\'accuracy\\': {\\'Loss\\': 0.4359, \\'Rmse\\': 0.4276}}, \\'description\\': \\'This model is a fine-tuned version of vinvino02/glpn-nyu on the diode-subset dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 187, "text": " We are creating an autonomous car and need to estimate the depth of objects in a given scene.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Depth Estimation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Depth Estimation\\', \\'api_name\\': \\'glpn-kitti\\', \\'api_call\\': \"GLPNForDepthEstimation.from_pretrained(\\'vinvino02/glpn-kitti\\')\", \\'api_arguments\\': \\'images, return_tensors\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'from transformers import GLPNFeatureExtractor, GLPNForDepthEstimation\\\\nimport torch\\\\nimport numpy as np\\\\nfrom PIL import Image\\\\nimport requests\\\\nurl = http://images.cocodataset.org/val2017/000000039769.jpg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nfeature_extractor = GLPNFeatureExtractor.from_pretrained(vinvino02/glpn-kitti)\\\\nmodel = GLPNForDepthEstimation.from_pretrained(vinvino02/glpn-kitti)\\\\ninputs = feature_extractor(images=image, return_tensors=pt)\\\\nwith torch.no_grad():\\\\n outputs = model(**inputs)\\\\n predicted_depth = outputs.predicted_depth\\\\nprediction = torch.nn.functional.interpolate(\\\\n predicted_depth.unsqueeze(1),\\\\n size=image.size[::-1],\\\\n mode=bicubic,\\\\n align_corners=False,\\\\n)\\\\noutput = prediction.squeeze().cpu().numpy()\\\\nformatted = (output * 255 / np.max(output)).astype(uint8)\\\\ndepth = Image.fromarray(formatted)\\', \\'performance\\': {\\'dataset\\': \\'KITTI\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Global-Local Path Networks (GLPN) model trained on KITTI for monocular depth estimation. It was introduced in the paper Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth by Kim et al. and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 188, "text": " Help us create an AI solution to automatically label images taken by a security camera.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'abhishek/autotrain-dog-vs-food\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'abhishek/autotrain-dog-vs-food\\')\", \\'api_arguments\\': \\'image_path\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'sasha/dog-food\\', \\'accuracy\\': 0.998}, \\'description\\': \"A pre-trained model for classifying images as either dog or food using Hugging Face\\'s AutoTrain framework.\"}', metadata={})]", "category": "generic"}
{"question_id": 189, "text": " Develop a software to classify an image from a URL into a thousand categories.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image-to-Image\\', \\'framework\\': \\'Keras\\', \\'functionality\\': \\'Image Deblurring\\', \\'api_name\\': \\'google/maxim-s3-deblurring-gopro\\', \\'api_call\\': \"from_pretrained_keras(\\'google/maxim-s3-deblurring-gopro\\')\", \\'api_arguments\\': [\\'image\\'], \\'python_environment_requirements\\': [\\'huggingface_hub\\', \\'PIL\\', \\'tensorflow\\', \\'numpy\\', \\'requests\\'], \\'example_code\\': \\'from huggingface_hub import from_pretrained_keras\\\\nfrom PIL import Image\\\\nimport tensorflow as tf\\\\nimport numpy as np\\\\nimport requests\\\\nurl = https://github.com/sayakpaul/maxim-tf/raw/main/images/Deblurring/input/1fromGOPR0950.png\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nimage = np.array(image)\\\\nimage = tf.convert_to_tensor(image)\\\\nimage = tf.image.resize(image, (256, 256))\\\\nmodel = from_pretrained_keras(google/maxim-s3-deblurring-gopro)\\\\npredictions = model.predict(tf.expand_dims(image, 0))\\', \\'performance\\': {\\'dataset\\': \\'GoPro\\', \\'accuracy\\': {\\'PSNR\\': 32.86, \\'SSIM\\': 0.961}}, \\'description\\': \\'MAXIM model pre-trained for image deblurring. It was introduced in the paper MAXIM: Multi-Axis MLP for Image Processing by Zhengzhong Tu, Hossein Talebi, Han Zhang, Feng Yang, Peyman Milanfar, Alan Bovik, Yinxiao Li and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 190, "text": " Our delivery drones need to detect and avoid obstacles while flying. Develop a solution for them to detect objects in their path.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Voice Activity Detection\\', \\'api_name\\': \\'d4data/Indian-voice-cloning\\', \\'api_call\\': \"pipeline(\\'voice-activity-detection\\', model=\\'d4data/Indian-voice-cloning\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A model for detecting voice activity in Indian languages.\\'}', metadata={})]", "category": "generic"}
{"question_id": 191, "text": " Develop a code to recognize objects in images using deformable-detr model.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Object Detection\\', \\'api_name\\': \\'deformable-detr\\', \\'api_call\\': \"DeformableDetrForObjectDetection.from_pretrained(\\'SenseTime/deformable-detr\\')\", \\'api_arguments\\': [\\'images\\', \\'return_tensors\\'], \\'python_environment_requirements\\': [\\'transformers\\', \\'torch\\', \\'PIL\\', \\'requests\\'], \\'example_code\\': \"from transformers import AutoImageProcessor, DeformableDetrForObjectDetection\\\\nimport torch\\\\nfrom PIL import Image\\\\nimport requests\\\\nurl = \\'http://images.cocodataset.org/val2017/000000039769.jpg\\'\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nprocessor = AutoImageProcessor.from_pretrained(\\'SenseTime/deformable-detr\\')\\\\nmodel = DeformableDetrForObjectDetection.from_pretrained(\\'SenseTime/deformable-detr\\')\\\\ninputs = processor(images=image, return_tensors=\\'pt\\')\\\\noutputs = model(**inputs)\", \\'performance\\': {\\'dataset\\': \\'COCO 2017\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Deformable DETR model with ResNet-50 backbone trained end-to-end on COCO 2017 object detection (118k annotated images). It was introduced in the paper Deformable DETR: Deformable Transformers for End-to-End Object Detection by Zhu et al. and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 192, "text": " I need to extract tables from a set of scanned document images to simplify data analysis.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Table Extraction\\', \\'api_name\\': \\'keremberke/yolov8n-table-extraction\\', \\'api_call\\': \"YOLO(\\'keremberke/yolov8n-table-extraction\\')\", \\'api_arguments\\': {\\'conf\\': 0.25, \\'iou\\': 0.45, \\'agnostic_nms\\': False, \\'max_det\\': 1000}, \\'python_environment_requirements\\': [\\'ultralyticsplus==0.0.23\\', \\'ultralytics==8.0.21\\'], \\'example_code\\': [\\'from ultralyticsplus import YOLO, render_result\\', \"model = YOLO(\\'keremberke/yolov8n-table-extraction\\')\", \"model.overrides[\\'conf\\'] = 0.25\", \"model.overrides[\\'iou\\'] = 0.45\", \"model.overrides[\\'agnostic_nms\\'] = False\", \"model.overrides[\\'max_det\\'] = 1000\", \"image = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\", \\'results = model.predict(image)\\', \\'print(results[0].boxes)\\', \\'render = render_result(model=model, image=image, result=results[0])\\', \\'render.show()\\'], \\'performance\\': {\\'dataset\\': \\'table-extraction\\', \\'accuracy\\': 0.967}, \\'description\\': \"An object detection model for extracting tables from documents. Supports two label types: \\'bordered\\' and \\'borderless\\'.\"}', metadata={})]", "category": "generic"}
{"question_id": 193, "text": " Our customers wish to automatically detect shoplifters in the store using a surveillance camera. Help them to implement object detection and identify potential thieves.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Object Detection\\', \\'api_name\\': \\'keremberke/yolov8m-forklift-detection\\', \\'api_call\\': \"YOLO(\\'keremberke/yolov8m-forklift-detection\\')\", \\'api_arguments\\': {\\'image\\': \\'URL or local path to the image\\'}, \\'python_environment_requirements\\': [\\'ultralyticsplus==0.0.23\\', \\'ultralytics==8.0.21\\'], \\'example_code\\': [\\'from ultralyticsplus import YOLO, render_result\\', \"model = YOLO(\\'keremberke/yolov8m-forklift-detection\\')\", \"model.overrides[\\'conf\\'] = 0.25\", \"model.overrides[\\'iou\\'] = 0.45\", \"model.overrides[\\'agnostic_nms\\'] = False\", \"model.overrides[\\'max_det\\'] = 1000\", \"image = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\", \\'results = model.predict(image)\\', \\'print(results[0].boxes)\\', \\'render = render_result(model=model, image=image, result=results[0])\\', \\'render.show()\\'], \\'performance\\': {\\'dataset\\': \\'forklift-object-detection\\', \\'accuracy\\': 0.846}, \\'description\\': \\'A YOLOv8 model for detecting forklifts and persons in images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 194, "text": " Create an object detector that can detect blood cells in an image, such as platelets, red blood cells, and white blood cells.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Blood Cell Detection\\', \\'api_name\\': \\'keremberke/yolov8n-blood-cell-detection\\', \\'api_call\\': \"YOLO(\\'keremberke/yolov8n-blood-cell-detection\\')\", \\'api_arguments\\': {\\'conf\\': 0.25, \\'iou\\': 0.45, \\'agnostic_nms\\': False, \\'max_det\\': 1000}, \\'python_environment_requirements\\': \\'ultralyticsplus==0.0.23 ultralytics==8.0.21\\', \\'example_code\\': \"from ultralyticsplus import YOLO, render_result\\\\nmodel = YOLO(\\'keremberke/yolov8n-blood-cell-detection\\')\\\\nmodel.overrides[\\'conf\\'] = 0.25\\\\nmodel.overrides[\\'iou\\'] = 0.45\\\\nmodel.overrides[\\'agnostic_nms\\'] = False\\\\nmodel.overrides[\\'max_det\\'] = 1000\\\\nimage = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\\\\nresults = model.predict(image)\\\\nprint(results[0].boxes)\\\\nrender = render_result(model=model, image=image, result=results[0])\\\\nrender.show()\", \\'performance\\': {\\'dataset\\': \\'blood-cell-object-detection\\', \\'accuracy\\': 0.893}, \\'description\\': \\'This model detects blood cells in images, specifically Platelets, RBC, and WBC. It is based on the YOLOv8 architecture and trained on the blood-cell-object-detection dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 195, "text": " I am a real-estate agent working on a project where I need to convert images of room plans to a better visual representation.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'Realistic_Vision_V1.4\\', \\'api_call\\': \"pipeline(\\'text-to-image\\', model=SG161222/Realistic_Vision_V1.4)\", \\'api_arguments\\': {\\'prompt\\': \\'string\\', \\'negative_prompt\\': \\'string\\'}, \\'python_environment_requirements\\': [\\'transformers\\', \\'torch\\'], \\'example_code\\': \"from transformers import pipeline\\\\n\\\\nmodel = pipeline(\\'text-to-image\\', model=\\'SG161222/Realistic_Vision_V1.4\\')\\\\n\\\\nprompt = \\'a close up portrait photo of 26 y.o woman in wastelander clothes, long haircut, pale skin, slim body, background is city ruins, (high detailed skin:1.2), 8k uhd, dslr, soft lighting, high quality, film grain, Fujifilm XT3\\'\\\\nnegative_prompt = \\'(deformed iris, deformed pupils, semi-realistic, cgi, 3d, render, sketch, cartoon, drawing, anime:1.4), text, close up, cropped, out of frame, worst quality, low quality, jpeg artifacts, ugly, duplicate, morbid, mutilated, extra fingers, mutated hands, poorly drawn hands, poorly drawn face, mutation, deformed, blurry, dehydrated, bad anatomy, bad proportions, extra limbs, cloned face, disfigured, gross proportions, malformed limbs, missing arms, missing legs, extra arms, extra legs, fused fingers, too many fingers, long neck\\'\\\\n\\\\nresult = model(prompt, negative_prompt=negative_prompt)\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'Realistic_Vision_V1.4 is a text-to-image model that generates high-quality and detailed images based on textual prompts. It can be used for various applications such as generating realistic portraits, landscapes, and other types of images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 196, "text": " We recently received low resolution images of newly released products and need to upscale them for better quality.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Image Upscaling\\', \\'api_name\\': \\'stabilityai/sd-x2-latent-upscaler\\', \\'api_call\\': \\'StableDiffusionLatentUpscalePipeline.from_pretrained(stabilityai/sd-x2-latent-upscaler, torch_dtype=torch.float16)\\', \\'api_arguments\\': {\\'prompt\\': \\'text prompt\\', \\'image\\': \\'low resolution latents\\', \\'num_inference_steps\\': 20, \\'guidance_scale\\': 0, \\'generator\\': \\'torch generator\\'}, \\'python_environment_requirements\\': [\\'git+https://github.com/huggingface/diffusers.git\\', \\'transformers\\', \\'accelerate\\', \\'scipy\\', \\'safetensors\\'], \\'example_code\\': \\'from diffusers import StableDiffusionLatentUpscalePipeline, StableDiffusionPipeline\\\\nimport torch\\\\npipeline = StableDiffusionPipeline.from_pretrained(CompVis/stable-diffusion-v1-4, torch_dtype=torch.float16)\\\\npipeline.to(cuda)\\\\nupscaler = StableDiffusionLatentUpscalePipeline.from_pretrained(stabilityai/sd-x2-latent-upscaler, torch_dtype=torch.float16)\\\\nupscaler.to(cuda)\\\\nprompt = a photo of an astronaut high resolution, unreal engine, ultra realistic\\\\ngenerator = torch.manual_seed(33)\\\\nlow_res_latents = pipeline(prompt, generator=generator, output_type=latent).images\\\\nupscaled_image = upscaler(prompt=prompt, image=low_res_latents, num_inference_steps=20, guidance_scale=0, generator=generator).images[0]\\\\nupscaled_image.save(astronaut_1024.png)\\', \\'performance\\': {\\'dataset\\': \\'LAION-2B\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \"Stable Diffusion x2 latent upscaler is a diffusion-based upscaler model developed by Katherine Crowson in collaboration with Stability AI. It is designed to upscale Stable Diffusion\\'s latent denoised image embeddings, allowing for fast text-to-image and upscaling pipelines. The model was trained on a high-resolution subset of the LAION-2B dataset and works with all Stable Diffusion checkpoints.\"}', metadata={})]", "category": "generic"}
{"question_id": 197, "text": " As a toy company, we are designing a new toy line. We'd like you to create an image of a toy robot using relevant text prompts as control input.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'prompthero/openjourney-v4\\', \\'api_call\\': \"pipeline(\\'text-to-image\\', model=\\'prompthero/openjourney-v4\\')\", \\'api_arguments\\': {\\'text\\': \\'string\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"generate_image(\\'your text here\\')\", \\'performance\\': {\\'dataset\\': \\'Midjourney v4 images\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Openjourney v4 is trained on +124k Midjourney v4 images by PromptHero. It is used for generating images based on text inputs.\\'}', metadata={})]", "category": "generic"}
{"question_id": 198, "text": " I'm developing a game that needs more Minecraft skins for some characters. How can I generete them with this model?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Unconditional Image Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Diffusers\\', \\'api_name\\': \\'Minecraft-Skin-Diffusion\\', \\'api_call\\': \"DDPMPipeline.from_pretrained(\\'WiNE-iNEFF/Minecraft-Skin-Diffusion\\')\", \\'api_arguments\\': {}, \\'python_environment_requirements\\': [\\'diffusers\\'], \\'example_code\\': \"from diffusers import DDPMPipeline\\\\npipeline = DDPMPipeline.from_pretrained(\\'WiNE-iNEFF/Minecraft-Skin-Diffusion\\')\\\\nimage = pipeline().images[0].convert(\\'RGBA\\')\\\\nimage\", \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Unconditional Image Generation model for generating Minecraft skins using diffusion-based methods.\\'}', metadata={})]", "category": "generic"}
{"question_id": 199, "text": " Find me a generative model to create cat images in 256x256 resolution.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Unconditional Image Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Unconditional Image Generation\\', \\'api_name\\': \\'google/ddpm-cat-256\\', \\'api_call\\': \"DDPMPipeline.from_pretrained(\\'google/ddpm-cat-256\\')\", \\'api_arguments\\': [\\'model_id\\'], \\'python_environment_requirements\\': [\\'diffusers\\'], \\'example_code\\': \\'!pip install diffusers\\\\nfrom diffusers import DDPMPipeline, DDIMPipeline, PNDMPipeline\\\\nmodel_id = google/ddpm-cat-256\\\\nddpm = DDPMPipeline.from_pretrained(model_id)\\\\nimage = ddpm().images[0]\\\\nimage.save(ddpm_generated_image.png)\\', \\'performance\\': {\\'dataset\\': \\'CIFAR10\\', \\'accuracy\\': {\\'Inception_score\\': 9.46, \\'FID_score\\': 3.17}}, \\'description\\': \\'Denoising Diffusion Probabilistic Models (DDPM) is a class of latent variable models inspired by considerations from nonequilibrium thermodynamics. It can generate high-quality images using discrete noise schedulers such as scheduling_ddpm, scheduling_ddim, and scheduling_pndm. The model is trained on the unconditional CIFAR10 dataset and 256x256 LSUN, obtaining an Inception score of 9.46 and a state-of-the-art FID score of 3.17.\\'}', metadata={})]", "category": "generic"}
{"question_id": 200, "text": " Our organization works with video surveillance. We need a system to analyze the videos and classify various events happening inside the video.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Video Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Video Classification\\', \\'api_name\\': \\'facebook/timesformer-base-finetuned-ssv2\\', \\'api_call\\': \"TimesformerForVideoClassification.from_pretrained(\\'facebook/timesformer-base-finetuned-ssv2\\')\", \\'api_arguments\\': [\\'images\\', \\'return_tensors\\'], \\'python_environment_requirements\\': [\\'transformers\\', \\'numpy\\', \\'torch\\'], \\'example_code\\': \\'from transformers import AutoImageProcessor, TimesformerForVideoClassification\\\\nimport numpy as np\\\\nimport torch\\\\nvideo = list(np.random.randn(8, 3, 224, 224))\\\\nprocessor = AutoImageProcessor.from_pretrained(facebook/timesformer-base-finetuned-ssv2)\\\\nmodel = TimesformerForVideoClassification.from_pretrained(facebook/timesformer-base-finetuned-ssv2)\\\\ninputs = processor(images=video, return_tensors=pt)\\\\nwith torch.no_grad():\\\\n outputs = model(**inputs)\\\\n logits = outputs.logits\\\\npredicted_class_idx = logits.argmax(-1).item()\\\\nprint(Predicted class:, model.config.id2label[predicted_class_idx])\\', \\'performance\\': {\\'dataset\\': \\'Something Something v2\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'TimeSformer model pre-trained on Something Something v2. It was introduced in the paper TimeSformer: Is Space-Time Attention All You Need for Video Understanding? by Tong et al. and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 201, "text": " People in my company need an automatic solution to categorize videos based on their content. The system should be able to recognize the main theme of a video with high accuracy.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'chavinlo/TempoFunk\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'chavinlo/TempoFunk\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Text-to-Video model using Hugging Face Transformers library. Model is capable of generating video content based on the input text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 202, "text": " We want to build a product to classify images of pets into different categories.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'abhishek/autotrain-dog-vs-food\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'abhishek/autotrain-dog-vs-food\\')\", \\'api_arguments\\': \\'image_path\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'sasha/dog-food\\', \\'accuracy\\': 0.998}, \\'description\\': \"A pre-trained model for classifying images as either dog or food using Hugging Face\\'s AutoTrain framework.\"}', metadata={})]", "category": "generic"}
{"question_id": 203, "text": " We are developing an app to classify food images. We have a set of images and want to use a pre-trained model for classification.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'abhishek/autotrain-dog-vs-food\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'abhishek/autotrain-dog-vs-food\\')\", \\'api_arguments\\': \\'image_path\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'sasha/dog-food\\', \\'accuracy\\': 0.998}, \\'description\\': \"A pre-trained model for classifying images as either dog or food using Hugging Face\\'s AutoTrain framework.\"}', metadata={})]", "category": "generic"}
{"question_id": 205, "text": " The company wants to use geolocalization techniques to identify the location of a given image.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Zero-Shot Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Geolocalization\\', \\'api_name\\': \\'geolocal/StreetCLIP\\', \\'api_call\\': \"CLIPModel.from_pretrained(\\'geolocal/StreetCLIP\\')\", \\'api_arguments\\': {\\'pretrained_model_name_or_path\\': \\'geolocal/StreetCLIP\\'}, \\'python_environment_requirements\\': [\\'transformers\\', \\'PIL\\', \\'requests\\'], \\'example_code\\': \\'from PIL import Image\\\\nimport requests\\\\nfrom transformers import CLIPProcessor, CLIPModel\\\\nmodel = CLIPModel.from_pretrained(geolocal/StreetCLIP)\\\\nprocessor = CLIPProcessor.from_pretrained(geolocal/StreetCLIP)\\\\nurl = https://huggingface.co/geolocal/StreetCLIP/resolve/main/sanfrancisco.jpeg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nchoices = [San Jose, San Diego, Los Angeles, Las Vegas, San Francisco]\\\\ninputs = processor(text=choices, images=image, return_tensors=pt, padding=True)\\\\noutputs = model(**inputs)\\\\nlogits_per_image = outputs.logits_per_image\\\\nprobs = logits_per_image.softmax(dim=1)\\', \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'IM2GPS\\', \\'accuracy\\': {\\'25km\\': 28.3, \\'200km\\': 45.1, \\'750km\\': 74.7, \\'2500km\\': 88.2}}, {\\'name\\': \\'IM2GPS3K\\', \\'accuracy\\': {\\'25km\\': 22.4, \\'200km\\': 37.4, \\'750km\\': 61.3, \\'2500km\\': 80.4}}]}, \\'description\\': \\'StreetCLIP is a robust foundation model for open-domain image geolocalization and other geographic and climate-related tasks. Trained on an original dataset of 1.1 million street-level urban and rural geo-tagged images, it achieves state-of-the-art performance on multiple open-domain image geolocalization benchmarks in zero-shot, outperforming supervised models trained on millions of images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 206, "text": " To improve customer service, the company is developing an AI-based chatbot that can help generate multiple paraphrases of a given customer query. We want to know if the generated paraphrases are adequate.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Paraphrasing\\', \\'api_name\\': \\'prithivida/parrot_paraphraser_on_T5\\', \\'api_call\\': \"Parrot(model_tag=\\'prithivida/parrot_paraphraser_on_T5\\', use_gpu=False)\", \\'api_arguments\\': [\\'input_phrase\\', \\'diversity_ranker\\', \\'do_diverse\\', \\'max_return_phrases\\', \\'max_length\\', \\'adequacy_threshold\\', \\'fluency_threshold\\'], \\'python_environment_requirements\\': [\\'torch\\', \\'transformers\\'], \\'example_code\\': \\'from parrot import Parrot\\\\nimport torch\\\\nimport warnings\\\\nwarnings.filterwarnings(ignore)\\\\n\\\\nparrot = Parrot(model_tag=prithivida/parrot_paraphraser_on_T5, use_gpu=False)\\\\nphrases = [Can you recommed some upscale restaurants in Newyork?,\\\\n What are the famous places we should not miss in Russia?\\\\n]\\\\nfor phrase in phrases:\\\\n print(-*100)\\\\n print(Input_phrase: , phrase)\\\\n print(-*100)\\\\n para_phrases = parrot.augment(input_phrase=phrase)\\\\n for para_phrase in para_phrases:\\\\n  print(para_phrase)\\', \\'performance\\': {\\'dataset\\': \\'Not mentioned\\', \\'accuracy\\': \\'Not mentioned\\'}, \\'description\\': \\'Parrot is a paraphrase based utterance augmentation framework purpose built to accelerate training NLU models. It offers knobs to control Adequacy, Fluency, and Diversity as per your needs. It mainly focuses on augmenting texts typed-into or spoken-to conversational interfaces for building robust NLU models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 207, "text": " We need to analyze user reviews for our app to determine whether they are positive, negative, or neutral.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'michellejieli/NSFW_text_classifier\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'michellejieli/NSFW_text_classification\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'classifier(I see you\u2019ve set aside this special time to humiliate yourself in public.)\\', \\'performance\\': {\\'dataset\\': \\'Reddit posts\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'DistilBERT is a transformer model that performs sentiment analysis. I fine-tuned the model on Reddit posts with the purpose of classifying not safe for work (NSFW) content, specifically text that is considered inappropriate and unprofessional. The model predicts 2 classes, which are NSFW or safe for work (SFW). The model is a fine-tuned version of DistilBERT. It was fine-tuned on 14317 Reddit posts pulled from the Reddit API.\\'}', metadata={})]", "category": "generic"}
{"question_id": 208, "text": " Give me a tool to detect named entities in multiple languages for my news portal analysis.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Token Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Named Entity Recognition\\', \\'api_name\\': \\'Davlan/bert-base-multilingual-cased-ner-hrl\\', \\'api_call\\': \"AutoModelForTokenClassification.from_pretrained(\\'Davlan/bert-base-multilingual-cased-ner-hrl\\')\", \\'api_arguments\\': {\\'model\\': \\'AutoModelForTokenClassification.from_pretrained(Davlan/bert-base-multilingual-cased-ner-hrl)\\', \\'tokenizer\\': \\'AutoTokenizer.from_pretrained(Davlan/bert-base-multilingual-cased-ner-hrl)\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'from transformers import AutoTokenizer, AutoModelForTokenClassification, pipeline\\'}, \\'example_code\\': \\'example = Nader Jokhadar had given Syria the lead with a well-struck header in the seventh minute.; ner_results = nlp(example); print(ner_results)\\', \\'performance\\': {\\'dataset\\': {\\'Arabic\\': \\'ANERcorp\\', \\'German\\': \\'conll 2003\\', \\'English\\': \\'conll 2003\\', \\'Spanish\\': \\'conll 2002\\', \\'French\\': \\'Europeana Newspapers\\', \\'Italian\\': \\'Italian I-CAB\\', \\'Latvian\\': \\'Latvian NER\\', \\'Dutch\\': \\'conll 2002\\', \\'Portuguese\\': \\'Paramopama + Second Harem\\', \\'Chinese\\': \\'MSRA\\'}, \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'bert-base-multilingual-cased-ner-hrl is a Named Entity Recognition model for 10 high resourced languages (Arabic, German, English, Spanish, French, Italian, Latvian, Dutch, Portuguese and Chinese) based on a fine-tuned mBERT base model. It has been trained to recognize three types of entities: location (LOC), organizations (ORG), and person (PER).\\'}', metadata={})]", "category": "generic"}
{"question_id": 209, "text": " We are building a fiction-sharing platform for multiple languages. Extract named entities from the given text.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'PyTorch Transformers\\', \\'functionality\\': \\'text2text-generation\\', \\'api_name\\': \\'facebook/nllb-200-distilled-600M\\', \\'api_call\\': \"pipeline(\\'translation_xx_to_yy\\', model=\\'facebook/nllb-200-distilled-600M\\')\", \\'api_arguments\\': [\\'model\\', \\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\', \\'torch\\'], \\'example_code\\': \"from transformers import pipeline; translator = pipeline(\\'translation_xx_to_yy\\', model=\\'facebook/nllb-200-distilled-600M\\'); translator(\\'Hello World\\')\", \\'performance\\': {\\'dataset\\': \\'Flores-200\\', \\'accuracy\\': \\'BLEU, spBLEU, chrF++\\'}, \\'description\\': \\'NLLB-200 is a machine translation model primarily intended for research in machine translation, especially for low-resource languages. It allows for single sentence translation among 200 languages. The model was trained on general domain text data and is not intended to be used with domain specific texts, such as medical domain or legal domain. The model is not intended to be used for document translation.\\'}', metadata={})]", "category": "generic"}
{"question_id": 210, "text": " Someone asked the question \\\"What is the capital of Sweden?\\\" in the context \\\"Stockholm is the beautiful capital of Sweden, which is known for its high living standards and great attractions.\\\"\\n###Input: {\\\"question\\\": \\\"What is the capital of Sweden?\\\", \\\"context\\\": \\\"Stockholm is the beautiful capital of Sweden, which is known for its high living standards and great attractions.\\\"}\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Question Answering\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'deepset/roberta-large-squad2\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=\\'deepset/roberta-large-squad2\\')\", \\'api_arguments\\': [\\'question\\', \\'context\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline; nlp = pipeline(\\'question-answering\\', model=\\'deepset/roberta-large-squad2\\'); nlp({\\'question\\': \\'What is the capital of Germany?\\', \\'context\\': \\'Berlin is the capital of Germany.\\'})\", \\'performance\\': {\\'dataset\\': \\'squad_v2\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'A pre-trained RoBERTa model for question answering tasks, specifically trained on the SQuAD v2 dataset. It can be used to answer questions based on a given context.\\'}', metadata={})]", "category": "generic"}
{"question_id": 211, "text": " Write an overview for astronomers about how Jupiter became the largest planet in our solar system.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'google/pegasus-large\\', \\'api_call\\': \"pipeline(\\'summarization\\', model=\\'google/pegasus-large\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"from transformers import pipeline\\\\nsummarizer = pipeline(\\'summarization\\', model=\\'google/pegasus-large\\')\\\\nsummary = summarizer(\\'your_text_here\\')\", \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'xsum\\', \\'accuracy\\': \\'47.60/24.83/39.64\\'}, {\\'name\\': \\'cnn_dailymail\\', \\'accuracy\\': \\'44.16/21.56/41.30\\'}, {\\'name\\': \\'newsroom\\', \\'accuracy\\': \\'45.98/34.20/42.18\\'}, {\\'name\\': \\'multi_news\\', \\'accuracy\\': \\'47.65/18.75/24.95\\'}, {\\'name\\': \\'gigaword\\', \\'accuracy\\': \\'39.65/20.47/36.76\\'}, {\\'name\\': \\'wikihow\\', \\'accuracy\\': \\'46.39/22.12/38.41\\'}, {\\'name\\': \\'reddit_tifu\\', \\'accuracy\\': \\'27.99/9.81/22.94\\'}, {\\'name\\': \\'big_patent\\', \\'accuracy\\': \\'52.29/33.08/41.66\\'}, {\\'name\\': \\'arxiv\\', \\'accuracy\\': \\'44.21/16.95/25.67\\'}, {\\'name\\': \\'pubmed\\', \\'accuracy\\': \\'45.97/20.15/28.25\\'}, {\\'name\\': \\'aeslc\\', \\'accuracy\\': \\'37.68/21.25/36.51\\'}, {\\'name\\': \\'billsum\\', \\'accuracy\\': \\'59.67/41.58/47.59\\'}]}, \\'description\\': \\'google/pegasus-large is a pre-trained model for abstractive text summarization based on the PEGASUS architecture. It is trained on a mixture of C4 and HugeNews datasets and uses a sentencepiece tokenizer that can encode newline characters. The model has been fine-tuned for various summarization tasks and achieves state-of-the-art performance on multiple benchmarks.\\'}', metadata={})]", "category": "generic"}
{"question_id": 212, "text": " Create a virtual assistant that can provide answers to trivia questions about history.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Table Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Table Question Answering\\', \\'api_name\\': \\'google/tapas-small-finetuned-sqa\\', \\'api_call\\': \"pipeline(\\'table-question-answering\\', model=\\'google/tapas-small-finetuned-sqa\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'msr_sqa\\', \\'accuracy\\': 0.6155}, \\'description\\': \\'TAPAS small model fine-tuned on Sequential Question Answering (SQA). It uses relative position embeddings (i.e. resetting the position index at every cell of the table).\\'}', metadata={})]", "category": "generic"}
{"question_id": 213, "text": " I want to build a system that can answer questions from users reading a book.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'distilbert-base-uncased-distilled-squad\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=\\'distilbert-base-uncased-distilled-squad\\')\", \\'api_arguments\\': [\\'question\\', \\'context\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nquestion_answerer = pipeline(question-answering, model=\\'distilbert-base-uncased-distilled-squad\\')\\\\ncontext = r\\\\n... Extractive Question Answering is the task of extracting an answer from a text given a question. An example of a\\\\n... question answering dataset is the SQuAD dataset, which is entirely based on that task. If you would like to fine-tune\\\\n... a model on a SQuAD task, you may leverage the examples/pytorch/question-answering/run_squad.py script.\\\\n... \\\\nresult = question_answerer(question=What is a good example of a question answering dataset?, context=context)\\\\nprint(\\\\n... fAnswer: \\'{result[\\'answer\\']}\\', score: {round(result[\\'score\\'], 4)}, start: {result[\\'start\\']}, end: {result[\\'end\\']}\\\\n...)\", \\'performance\\': {\\'dataset\\': \\'SQuAD v1.1\\', \\'accuracy\\': \\'86.9 F1 score\\'}, \\'description\\': \"DistilBERT base uncased distilled SQuAD is a fine-tuned version of DistilBERT-base-uncased for the task of question answering. It has 40% less parameters than bert-base-uncased, runs 60% faster while preserving over 95% of BERT\\'s performances as measured on the GLUE language understanding benchmark.\"}', metadata={})]", "category": "generic"}
{"question_id": 214, "text": " Our customer support team needs an AI assistant to handle customer inquiries. It should be able to understand and evaluate if the answer provided by the assistant is contradictory, neutral or entails the customer's question.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 215, "text": " Use a zero-shot classifier to classify news headlines into three categories: sports, technology, and politics.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Zero-Shot Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Zero-Shot Classification\\', \\'api_name\\': \\'cross-encoder/nli-deberta-v3-xsmall\\', \\'api_call\\': \"pipeline(\\'zero-shot-classification\\', model=\\'cross-encoder/nli-deberta-v3-xsmall\\')\", \\'api_arguments\\': [\\'sent\\', \\'candidate_labels\\'], \\'python_environment_requirements\\': [\\'sentence_transformers\\', \\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nclassifier = pipeline(\\'zero-shot-classification\\', model=\\'cross-encoder/nli-deberta-v3-xsmall\\')\\\\nsent = \\'Apple just announced the newest iPhone X\\'\\\\ncandidate_labels = [\\'technology\\', \\'sports\\', \\'politics\\']\\\\nres = classifier(sent, candidate_labels)\\\\nprint(res)\", \\'performance\\': {\\'dataset\\': {\\'SNLI-test\\': \\'91.64\\', \\'MNLI_mismatched\\': \\'87.77\\'}}, \\'description\\': \\'This model is a Cross-Encoder for Natural Language Inference, trained on the SNLI and MultiNLI datasets. It can be used for zero-shot classification tasks.\\'}', metadata={})]", "category": "generic"}
{"question_id": 216, "text": " Analyze a German text and classify it into different categories like crime, tragedy, or theft.\\n \n Use this API documentation for reference: [Document(page_content=\"{'domain': 'Natural Language Processing Text Classification', 'framework': 'Hugging Face Transformers', 'functionality': 'German Sentiment Classification', 'api_name': 'oliverguhr/german-sentiment-bert', 'api_call': 'SentimentModel()', 'api_arguments': ['texts'], 'python_environment_requirements': 'pip install germansentiment', 'example_code': ['from germansentiment import SentimentModel', 'model = SentimentModel()', 'texts = [', ' Mit keinem guten Ergebniss,Das ist gar nicht mal so gut,', ' Total awesome!,nicht so schlecht wie erwartet,', ' Der Test verlief positiv.,Sie fhrt ein grnes Auto.]', 'result = model.predict_sentiment(texts)', 'print(result)'], 'performance': {'dataset': ['holidaycheck', 'scare', 'filmstarts', 'germeval', 'PotTS', 'emotions', 'sb10k', 'Leipzig Wikipedia Corpus 2016', 'all'], 'accuracy': [0.9568, 0.9418, 0.9021, 0.7536, 0.678, 0.9649, 0.7376, 0.9967, 0.9639]}, 'description': 'This model was trained for sentiment classification of German language texts. The model uses the Googles Bert architecture and was trained on 1.834 million German-language samples. The training data contains texts from various domains like Twitter, Facebook and movie, app and hotel reviews.'}\", metadata={})]", "category": "generic"}
{"question_id": 217, "text": " Create a tool to determine the logical relationship between two given sentences.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Sentence Similarity\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Sentence Transformers\\', \\'api_name\\': \\'sentence-transformers/distiluse-base-multilingual-cased-v2\\', \\'api_call\\': \"SentenceTransformer(\\'sentence-transformers/distiluse-base-multilingual-cased-v2\\')\", \\'api_arguments\\': [\\'sentences\\'], \\'python_environment_requirements\\': \\'pip install -U sentence-transformers\\', \\'example_code\\': \"from sentence_transformers import SentenceTransformer\\\\nsentences = [This is an example sentence, Each sentence is converted]\\\\nmodel = SentenceTransformer(\\'sentence-transformers/distiluse-base-multilingual-cased-v2\\')\\\\nembeddings = model.encode(sentences)\\\\nprint(embeddings)\", \\'performance\\': {\\'dataset\\': \\'https://seb.sbert.net\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This is a sentence-transformers model: It maps sentences & paragraphs to a 512 dimensional dense vector space and can be used for tasks like clustering or semantic search.\\'}', metadata={})]", "category": "generic"}
{"question_id": 218, "text": " A media company needs to summarize a news article in order to make it easy for their audience to understand the main points quickly.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'it5-base-news-summarization\\', \\'api_call\\': \"pipeline(\\'summarization\\', model=\\'it5/it5-base-news-summarization\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'newsum(Dal 31 maggio \u00e8 infine partita la piattaforma ITsART, a pi\u00f9 di un anno da quando \u2013 durante il primo lockdown \u2013 il ministro della Cultura Dario Franceschini ne aveva parlato come di \u00abuna sorta di Netflix della cultura\u00bb, pensata per \u00aboffrire a tutto il mondo la cultura italiana a pagamento\u00bb. \u00c8 presto per dare giudizi definitivi sulla piattaforma, e di certo sar\u00e0 difficile farlo anche pi\u00f9 avanti senza numeri precisi. Al momento, l\u2019unica cosa che si pu\u00f2 fare \u00e8 guardare com\u2019\u00e8 fatto il sito, contare quanti contenuti ci sono (circa 700 \u201ctitoli\u201d, tra film, documentari, spettacoli teatrali e musicali e altri eventi) e provare a dare un giudizio sul loro valore e sulla loro variet\u00e0. Intanto, una cosa notata da pi\u00f9 parti \u00e8 che diversi contenuti di ITsART sono a pagamento sulla piattaforma sebbene altrove, per esempio su RaiPlay, siano invece disponibili gratuitamente.)\\', \\'performance\\': {\\'dataset\\': \\'NewsSum-IT\\', \\'accuracy\\': {\\'Rouge1\\': 0.339, \\'Rouge2\\': 0.16, \\'RougeL\\': 0.263}}, \\'description\\': \\'IT5 Base model fine-tuned on news summarization on the Fanpage and Il Post corpora for Italian Language Understanding and Generation.\\'}', metadata={})]", "category": "generic"}
{"question_id": 219, "text": " We want to implement a customer service chatbot to engage with website visitors and provide support.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Generation\\', \\'api_name\\': \\'mywateriswet/ShuanBot\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'mywateriswet/ShuanBot\\')\", \\'api_arguments\\': \\'message\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"response = chatbot(\\'What is your name?\\')\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'ShuanBot is a conversational chatbot model based on the GPT-2 architecture. It can be used for generating human-like responses in a chat context.\\'}', metadata={})]", "category": "generic"}
{"question_id": 220, "text": " As the project manager of a company who receives long emails, my goal is to summarize them and extract the most important information.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'text2text-generation\\', \\'api_name\\': \\'google/pegasus-newsroom\\', \\'api_call\\': \"pipeline(\\'summarization\\', model=\\'google/pegasus-newsroom\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'This model can be loaded on the Inference API on-demand.\\', \\'performance\\': {\\'dataset\\': \\'newsroom\\', \\'accuracy\\': \\'45.98/34.20/42.18\\'}, \\'description\\': \\'PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization by Jingqing Zhang, Yao Zhao, Mohammad Saleh, and Peter J. Liu. The model is trained on both C4 and HugeNews datasets and is designed for summarization tasks.\\'}', metadata={})]", "category": "generic"}
{"question_id": 221, "text": " I am the user and I want to start a multi-turn conversation with this model.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Generation\\', \\'api_name\\': \\'mywateriswet/ShuanBot\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'mywateriswet/ShuanBot\\')\", \\'api_arguments\\': \\'message\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"response = chatbot(\\'What is your name?\\')\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'ShuanBot is a conversational chatbot model based on the GPT-2 architecture. It can be used for generating human-like responses in a chat context.\\'}', metadata={})]", "category": "generic"}
{"question_id": 222, "text": " Create a Python function named \\\"summarize_russian_text\\\" that accepts a Russian text as input and returns a brief summary of the text.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'Kirili4ik/mbart_ruDialogSum\\', \\'api_call\\': \"MBartForConditionalGeneration.from_pretrained(\\'Kirili4ik/mbart_ruDialogSum\\')\", \\'api_arguments\\': {\\'model_name\\': \\'Kirili4ik/mbart_ruDialogSum\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'from transformers import MBartTokenizer, MBartForConditionalGeneration\\\\nmodel_name = Kirili4ik/mbart_ruDialogSum\\\\ntokenizer = AutoTokenizer.from_pretrained(model_name)\\\\nmodel = MBartForConditionalGeneration.from_pretrained(model_name)\\\\nmodel.eval()\\\\narticle_text = ...\\\\ninput_ids = tokenizer(\\\\n [article_text],\\\\n max_length=600,\\\\n padding=max_length,\\\\n truncation=True,\\\\n return_tensors=pt,\\\\n)[input_ids]\\\\noutput_ids = model.generate(\\\\n input_ids=input_ids,\\\\n top_k=0,\\\\n num_beams=3,\\\\n no_repeat_ngram_size=3\\\\n)[0]\\\\nsummary = tokenizer.decode(output_ids, skip_special_tokens=True)\\\\nprint(summary)\\', \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'SAMSum Corpus (translated to Russian)\\', \\'accuracy\\': {\\'Validation ROGUE-1\\': 34.5, \\'Validation ROGUE-L\\': 33, \\'Test ROGUE-1\\': 31, \\'Test ROGUE-L\\': 28}}]}, \\'description\\': \\'MBart for Russian summarization fine-tuned for dialogues summarization. This model was firstly fine-tuned by Ilya Gusev on Gazeta dataset. We have fine tuned that model on SamSum dataset translated to Russian using GoogleTranslateAPI. Moreover! We have implemented a ! telegram bot @summarization_bot ! with the inference of this model. Add it to the chat and get summaries instead of dozens spam messages!\\'}', metadata={})]", "category": "generic"}
{"question_id": 223, "text": " We want to implement a chatbot on our website to provide quick responses to customer inquiries.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Generation\\', \\'api_name\\': \\'mywateriswet/ShuanBot\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'mywateriswet/ShuanBot\\')\", \\'api_arguments\\': \\'message\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"response = chatbot(\\'What is your name?\\')\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'ShuanBot is a conversational chatbot model based on the GPT-2 architecture. It can be used for generating human-like responses in a chat context.\\'}', metadata={})]", "category": "generic"}
{"question_id": 224, "text": " We want the AI to have an interesting conversation and ask it general knowledge questions.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'hyunwoongko/blenderbot-9B\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'hyunwoongko/blenderbot-9B\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'Input a message to start chatting with hyunwoongko/blenderbot-9B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high-performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi-turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 225, "text": " We need to create a dialogue in Russian for our educational app. It should cover a general greeting and asking about the users' well-being.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'Kirili4ik/mbart_ruDialogSum\\', \\'api_call\\': \"MBartForConditionalGeneration.from_pretrained(\\'Kirili4ik/mbart_ruDialogSum\\')\", \\'api_arguments\\': {\\'model_name\\': \\'Kirili4ik/mbart_ruDialogSum\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'from transformers import MBartTokenizer, MBartForConditionalGeneration\\\\nmodel_name = Kirili4ik/mbart_ruDialogSum\\\\ntokenizer = AutoTokenizer.from_pretrained(model_name)\\\\nmodel = MBartForConditionalGeneration.from_pretrained(model_name)\\\\nmodel.eval()\\\\narticle_text = ...\\\\ninput_ids = tokenizer(\\\\n [article_text],\\\\n max_length=600,\\\\n padding=max_length,\\\\n truncation=True,\\\\n return_tensors=pt,\\\\n)[input_ids]\\\\noutput_ids = model.generate(\\\\n input_ids=input_ids,\\\\n top_k=0,\\\\n num_beams=3,\\\\n no_repeat_ngram_size=3\\\\n)[0]\\\\nsummary = tokenizer.decode(output_ids, skip_special_tokens=True)\\\\nprint(summary)\\', \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'SAMSum Corpus (translated to Russian)\\', \\'accuracy\\': {\\'Validation ROGUE-1\\': 34.5, \\'Validation ROGUE-L\\': 33, \\'Test ROGUE-1\\': 31, \\'Test ROGUE-L\\': 28}}]}, \\'description\\': \\'MBart for Russian summarization fine-tuned for dialogues summarization. This model was firstly fine-tuned by Ilya Gusev on Gazeta dataset. We have fine tuned that model on SamSum dataset translated to Russian using GoogleTranslateAPI. Moreover! We have implemented a ! telegram bot @summarization_bot ! with the inference of this model. Add it to the chat and get summaries instead of dozens spam messages!\\'}', metadata={})]", "category": "generic"}
{"question_id": 226, "text": " The company wants to create a chatbot to help answer customer questions regarding the chatbot's consciousness. We need to be able to generate sensible responses.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Generation\\', \\'api_name\\': \\'mywateriswet/ShuanBot\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'mywateriswet/ShuanBot\\')\", \\'api_arguments\\': \\'message\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"response = chatbot(\\'What is your name?\\')\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'ShuanBot is a conversational chatbot model based on the GPT-2 architecture. It can be used for generating human-like responses in a chat context.\\'}', metadata={})]", "category": "generic"}
{"question_id": 227, "text": " Create interesting variations of the given phrase 'How can I improve my time management skills?'.\\n###Input: How can I improve my time management skills?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Paraphrasing\\', \\'api_name\\': \\'prithivida/parrot_paraphraser_on_T5\\', \\'api_call\\': \"Parrot(model_tag=\\'prithivida/parrot_paraphraser_on_T5\\', use_gpu=False)\", \\'api_arguments\\': [\\'input_phrase\\', \\'diversity_ranker\\', \\'do_diverse\\', \\'max_return_phrases\\', \\'max_length\\', \\'adequacy_threshold\\', \\'fluency_threshold\\'], \\'python_environment_requirements\\': [\\'torch\\', \\'transformers\\'], \\'example_code\\': \\'from parrot import Parrot\\\\nimport torch\\\\nimport warnings\\\\nwarnings.filterwarnings(ignore)\\\\n\\\\nparrot = Parrot(model_tag=prithivida/parrot_paraphraser_on_T5, use_gpu=False)\\\\nphrases = [Can you recommed some upscale restaurants in Newyork?,\\\\n What are the famous places we should not miss in Russia?\\\\n]\\\\nfor phrase in phrases:\\\\n print(-*100)\\\\n print(Input_phrase: , phrase)\\\\n print(-*100)\\\\n para_phrases = parrot.augment(input_phrase=phrase)\\\\n for para_phrase in para_phrases:\\\\n  print(para_phrase)\\', \\'performance\\': {\\'dataset\\': \\'Not mentioned\\', \\'accuracy\\': \\'Not mentioned\\'}, \\'description\\': \\'Parrot is a paraphrase based utterance augmentation framework purpose built to accelerate training NLU models. It offers knobs to control Adequacy, Fluency, and Diversity as per your needs. It mainly focuses on augmenting texts typed-into or spoken-to conversational interfaces for building robust NLU models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 228, "text": " A customer wants to convert an input Korean text into a summary. Provide a solution for it.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'pszemraj/long-t5-tglobal-base-16384-book-summary\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'pszemraj/long-t5-tglobal-base-16384-book-summary\\')\", \\'api_arguments\\': [\\'long_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nimport torch\\\\nsummarizer = pipeline(\\\\n summarization,\\\\n pszemraj/long-t5-tglobal-base-16384-book-summary,\\\\n device=0 if torch.cuda.is_available() else -1,\\\\n)\\\\nlong_text = Here is a lot of text I don\\'t want to read. Replace me\\\\nresult = summarizer(long_text)\\\\nprint(result[0][summary_text])\", \\'performance\\': {\\'dataset\\': \\'kmfoda/booksum\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.408, \\'ROUGE-2\\': 6.065, \\'ROUGE-L\\': 16.721, \\'ROUGE-LSUM\\': 33.34}}, \\'description\\': \\'A fine-tuned version of google/long-t5-tglobal-base on the kmfoda/booksum dataset, which can be used to summarize long text and generate SparkNotes-esque summaries of arbitrary topics. The model generalizes reasonably well to academic and narrative text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 229, "text": " Our company is providing translation services for business clients. We need to translate English contracts to French.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Translation\\', \\'api_name\\': \\'opus-mt-fr-en\\', \\'api_call\\': \"pipeline(\\'translation_fr_to_en\\', model=\\'Helsinki-NLP/opus-mt-fr-en\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\', \\'torch\\'], \\'example_code\\': \"translation_pipeline(\\'Bonjour, comment \u00e7a va?\\')\", \\'performance\\': {\\'dataset\\': \\'opus\\', \\'accuracy\\': {\\'BLEU\\': {\\'newsdiscussdev2015-enfr.fr.en\\': 33.1, \\'newsdiscusstest2015-enfr.fr.en\\': 38.7, \\'newssyscomb2009.fr.en\\': 30.3, \\'news-test2008.fr.en\\': 26.2, \\'newstest2009.fr.en\\': 30.2, \\'newstest2010.fr.en\\': 32.2, \\'newstest2011.fr.en\\': 33.0, \\'newstest2012.fr.en\\': 32.8, \\'newstest2013.fr.en\\': 33.9, \\'newstest2014-fren.fr.en\\': 37.8, \\'Tatoeba.fr.en\\': 57.5}}}, \\'description\\': \\'Helsinki-NLP/opus-mt-fr-en is a machine translation model trained to translate from French to English. It is based on the Marian NMT framework and trained on the OPUS dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 230, "text": " We are an online platform that offers various Chinese language courses for students. Now we need to autoplay a fill-in-the-blank video game for our users.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Generation\\', \\'api_name\\': \\'mywateriswet/ShuanBot\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'mywateriswet/ShuanBot\\')\", \\'api_arguments\\': \\'message\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"response = chatbot(\\'What is your name?\\')\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'ShuanBot is a conversational chatbot model based on the GPT-2 architecture. It can be used for generating human-like responses in a chat context.\\'}', metadata={})]", "category": "generic"}
{"question_id": 231, "text": " We want to determine the similarity between two sentences, \\\"I love going to the park\\\" and \\\"My favorite activity is visiting the park\\\".\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Sentence Similarity\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Sentence Embeddings\\', \\'api_name\\': \\'sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2\\', \\'api_call\\': \"SentenceTransformer(\\'sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2\\')\", \\'api_arguments\\': [\\'sentences\\'], \\'python_environment_requirements\\': \\'pip install -U sentence-transformers\\', \\'example_code\\': \"from sentence_transformers import SentenceTransformer\\\\nsentences = [This is an example sentence, Each sentence is converted]\\\\nmodel = SentenceTransformer(\\'sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2\\')\\\\nembeddings = model.encode(sentences)\\\\nprint(embeddings)\", \\'performance\\': {\\'dataset\\': \\'https://seb.sbert.net\\', \\'accuracy\\': \\'Automated evaluation\\'}, \\'description\\': \\'This is a sentence-transformers model: It maps sentences & paragraphs to a 384 dimensional dense vector space and can be used for tasks like clustering or semantic search.\\'}', metadata={})]", "category": "generic"}
{"question_id": 232, "text": " We have an audiobook app in the Chinese language, and we want to convert the Chinese text to speech.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'kan-bayashi_csmsc_tts_train_tacotron2_raw_phn_pypinyin_g2p_phone_train.loss.best\\', \\'api_call\\': \"Text2Speech.from_pretrained(\\'espnet/kan-bayashi_csmsc_tts_train_tacotron2_raw_phn_pypinyin_g2p_phone_train.loss.best\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'torch\\', \\'espnet_model_zoo\\'], \\'example_code\\': \\'import soundfile\\\\nfrom espnet2.bin.tts_inference import Text2Speech\\\\ntext2speech = Text2Speech.from_pretrained(espnet/kan-bayashi_csmsc_tts_train_tacotron2_raw_phn_pypinyin_g2p_phone_train.loss.best)\\\\ntext = \u6625\u6c5f\u6f6e\u6c34\u8fde\u6d77\u5e73\uff0c\u6d77\u4e0a\u660e\u6708\u5171\u6f6e\u751f\\\\nspeech = text2speech(text)[wav]\\\\nsoundfile.write(out.wav, speech.numpy(), text2speech.fs, PCM_16)\\', \\'performance\\': {\\'dataset\\': \\'csmsc\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'A pre-trained Text-to-Speech model for Chinese language using ESPnet framework. It can be used to convert text input into speech output in Chinese.\\'}', metadata={})]", "category": "generic"}
{"question_id": 233, "text": " We want to build an audiobook service. We need to convert text to speech for the user to listen.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'lakahaga/novel_reading_tts\\', \\'api_call\\': \"AutoModelForTTS.from_pretrained(\\'lakahaga/novel_reading_tts\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"inputs = processor(text, return_tensors=\\'pt\\'); generated_audio = model.generate(**inputs);\", \\'performance\\': {\\'dataset\\': \\'novelspeech\\', \\'accuracy\\': None}, \\'description\\': \\'This model was trained by lakahaga using novelspeech recipe in espnet. It is designed for Korean text-to-speech tasks.\\'}', metadata={})]", "category": "generic"}
{"question_id": 234, "text": " I want to convert English speech into different language speech. I have some audio in English language and I would like to convert that into Spanish.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'SYSPIN/Marathi_Male_TTS\\', \\'api_call\\': \"api.load(\\'ESPnet/espnet_model_zoo:SYSPIN/Marathi_Male_TTS\\').\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'huggingface_hub\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Marathi Male Text-to-Speech model using ESPnet framework.\\'}', metadata={})]", "category": "generic"}
{"question_id": 235, "text": " We are trying to transcribe audios. Can you help with giving us an ASR model?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'CTranslate2\\', \\'functionality\\': \\'Automatic Speech Recognition\\', \\'api_name\\': \\'guillaumekln/faster-whisper-large-v2\\', \\'api_call\\': \"WhisperModel(\\'large-v2\\')\", \\'api_arguments\\': [\\'audio.mp3\\'], \\'python_environment_requirements\\': [\\'faster_whisper\\'], \\'example_code\\': \\'from faster_whisper import WhisperModel\\\\nmodel = WhisperModel(large-v2)\\\\nsegments, info = model.transcribe(audio.mp3)\\\\nfor segment in segments:\\\\n print([%.2fs -&gt; %.2fs] %s % (segment.start, segment.end, segment.text))\\', \\'performance\\': {\\'dataset\\': \\'99 languages\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Whisper large-v2 model for CTranslate2. This model can be used in CTranslate2 or projets based on CTranslate2 such as faster-whisper.\\'}', metadata={})]", "category": "generic"}
{"question_id": 236, "text": " Our blog platform accepts voice notes, and we want to transcribe these notes to text.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'kan-bayashi_ljspeech_vits\\', \\'api_call\\': \"pipeline(\\'text-to-speech\\', model=\\'espnet/kan-bayashi_ljspeech_vits\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"from transformers import pipeline; tts = pipeline(\\'text-to-speech\\', model=\\'espnet/kan-bayashi_ljspeech_vits\\'); tts(\\'Hello World\\')\", \\'performance\\': {\\'dataset\\': \\'ljspeech\\', \\'accuracy\\': \\'Not mentioned\\'}, \\'description\\': \\'A Text-to-Speech model trained on the ljspeech dataset using the ESPnet toolkit. This model can be used to convert text input into synthesized speech.\\'}', metadata={})]", "category": "generic"}
{"question_id": 237, "text": " Create a program to separate music and vocals from an audio file using a pretrained model.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Audio-to-Audio\\', \\'framework\\': \\'SpeechBrain\\', \\'functionality\\': \\'Audio Source Separation\\', \\'api_name\\': \\'sepformer-wsj02mix\\', \\'api_call\\': \"separator.from_hparams(source=\\'speechbrain/sepformer-wsj02mix\\')\", \\'api_arguments\\': [\\'path\\'], \\'python_environment_requirements\\': [\\'pip install speechbrain\\'], \\'example_code\\': [\\'from speechbrain.pretrained import SepformerSeparation as separator\\', \\'import torchaudio\\', \"model = separator.from_hparams(source=speechbrain/sepformer-wsj02mix, savedir=\\'pretrained_models/sepformer-wsj02mix\\')\", \"est_sources = model.separate_file(path=\\'speechbrain/sepformer-wsj02mix/test_mixture.wav\\')\", \\'torchaudio.save(source1hat.wav, est_sources[:, :, 0].detach().cpu(), 8000)\\', \\'torchaudio.save(source2hat.wav, est_sources[:, :, 1].detach().cpu(), 8000)\\'], \\'performance\\': {\\'dataset\\': \\'WSJ0-2Mix\\', \\'accuracy\\': \\'22.4 dB\\'}, \\'description\\': \\'This repository provides all the necessary tools to perform audio source separation with a SepFormer model, implemented with SpeechBrain, and pretrained on WSJ0-2Mix dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 238, "text": " \\nCreate an application that automatically separates vocals from a song for karaoke nights.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Audio-to-Audio\\', \\'framework\\': \\'SpeechBrain\\', \\'functionality\\': \\'Audio Source Separation\\', \\'api_name\\': \\'sepformer-wsj02mix\\', \\'api_call\\': \"separator.from_hparams(source=\\'speechbrain/sepformer-wsj02mix\\')\", \\'api_arguments\\': [\\'path\\'], \\'python_environment_requirements\\': [\\'pip install speechbrain\\'], \\'example_code\\': [\\'from speechbrain.pretrained import SepformerSeparation as separator\\', \\'import torchaudio\\', \"model = separator.from_hparams(source=speechbrain/sepformer-wsj02mix, savedir=\\'pretrained_models/sepformer-wsj02mix\\')\", \"est_sources = model.separate_file(path=\\'speechbrain/sepformer-wsj02mix/test_mixture.wav\\')\", \\'torchaudio.save(source1hat.wav, est_sources[:, :, 0].detach().cpu(), 8000)\\', \\'torchaudio.save(source2hat.wav, est_sources[:, :, 1].detach().cpu(), 8000)\\'], \\'performance\\': {\\'dataset\\': \\'WSJ0-2Mix\\', \\'accuracy\\': \\'22.4 dB\\'}, \\'description\\': \\'This repository provides all the necessary tools to perform audio source separation with a SepFormer model, implemented with SpeechBrain, and pretrained on WSJ0-2Mix dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 239, "text": " One of our clients need high recognition capability of denoising audio streams for improving sound quality.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'pyannote.audio\\', \\'functionality\\': \\'Speaker diarization\\', \\'api_name\\': \\'johnislarry/cloned-pyannote-speaker-diarization-endpoint\\', \\'api_call\\': \"Pipeline.from_pretrained(\\'pyannote/speaker-diarization@2.1\\',use_auth_token=\\'ACCESS_TOKEN_GOES_HERE\\')\", \\'api_arguments\\': [\\'num_speakers\\', \\'min_speakers\\', \\'max_speakers\\', \\'segmentation_onset\\'], \\'python_environment_requirements\\': \\'pyannote.audio 2.0\\', \\'example_code\\': {\\'load_pipeline\\': \\'from pyannote.audio import Pipeline\\\\npipeline = Pipeline.from_pretrained(pyannote/speaker-diarization@2022.07)\\', \\'apply_pipeline\\': \\'diarization = pipeline(audio.wav)\\', \\'save_output\\': \\'with open(audio.rttm, w) as rttm:\\\\n  diarization.write_rttm(rttm)\\'}, \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'AISHELL-4\\', \\'accuracy\\': {\\'DER%\\': 14.61, \\'FA%\\': 3.31, \\'Miss%\\': 4.35, \\'Conf%\\': 6.95}}, {\\'name\\': \\'AMI Mix-Headset only_words\\', \\'accuracy\\': {\\'DER%\\': 18.21, \\'FA%\\': 3.28, \\'Miss%\\': 11.07, \\'Conf%\\': 3.87}}, {\\'name\\': \\'AMI Array1-01 only_words\\', \\'accuracy\\': {\\'DER%\\': 29.0, \\'FA%\\': 2.71, \\'Miss%\\': 21.61, \\'Conf%\\': 4.68}}, {\\'name\\': \\'CALLHOME Part2\\', \\'accuracy\\': {\\'DER%\\': 30.24, \\'FA%\\': 3.71, \\'Miss%\\': 16.86, \\'Conf%\\': 9.66}}, {\\'name\\': \\'DIHARD 3 Full\\', \\'accuracy\\': {\\'DER%\\': 20.99, \\'FA%\\': 4.25, \\'Miss%\\': 10.74, \\'Conf%\\': 6.0}}, {\\'name\\': \\'REPERE Phase 2\\', \\'accuracy\\': {\\'DER%\\': 12.62, \\'FA%\\': 1.55, \\'Miss%\\': 3.3, \\'Conf%\\': 7.76}}, {\\'name\\': \\'VoxConverse v0.0.2\\', \\'accuracy\\': {\\'DER%\\': 12.76, \\'FA%\\': 3.45, \\'Miss%\\': 3.85, \\'Conf%\\': 5.46}}]}, \\'description\\': \\'This API provides speaker diarization functionality using the pyannote.audio framework. It is capable of processing audio files and outputting speaker diarization results in RTTM format. The API supports providing the number of speakers, minimum and maximum number of speakers, and adjusting the segmentation onset threshold.\\'}', metadata={})]", "category": "generic"}
{"question_id": 240, "text": " A podcast company reached out to us wanting to change their speaker's voice in a recorded podcast.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'pyannote.audio\\', \\'functionality\\': \\'Speaker diarization\\', \\'api_name\\': \\'johnislarry/cloned-pyannote-speaker-diarization-endpoint\\', \\'api_call\\': \"Pipeline.from_pretrained(\\'pyannote/speaker-diarization@2.1\\',use_auth_token=\\'ACCESS_TOKEN_GOES_HERE\\')\", \\'api_arguments\\': [\\'num_speakers\\', \\'min_speakers\\', \\'max_speakers\\', \\'segmentation_onset\\'], \\'python_environment_requirements\\': \\'pyannote.audio 2.0\\', \\'example_code\\': {\\'load_pipeline\\': \\'from pyannote.audio import Pipeline\\\\npipeline = Pipeline.from_pretrained(pyannote/speaker-diarization@2022.07)\\', \\'apply_pipeline\\': \\'diarization = pipeline(audio.wav)\\', \\'save_output\\': \\'with open(audio.rttm, w) as rttm:\\\\n  diarization.write_rttm(rttm)\\'}, \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'AISHELL-4\\', \\'accuracy\\': {\\'DER%\\': 14.61, \\'FA%\\': 3.31, \\'Miss%\\': 4.35, \\'Conf%\\': 6.95}}, {\\'name\\': \\'AMI Mix-Headset only_words\\', \\'accuracy\\': {\\'DER%\\': 18.21, \\'FA%\\': 3.28, \\'Miss%\\': 11.07, \\'Conf%\\': 3.87}}, {\\'name\\': \\'AMI Array1-01 only_words\\', \\'accuracy\\': {\\'DER%\\': 29.0, \\'FA%\\': 2.71, \\'Miss%\\': 21.61, \\'Conf%\\': 4.68}}, {\\'name\\': \\'CALLHOME Part2\\', \\'accuracy\\': {\\'DER%\\': 30.24, \\'FA%\\': 3.71, \\'Miss%\\': 16.86, \\'Conf%\\': 9.66}}, {\\'name\\': \\'DIHARD 3 Full\\', \\'accuracy\\': {\\'DER%\\': 20.99, \\'FA%\\': 4.25, \\'Miss%\\': 10.74, \\'Conf%\\': 6.0}}, {\\'name\\': \\'REPERE Phase 2\\', \\'accuracy\\': {\\'DER%\\': 12.62, \\'FA%\\': 1.55, \\'Miss%\\': 3.3, \\'Conf%\\': 7.76}}, {\\'name\\': \\'VoxConverse v0.0.2\\', \\'accuracy\\': {\\'DER%\\': 12.76, \\'FA%\\': 3.45, \\'Miss%\\': 3.85, \\'Conf%\\': 5.46}}]}, \\'description\\': \\'This API provides speaker diarization functionality using the pyannote.audio framework. It is capable of processing audio files and outputting speaker diarization results in RTTM format. The API supports providing the number of speakers, minimum and maximum number of speakers, and adjusting the segmentation onset threshold.\\'}', metadata={})]", "category": "generic"}
{"question_id": 241, "text": " Our research team is focused on the analysis and separation of complex audio recordings. We need a model to be able to separate speaker voices from mixed sound.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Audio-to-Audio\\', \\'framework\\': \\'SpeechBrain\\', \\'functionality\\': \\'Audio Source Separation\\', \\'api_name\\': \\'sepformer-wsj02mix\\', \\'api_call\\': \"separator.from_hparams(source=\\'speechbrain/sepformer-wsj02mix\\')\", \\'api_arguments\\': [\\'path\\'], \\'python_environment_requirements\\': [\\'pip install speechbrain\\'], \\'example_code\\': [\\'from speechbrain.pretrained import SepformerSeparation as separator\\', \\'import torchaudio\\', \"model = separator.from_hparams(source=speechbrain/sepformer-wsj02mix, savedir=\\'pretrained_models/sepformer-wsj02mix\\')\", \"est_sources = model.separate_file(path=\\'speechbrain/sepformer-wsj02mix/test_mixture.wav\\')\", \\'torchaudio.save(source1hat.wav, est_sources[:, :, 0].detach().cpu(), 8000)\\', \\'torchaudio.save(source2hat.wav, est_sources[:, :, 1].detach().cpu(), 8000)\\'], \\'performance\\': {\\'dataset\\': \\'WSJ0-2Mix\\', \\'accuracy\\': \\'22.4 dB\\'}, \\'description\\': \\'This repository provides all the necessary tools to perform audio source separation with a SepFormer model, implemented with SpeechBrain, and pretrained on WSJ0-2Mix dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 242, "text": " A Romanian-speaking person wants to communicate with an English-speaking friend over the phone using our platform. Please provide a method to translate their speech in real-time.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Audio-to-Audio\\', \\'framework\\': \\'Fairseq\\', \\'functionality\\': \\'speech-to-speech-translation\\', \\'api_name\\': \\'facebook/textless_sm_ro_en\\', \\'api_call\\': \"pipeline(\\'audio-to-audio\\', model=\\'facebook/textless_sm_ro_en\\')\", \\'api_arguments\\': \\'audio file or recording\\', \\'python_environment_requirements\\': \\'fairseq, huggingface_hub\\', \\'example_code\\': \\'https://huggingface.co/facebook/textless_sm_cs_en\\', \\'performance\\': {\\'dataset\\': \\'unknown\\', \\'accuracy\\': \\'unknown\\'}, \\'description\\': \\'A speech-to-speech translation model for Romanian to English developed by Facebook AI\\'}', metadata={})]", "category": "generic"}
{"question_id": 243, "text": " Implement an application that recognizes languages by analyzing an online audio file.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'CTranslate2\\', \\'functionality\\': \\'Automatic Speech Recognition\\', \\'api_name\\': \\'guillaumekln/faster-whisper-large-v2\\', \\'api_call\\': \"WhisperModel(\\'large-v2\\')\", \\'api_arguments\\': [\\'audio.mp3\\'], \\'python_environment_requirements\\': [\\'faster_whisper\\'], \\'example_code\\': \\'from faster_whisper import WhisperModel\\\\nmodel = WhisperModel(large-v2)\\\\nsegments, info = model.transcribe(audio.mp3)\\\\nfor segment in segments:\\\\n print([%.2fs -&gt; %.2fs] %s % (segment.start, segment.end, segment.text))\\', \\'performance\\': {\\'dataset\\': \\'99 languages\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Whisper large-v2 model for CTranslate2. This model can be used in CTranslate2 or projets based on CTranslate2 such as faster-whisper.\\'}', metadata={})]", "category": "generic"}
{"question_id": 244, "text": " A toy company wants to use your software to recognize spoken numbers (0-9) in English by young children for an interactive game. Please create an identifier.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'imdanboy/jets\\', \\'api_call\\': \"pipeline(\\'text-to-speech\\', model=\\'imdanboy/jets\\')\", \\'api_arguments\\': None, \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"from transformers import pipeline; tts = pipeline(\\'text-to-speech\\', model=\\'imdanboy/jets\\'); tts(\\'Hello world\\')\", \\'performance\\': {\\'dataset\\': \\'ljspeech\\', \\'accuracy\\': None}, \\'description\\': \\'This model was trained by imdanboy using ljspeech recipe in espnet.\\'}', metadata={})]", "category": "generic"}
{"question_id": 245, "text": " Develop an automated podcast recommender system to suggest content that matches user's taste based on their favorite podcast speakers.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'pyannote.audio\\', \\'functionality\\': \\'Speaker Diarization\\', \\'api_name\\': \\'pyannote/speaker-diarization\\', \\'api_call\\': \"Pipeline.from_pretrained(\\'pyannote/speaker-diarization@2.1\\', use_auth_token=\\'ACCESS_TOKEN_GOES_HERE\\')\", \\'api_arguments\\': {\\'num_speakers\\': \\'int (optional)\\', \\'min_speakers\\': \\'int (optional)\\', \\'max_speakers\\': \\'int (optional)\\'}, \\'python_environment_requirements\\': \\'pyannote.audio 2.1.1\\', \\'example_code\\': [\\'from pyannote.audio import Pipeline\\', \\'pipeline = Pipeline.from_pretrained(pyannote/speaker-diarization@2.1, use_auth_token=ACCESS_TOKEN_GOES_HERE)\\', \\'diarization = pipeline(audio.wav)\\', \\'with open(audio.rttm, w) as rttm:\\', \\'  diarization.write_rttm(rttm)\\'], \\'performance\\': {\\'dataset\\': \\'ami\\', \\'accuracy\\': {\\'DER%\\': \\'18.91\\', \\'FA%\\': \\'4.48\\', \\'Miss%\\': \\'9.51\\', \\'Conf%\\': \\'4.91\\'}}, \\'description\\': \\'This API provides an automatic speaker diarization pipeline using the pyannote.audio framework. It can process audio files and output speaker diarization results in RTTM format. The pipeline can also handle cases where the number of speakers is known in advance or when providing lower and/or upper bounds on the number of speakers.\\'}', metadata={})]", "category": "generic"}
{"question_id": 246, "text": " We are developing an app that can help mediate virtual therapy sessions. We need a feature that can detect emotions in the voice of the user.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Voice Activity Detection\\', \\'api_name\\': \\'d4data/Indian-voice-cloning\\', \\'api_call\\': \"pipeline(\\'voice-activity-detection\\', model=\\'d4data/Indian-voice-cloning\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A model for detecting voice activity in Indian languages.\\'}', metadata={})]", "category": "generic"}
{"question_id": 247, "text": " We are engineering a green consulting firm. We need to predict carbon emissions with input features from customer's dataset.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'kochetkovIT/autotrain-ironhack-49741119788\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'json\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'kochetkovIT/autotrain-data-ironhack\\', \\'accuracy\\': {\\'Loss\\': 2.603, \\'R2\\': 0.013, \\'MSE\\': 6.776, \\'MAE\\': 1.666, \\'RMSLE\\': 0.502}}, \\'description\\': \\'A tabular regression model trained using AutoTrain to predict carbon emissions (in grams) with an R2 score of 0.013.\\'}', metadata={})]", "category": "generic"}
{"question_id": 248, "text": " Develop a system that can detect voices in a podcast to find out if guests are speaking or not. \\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'pyannote.audio\\', \\'functionality\\': \\'Automatic Speech Recognition\\', \\'api_name\\': \\'pyannote/voice-activity-detection\\', \\'api_call\\': \"Pipeline.from_pretrained(\\'pyannote/voice-activity-detection\\')\", \\'api_arguments\\': [\\'audio.wav\\'], \\'python_environment_requirements\\': [\\'pyannote.audio 2.1\\'], \\'example_code\\': \\'from pyannote.audio import Pipeline\\\\npipeline = Pipeline.from_pretrained(pyannote/voice-activity-detection, use_auth_token=ACCESS_TOKEN_GOES_HERE)\\\\noutput = pipeline(audio.wav)\\\\nfor speech in output.get_timeline().support():\\\\n  # active speech between speech.start and speech.end\\', \\'performance\\': {\\'dataset\\': \\'ami\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'A pretrained voice activity detection pipeline that detects active speech in audio files.\\'}', metadata={})]", "category": "generic"}
{"question_id": 249, "text": " Develop a machine learning product to help home buyers predict house prices in the US.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Joblib\\', \\'functionality\\': \\'Single Column Regression\\', \\'api_name\\': \\'jwan2021/autotrain-us-housing-prices-1771761512\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'joblib\\', \\'pandas\\', \\'json\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'jwan2021/autotrain-data-us-housing-prices\\', \\'accuracy\\': {\\'Loss\\': 122809.223, \\'R2\\': 0.884, \\'MSE\\': 15082105200.447, \\'MAE\\': 95586.887, \\'RMSLE\\': 0.13}}, \\'description\\': \\'A model trained using AutoTrain for predicting US housing prices with single column regression. The model is based on the jwan2021/autotrain-data-us-housing-prices dataset and has a CO2 Emissions of 50.5369 grams.\\'}', metadata={})]", "category": "generic"}
{"question_id": 250, "text": " A game developer needs to predict the HP of a new Pokemon character based on several attributes. Provide the code to make that prediction.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Predicting Pokemon HP\\', \\'api_name\\': \\'julien-c/pokemon-predict-hp\\', \\'api_call\\': \"pipeline(\\'regression\\', model=\\'julien-c/pokemon-predict-hp\\')\", \\'api_arguments\\': \\'input_data\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'julien-c/kaggle-rounakbanik-pokemon\\', \\'accuracy\\': {\\'mean_absolute_error\\': 15.909, \\'model_loss\\': 647.605}}, \\'description\\': \\'A tabular regression model trained on the julien-c/kaggle-rounakbanik-pokemon dataset to predict the HP of Pokemon.\\'}', metadata={})]", "category": "generic"}
{"question_id": 251, "text": " Let's create a smart agent that can learn to play CartPole-v0 using reinforcement learning.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Reinforcement Learning\\', \\'framework\\': \\'Stable-Baselines3\\', \\'functionality\\': \\'CartPole-v1\\', \\'api_name\\': \\'dqn-CartPole-v1\\', \\'api_call\\': \"load_from_hub(repo_id=\\'sb3/dqn-CartPole-v1\\',filename=\\'{MODEL FILENAME}.zip\\',)\", \\'api_arguments\\': [\\'algo\\', \\'env\\', \\'logs\\'], \\'python_environment_requirements\\': [\\'rl_zoo3\\', \\'stable-baselines3\\', \\'stable-baselines3-contrib\\'], \\'example_code\\': \\'python train.py --algo dqn --env CartPole-v1 -f logs/\\', \\'performance\\': {\\'dataset\\': \\'CartPole-v1\\', \\'accuracy\\': \\'500.00 +/- 0.00\\'}, \\'description\\': \\'This is a trained model of a DQN agent playing CartPole-v1 using the stable-baselines3 library and the RL Zoo. The RL Zoo is a training framework for Stable Baselines3 reinforcement learning agents, with hyperparameter optimization and pre-trained agents included.\\'}', metadata={})]", "category": "generic"}
{"question_id": 252, "text": " The sports equipment company I work for wants to build a virtual soccer training game. The agent must be able to effectively play soccer against another team.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Reinforcement Learning\\', \\'framework\\': \\'Unity ML-Agents Library\\', \\'functionality\\': \\'Train and play SoccerTwos\\', \\'api_name\\': \\'poca-SoccerTwosv2\\', \\'api_call\\': \"mlagents-load-from-hf --repo-id=\\'Raiden-1001/poca-SoccerTwosv2\\' --local-dir=\\'./downloads\\'\", \\'api_arguments\\': [\\'your_configuration_file_path.yaml\\', \\'run_id\\'], \\'python_environment_requirements\\': [\\'ml-agents\\'], \\'example_code\\': \\'mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume\\', \\'performance\\': {\\'dataset\\': \\'SoccerTwos\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'A trained model of a poca agent playing SoccerTwos using the Unity ML-Agents Library.\\'}', metadata={})]", "category": "generic"}
{"question_id": 253, "text": " We want to build a smart-cars training system for our team. I need you to solve the gym environment 'MountainCar-v0'.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Reinforcement Learning\\', \\'framework\\': \\'Stable-Baselines3\\', \\'functionality\\': \\'MountainCar-v0\\', \\'api_name\\': \\'sb3/dqn-MountainCar-v0\\', \\'api_call\\': \"load_from_hub(repo_id=\\'sb3/dqn-MountainCar-v0\\',filename=\\'{MODEL FILENAME}.zip\\',)\", \\'api_arguments\\': [\\'algo\\', \\'env\\', \\'f\\'], \\'python_environment_requirements\\': [\\'RL Zoo\\', \\'SB3\\', \\'SB3 Contrib\\'], \\'example_code\\': [\\'python -m rl_zoo3.load_from_hub --algo dqn --env MountainCar-v0 -orga sb3 -f logs/\\', \\'python train.py --algo dqn --env MountainCar-v0 -f logs/\\', \\'python -m rl_zoo3.push_to_hub --algo dqn --env MountainCar-v0 -f logs/ -orga sb3\\'], \\'performance\\': {\\'dataset\\': \\'MountainCar-v0\\', \\'accuracy\\': \\'-103.40 +/- 7.49\\'}, \\'description\\': \\'This is a trained model of a DQN agent playing MountainCar-v0 using the stable-baselines3 library and the RL Zoo. The RL Zoo is a training framework for Stable Baselines3 reinforcement learning agents, with hyperparameter optimization and pre-trained agents included.\\'}', metadata={})]", "category": "generic"}
{"question_id": 254, "text": " We are a group of teachers looking to create an automated summary based on the student's essays.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'pszemraj/long-t5-tglobal-base-16384-book-summary\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'pszemraj/long-t5-tglobal-base-16384-book-summary\\')\", \\'api_arguments\\': [\\'long_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nimport torch\\\\nsummarizer = pipeline(\\\\n summarization,\\\\n pszemraj/long-t5-tglobal-base-16384-book-summary,\\\\n device=0 if torch.cuda.is_available() else -1,\\\\n)\\\\nlong_text = Here is a lot of text I don\\'t want to read. Replace me\\\\nresult = summarizer(long_text)\\\\nprint(result[0][summary_text])\", \\'performance\\': {\\'dataset\\': \\'kmfoda/booksum\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.408, \\'ROUGE-2\\': 6.065, \\'ROUGE-L\\': 16.721, \\'ROUGE-LSUM\\': 33.34}}, \\'description\\': \\'A fine-tuned version of google/long-t5-tglobal-base on the kmfoda/booksum dataset, which can be used to summarize long text and generate SparkNotes-esque summaries of arbitrary topics. The model generalizes reasonably well to academic and narrative text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 255, "text": " An important project is under development, and we need to know the positional relations between various keywords of a given medical text.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Text2Text Generation\\', \\'api_name\\': \\'castorini/doc2query-t5-base-msmarco\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'castorini/doc2query-t5-base-msmarco\\')\", \\'api_arguments\\': \\'text, max_length\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'MS MARCO\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'A T5 model trained on the MS MARCO dataset for generating queries from documents.\\'}', metadata={})]", "category": "generic"}
{"question_id": 256, "text": " Build a system for detecting hate speech from social media comments in Korean.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'lakahaga/novel_reading_tts\\', \\'api_call\\': \"AutoModelForTTS.from_pretrained(\\'lakahaga/novel_reading_tts\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"inputs = processor(text, return_tensors=\\'pt\\'); generated_audio = model.generate(**inputs);\", \\'performance\\': {\\'dataset\\': \\'novelspeech\\', \\'accuracy\\': None}, \\'description\\': \\'This model was trained by lakahaga using novelspeech recipe in espnet. It is designed for Korean text-to-speech tasks.\\'}', metadata={})]", "category": "generic"}
{"question_id": 257, "text": " Our team is working on a recommendation system for a news article app. We should understand the semantic similarity of various texts.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'text2text-generation\\', \\'api_name\\': \\'google/pegasus-newsroom\\', \\'api_call\\': \"pipeline(\\'summarization\\', model=\\'google/pegasus-newsroom\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'This model can be loaded on the Inference API on-demand.\\', \\'performance\\': {\\'dataset\\': \\'newsroom\\', \\'accuracy\\': \\'45.98/34.20/42.18\\'}, \\'description\\': \\'PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization by Jingqing Zhang, Yao Zhao, Mohammad Saleh, and Peter J. Liu. The model is trained on both C4 and HugeNews datasets and is designed for summarization tasks.\\'}', metadata={})]", "category": "generic"}
{"question_id": 258, "text": " We are developing an AI chatbot for a Russian company. I want to extract features from the clients' text messages in Russian language.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'Kirili4ik/mbart_ruDialogSum\\', \\'api_call\\': \"MBartForConditionalGeneration.from_pretrained(\\'Kirili4ik/mbart_ruDialogSum\\')\", \\'api_arguments\\': {\\'model_name\\': \\'Kirili4ik/mbart_ruDialogSum\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'from transformers import MBartTokenizer, MBartForConditionalGeneration\\\\nmodel_name = Kirili4ik/mbart_ruDialogSum\\\\ntokenizer = AutoTokenizer.from_pretrained(model_name)\\\\nmodel = MBartForConditionalGeneration.from_pretrained(model_name)\\\\nmodel.eval()\\\\narticle_text = ...\\\\ninput_ids = tokenizer(\\\\n [article_text],\\\\n max_length=600,\\\\n padding=max_length,\\\\n truncation=True,\\\\n return_tensors=pt,\\\\n)[input_ids]\\\\noutput_ids = model.generate(\\\\n input_ids=input_ids,\\\\n top_k=0,\\\\n num_beams=3,\\\\n no_repeat_ngram_size=3\\\\n)[0]\\\\nsummary = tokenizer.decode(output_ids, skip_special_tokens=True)\\\\nprint(summary)\\', \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'SAMSum Corpus (translated to Russian)\\', \\'accuracy\\': {\\'Validation ROGUE-1\\': 34.5, \\'Validation ROGUE-L\\': 33, \\'Test ROGUE-1\\': 31, \\'Test ROGUE-L\\': 28}}]}, \\'description\\': \\'MBart for Russian summarization fine-tuned for dialogues summarization. This model was firstly fine-tuned by Ilya Gusev on Gazeta dataset. We have fine tuned that model on SamSum dataset translated to Russian using GoogleTranslateAPI. Moreover! We have implemented a ! telegram bot @summarization_bot ! with the inference of this model. Add it to the chat and get summaries instead of dozens spam messages!\\'}', metadata={})]", "category": "generic"}
{"question_id": 259, "text": " I am an artist who wants to create a new piece of artwork based on the prompt \\\"A futuristic city under the ocean\\\".\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'dreamlike-art/dreamlike-anime-1.0\\', \\'api_call\\': \"StableDiffusionPipeline.from_pretrained(\\'dreamlike-art/dreamlike-anime-1.0\\', torch_dtype=torch.float16)(prompt, negative_prompt=negative_prompt)\", \\'api_arguments\\': [\\'prompt\\', \\'negative_prompt\\'], \\'python_environment_requirements\\': [\\'diffusers\\', \\'torch\\'], \\'example_code\\': \"from diffusers import StableDiffusionPipeline\\\\nimport torch\\\\nmodel_id = dreamlike-art/dreamlike-anime-1.0\\\\npipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16)\\\\npipe = pipe.to(cuda)\\\\nprompt = anime, masterpiece, high quality, 1girl, solo, long hair, looking at viewer, blush, smile, bangs, blue eyes, skirt, medium breasts, iridescent, gradient, colorful, besides a cottage, in the country\\\\nnegative_prompt = \\'simple background, duplicate, retro style, low quality, lowest quality, 1980s, 1990s, 2000s, 2005 2006 2007 2008 2009 2010 2011 2012 2013, bad anatomy, bad proportions, extra digits, lowres, username, artist name, error, duplicate, watermark, signature, text, extra digit, fewer digits, worst quality, jpeg artifacts, blurry\\'\\\\nimage = pipe(prompt, negative_prompt=negative_prompt).images[0]\\\\nimage.save(./result.jpg)\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \"Dreamlike Anime 1.0 is a high quality anime model, made by dreamlike.art. It can be used to generate anime-style images based on text prompts. The model is trained on 768x768px images and works best with prompts that include \\'photo anime, masterpiece, high quality, absurdres\\'. It can be used with the Stable Diffusion Pipeline from the diffusers library.\"}', metadata={})]", "category": "generic"}
{"question_id": 260, "text": " Create a character-like artwork image based on the phrase 'anime-style girl with a guitar'.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'dreamlike-art/dreamlike-anime-1.0\\', \\'api_call\\': \"StableDiffusionPipeline.from_pretrained(\\'dreamlike-art/dreamlike-anime-1.0\\', torch_dtype=torch.float16)(prompt, negative_prompt=negative_prompt)\", \\'api_arguments\\': [\\'prompt\\', \\'negative_prompt\\'], \\'python_environment_requirements\\': [\\'diffusers\\', \\'torch\\'], \\'example_code\\': \"from diffusers import StableDiffusionPipeline\\\\nimport torch\\\\nmodel_id = dreamlike-art/dreamlike-anime-1.0\\\\npipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16)\\\\npipe = pipe.to(cuda)\\\\nprompt = anime, masterpiece, high quality, 1girl, solo, long hair, looking at viewer, blush, smile, bangs, blue eyes, skirt, medium breasts, iridescent, gradient, colorful, besides a cottage, in the country\\\\nnegative_prompt = \\'simple background, duplicate, retro style, low quality, lowest quality, 1980s, 1990s, 2000s, 2005 2006 2007 2008 2009 2010 2011 2012 2013, bad anatomy, bad proportions, extra digits, lowres, username, artist name, error, duplicate, watermark, signature, text, extra digit, fewer digits, worst quality, jpeg artifacts, blurry\\'\\\\nimage = pipe(prompt, negative_prompt=negative_prompt).images[0]\\\\nimage.save(./result.jpg)\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \"Dreamlike Anime 1.0 is a high quality anime model, made by dreamlike.art. It can be used to generate anime-style images based on text prompts. The model is trained on 768x768px images and works best with prompts that include \\'photo anime, masterpiece, high quality, absurdres\\'. It can be used with the Stable Diffusion Pipeline from the diffusers library.\"}', metadata={})]", "category": "generic"}
{"question_id": 261, "text": " We are designing an application that can generate images based on textual descriptions. We need to load a model capable of performing text-to-image conversion.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'Linaqruf/anything-v3.0\\', \\'api_call\\': \"Text2ImagePipeline(model=\\'Linaqruf/anything-v3.0\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A text-to-image model that generates images from text descriptions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 262, "text": " A food delivery app wants to help users understand ingredients in the food item by analyzing the images. We need a solution to process the food images and give textual information about the items.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'Linaqruf/anything-v3.0\\', \\'api_call\\': \"Text2ImagePipeline(model=\\'Linaqruf/anything-v3.0\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A text-to-image model that generates images from text descriptions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 263, "text": " The company wants to create a social media application like Pinterest that generates captions for users' images. Show us how to create this functionality.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'Lykon/DreamShaper\\', \\'api_call\\': \"pipeline(\\'text-to-image\\', model=Lykon/DreamShaper)\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers, torch\\', \\'example_code\\': \\'https://huggingface.co/spaces/Lykon/DreamShaper-webui\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Dream Shaper is a text-to-image model that generates artistic images based on the given input text. Read more about this model here: https://civitai.com/models/4384/dreamshaper\\'}', metadata={})]", "category": "generic"}
{"question_id": 264, "text": " We are an analytics company that needs help in extracting data tables from plots and charts efficiently.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'kochetkovIT/autotrain-ironhack-49741119788\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'json\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'kochetkovIT/autotrain-data-ironhack\\', \\'accuracy\\': {\\'Loss\\': 2.603, \\'R2\\': 0.013, \\'MSE\\': 6.776, \\'MAE\\': 1.666, \\'RMSLE\\': 0.502}}, \\'description\\': \\'A tabular regression model trained using AutoTrain to predict carbon emissions (in grams) with an R2 score of 0.013.\\'}', metadata={})]", "category": "generic"}
{"question_id": 265, "text": " Our marketing team wants a video commercial for our product. They are looking for a video with a couple sitting in a cafe and laughing while using our product.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'chavinlo/TempoFunk\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'chavinlo/TempoFunk\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Text-to-Video model using Hugging Face Transformers library. Model is capable of generating video content based on the input text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 266, "text": " A company is having an ad campaign with a superhero theme. They want to generate a video of Spiderman surfing to showcase as a social media ad.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'chavinlo/TempoFunk\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'chavinlo/TempoFunk\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Text-to-Video model using Hugging Face Transformers library. Model is capable of generating video content based on the input text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 267, "text": " We are developing a solution where we can visually detect the medical conditions of patients. Can you write code to determine the condition based on the image provided?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'martinezomg/vit-base-patch16-224-diabetic-retinopathy\\', \\'api_call\\': \"pipeline(\\'image-classification\\', \\'martinezomg/vit-base-patch16-224-diabetic-retinopathy\\')\", \\'api_arguments\\': {\\'model_name\\': \\'martinezomg/vit-base-patch16-224-diabetic-retinopathy\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'4.28.1\\', \\'pytorch\\': \\'2.0.0+cu118\\', \\'datasets\\': \\'2.11.0\\', \\'tokenizers\\': \\'0.13.3\\'}, \\'example_code\\': \"from transformers import pipeline\\\\nimage_classifier = pipeline(\\'image-classification\\', \\'martinezomg/vit-base-patch16-224-diabetic-retinopathy\\')\\\\nresult = image_classifier(\\'path/to/image.jpg\\')\", \\'performance\\': {\\'dataset\\': \\'None\\', \\'accuracy\\': 0.7744}, \\'description\\': \\'This model is a fine-tuned version of google/vit-base-patch16-224 on the None dataset. It is designed for image classification tasks, specifically for diabetic retinopathy detection.\\'}', metadata={})]", "category": "generic"}
{"question_id": 268, "text": " Help me to build a visual-question-answering model that takes a question and an image as input and returns an answer based on the image.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 269, "text": " Implement an optical text recognition system for documents that can answer a simple question about the document's content.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 270, "text": " Our company deals with insurance claims. We need a smart assistant who can retrieve information from documents, especially invoices, such as total amount, date of invoice, and name of the service provider.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'layoutlm-invoices\\', \\'api_call\\': \"AutoModelForDocumentQuestionAnswering.from_pretrained(\\'impira/layoutlm-invoices\\')\", \\'api_arguments\\': \\'question, context\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"nlp(question=\\'What is the total amount?\\', context=\\'your_invoice_text\\')\", \\'performance\\': {\\'dataset\\': \\'proprietary dataset of invoices, SQuAD2.0, and DocVQA\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'A fine-tuned version of the multi-modal LayoutLM model for the task of question answering on invoices and other documents. It has been fine-tuned on a proprietary dataset of invoices as well as both SQuAD2.0 and DocVQA for general comprehension. Unlike other QA models, which can only extract consecutive tokens, this model can predict longer-range, non-consecutive sequences with an additional classifier head.\\'}', metadata={})]", "category": "generic"}
{"question_id": 271, "text": " We have an AI-powered document management system that needs to answer questions based on the content of a given document.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 272, "text": " Take a look at this document image and tell me the answer to my question: \\\"What is the total amount due?\\\".\\n###Input: {\\\"image_url\\\": \\\"https://example.com/document_invoice.jpg\\\", \\\"question\\\": \\\"What is the total amount due?\\\"}\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 273, "text": " I am designing a parking spot detector for a car parking management solution. I want the model to estimage the depth of the car park.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Depth Estimation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'glpn-nyu-finetuned-diode-221215-092352\\', \\'api_call\\': \"AutoModel.from_pretrained(\\'sayakpaul/glpn-nyu-finetuned-diode-221215-092352\\')\", \\'api_arguments\\': {}, \\'python_environment_requirements\\': {\\'huggingface_transformers\\': \\'4.13.0\\'}, \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'DIODE\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A depth estimation model fine-tuned on the DIODE dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 274, "text": " We are researching multi-view 3D scanning, and we would like to improve our depth estimation with a pretrained model.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Depth Estimation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'dpt-large-redesign\\', \\'api_call\\': \"AutoModelForDepthEstimation.from_pretrained(\\'nielsr/dpt-large-redesign\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'torch\\', \\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A depth estimation model based on the DPT architecture.\\'}', metadata={})]", "category": "generic"}
{"question_id": 275, "text": " I am building an app to detect objects in an image from a URL. How can I experiment with object detection models using transformers?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Object Detection\\', \\'api_name\\': \\'facebook/detr-resnet-101\\', \\'api_call\\': \"DetrForObjectDetection.from_pretrained(\\'facebook/detr-resnet-101\\')\", \\'api_arguments\\': [\\'image\\'], \\'python_environment_requirements\\': [\\'transformers\\', \\'torch\\', \\'PIL\\', \\'requests\\'], \\'example_code\\': \\'from transformers import DetrImageProcessor, DetrForObjectDetection\\\\nimport torch\\\\nfrom PIL import Image\\\\nimport requests\\\\nurl = http://images.cocodataset.org/val2017/000000039769.jpg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nprocessor = DetrImageProcessor.from_pretrained(facebook/detr-resnet-101)\\\\nmodel = DetrForObjectDetection.from_pretrained(facebook/detr-resnet-101)\\\\ninputs = processor(images=image, return_tensors=pt)\\\\noutputs = model(**inputs)\\', \\'performance\\': {\\'dataset\\': \\'COCO 2017\\', \\'accuracy\\': \\'43.5 AP\\'}, \\'description\\': \\'DEtection TRansformer (DETR) model trained end-to-end on COCO 2017 object detection (118k annotated images). It was introduced in the paper End-to-End Object Detection with Transformers by Carion et al. and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 276, "text": " I have a picture of my backyard, but I am not sure if a cat sneaked in. Can you help me find out if it is a cat in the picture?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Unconditional Image Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Denoising Diffusion Probabilistic Models (DDPM)\\', \\'api_name\\': \\'google/ddpm-ema-cat-256\\', \\'api_call\\': \"DDPMPipeline.from_pretrained(\\'google/ddpm-ema-cat-256\\')\", \\'api_arguments\\': [\\'model_id\\'], \\'python_environment_requirements\\': [\\'!pip install diffusers\\'], \\'example_code\\': \\'from diffusers import DDPMPipeline, DDIMPipeline, PNDMPipeline\\\\nmodel_id = google/ddpm-ema-cat-256\\\\nddpm = DDPMPipeline.from_pretrained(model_id)\\\\nimage = ddpm().images[0]\\\\nimage.save(ddpm_generated_image.png)\\', \\'performance\\': {\\'dataset\\': \\'CIFAR10\\', \\'accuracy\\': {\\'Inception_score\\': 9.46, \\'FID_score\\': 3.17}}, \\'description\\': \\'Denoising Diffusion Probabilistic Models (DDPM) is a class of latent variable models inspired by considerations from nonequilibrium thermodynamics. It can generate high-quality images, and supports different noise schedulers such as scheduling_ddpm, scheduling_ddim, and scheduling_pndm. On the unconditional CIFAR10 dataset, it achieves an Inception score of 9.46 and a state-of-the-art FID score of 3.17.\\'}', metadata={})]", "category": "generic"}
{"question_id": 277, "text": " Our hospital needs to analyze digital blood samples in order to detect and count platelets, red blood cells, and white blood cells.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Blood Cell Detection\\', \\'api_name\\': \\'keremberke/yolov8n-blood-cell-detection\\', \\'api_call\\': \"YOLO(\\'keremberke/yolov8n-blood-cell-detection\\')\", \\'api_arguments\\': {\\'conf\\': 0.25, \\'iou\\': 0.45, \\'agnostic_nms\\': False, \\'max_det\\': 1000}, \\'python_environment_requirements\\': \\'ultralyticsplus==0.0.23 ultralytics==8.0.21\\', \\'example_code\\': \"from ultralyticsplus import YOLO, render_result\\\\nmodel = YOLO(\\'keremberke/yolov8n-blood-cell-detection\\')\\\\nmodel.overrides[\\'conf\\'] = 0.25\\\\nmodel.overrides[\\'iou\\'] = 0.45\\\\nmodel.overrides[\\'agnostic_nms\\'] = False\\\\nmodel.overrides[\\'max_det\\'] = 1000\\\\nimage = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\\\\nresults = model.predict(image)\\\\nprint(results[0].boxes)\\\\nrender = render_result(model=model, image=image, result=results[0])\\\\nrender.show()\", \\'performance\\': {\\'dataset\\': \\'blood-cell-object-detection\\', \\'accuracy\\': 0.893}, \\'description\\': \\'This model detects blood cells in images, specifically Platelets, RBC, and WBC. It is based on the YOLOv8 architecture and trained on the blood-cell-object-detection dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 278, "text": " Develop a script that extracts a table from a given document and then visualize the results.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Table Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Table Question Answering\\', \\'api_name\\': \\'google/tapas-large-finetuned-wikisql-supervised\\', \\'api_call\\': \"pipeline(\\'table-question-answering\\', model=\\'google/tapas-large-finetuned-wikisql-supervised\\')\", \\'api_arguments\\': {\\'model\\': \\'google/tapas-large-finetuned-wikisql-supervised\\'}, \\'python_environment_requirements\\': [\\'transformers\\', \\'torch\\', \\'tensorflow\\'], \\'example_code\\': \"from transformers import pipeline\\\\n\\\\nqa_pipeline = pipeline(\\'table-question-answering\\', model=\\'google/tapas-large-finetuned-wikisql-supervised\\')\\\\n\\\\nresult = qa_pipeline(question=\\'What is the capital of France?\\', table=table)\\\\nprint(result)\", \\'performance\\': {\\'dataset\\': \\'wikisql\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'TAPAS is a BERT-like transformers model pretrained on a large corpus of English data from Wikipedia in a self-supervised fashion. It can be used for answering questions related to a table.\\'}', metadata={})]", "category": "generic"}
{"question_id": 279, "text": " We are now working on an architectural image generation project to generate different images of architecture.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Unconditional Image Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Unconditional Image Generation\\', \\'api_name\\': \\'Apocalypse-19/shoe-generator\\', \\'api_call\\': \"DDPMPipeline.from_pretrained(\\'Apocalypse-19/shoe-generator\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'diffusers\\'], \\'example_code\\': \"from diffusers import DDPMPipeline\\\\npipeline = DDPMPipeline.from_pretrained(\\'Apocalypse-19/shoe-generator\\')\\\\nimage = pipeline().images[0]\\\\nimage\", \\'performance\\': {\\'dataset\\': \\'custom dataset\\', \\'accuracy\\': \\'128x128 resolution\\'}, \\'description\\': \\'This model is a diffusion model for unconditional image generation of shoes trained on a custom dataset at 128x128 resolution.\\'}', metadata={})]", "category": "generic"}
{"question_id": 280, "text": " We are designing a book cover where the book is about love and roses. We have the idea of \\\"A head full of roses.\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Image-to-Text\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'kha-white/manga-ocr-base\\', \\'api_call\\': \"pipeline(\\'ocr\\', model=\\'kha-white/manga-ocr-base\\')\", \\'api_arguments\\': \\'image\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'manga109s\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Optical character recognition for Japanese text, with the main focus being Japanese manga. It uses Vision Encoder Decoder framework. Manga OCR can be used as a general purpose printed Japanese OCR, but its main goal was to provide a high quality text recognition, robust against various scenarios specific to manga: both vertical and horizontal text, text with furigana, text overlaid on images, wide variety of fonts and font styles, and low quality images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 281, "text": " Help me generate a realistic bedroom interior image that can be used as reference for a 3D model being created for a virtual reality game.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Unconditional Image Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Denoising Diffusion Probabilistic Models (DDPM)\\', \\'api_name\\': \\'google/ddpm-bedroom-256\\', \\'api_call\\': \"DDPMPipeline.from_pretrained(\\'google/ddpm-bedroom-256\\')\", \\'api_arguments\\': \\'None\\', \\'python_environment_requirements\\': \\'diffusers\\', \\'example_code\\': \\'!pip install diffusers\\\\nfrom diffusers import DDPMPipeline, DDIMPipeline, PNDMPipeline\\\\nmodel_id = google/ddpm-bedroom-256\\\\nddpm = DDPMPipeline.from_pretrained(model_id)\\\\nimage = ddpm().images[0]\\\\nimage.save(ddpm_generated_image.png)\\', \\'performance\\': {\\'dataset\\': \\'CIFAR10\\', \\'accuracy\\': {\\'Inception score\\': 9.46, \\'FID score\\': 3.17}}, \\'description\\': \\'We present high quality image synthesis results using diffusion probabilistic models, a class of latent variable models inspired by considerations from nonequilibrium thermodynamics. Our best results are obtained by training on a weighted variational bound designed according to a novel connection between diffusion probabilistic models and denoising score matching with Langevin dynamics, and our models naturally admit a progressive lossy decompression scheme that can be interpreted as a generalization of autoregressive decoding. On the unconditional CIFAR10 dataset, we obtain an Inception score of 9.46 and a state-of-the-art FID score of 3.17. On 256x256 LSUN, we obtain sample quality similar to ProgressiveGAN.\\'}', metadata={})]", "category": "generic"}
{"question_id": 282, "text": " You feel lonely and need to generate a picture of a butterfly. How would you proceed?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Unconditional Image Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Unconditional Image Generation\\', \\'api_name\\': \\'ntrant7/sd-class-butterflies-32\\', \\'api_call\\': \"DDPMPipeline.from_pretrained(\\'ntrant7/sd-class-butterflies-32\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'diffusers\\'], \\'example_code\\': \"from diffusers import DDPMPipeline\\\\npipeline = DDPMPipeline.from_pretrained(\\'ntrant7/sd-class-butterflies-32\\')\\\\nimage = pipeline().images[0]\\\\nimage\", \\'performance\\': {\\'dataset\\': \\'Not specified\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'This model is a diffusion model for unconditional image generation of cute butterflies.\\'}', metadata={})]", "category": "generic"}
{"question_id": 283, "text": " We need an AI-generated insect image for our biology article about African habitats.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Unconditional Image Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Unconditional Image Generation\\', \\'api_name\\': \\'ntrant7/sd-class-butterflies-32\\', \\'api_call\\': \"DDPMPipeline.from_pretrained(\\'ntrant7/sd-class-butterflies-32\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'diffusers\\'], \\'example_code\\': \"from diffusers import DDPMPipeline\\\\npipeline = DDPMPipeline.from_pretrained(\\'ntrant7/sd-class-butterflies-32\\')\\\\nimage = pipeline().images[0]\\\\nimage\", \\'performance\\': {\\'dataset\\': \\'Not specified\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'This model is a diffusion model for unconditional image generation of cute butterflies.\\'}', metadata={})]", "category": "generic"}
{"question_id": 284, "text": " Our client is involved in a sports organization, and they require a solution to classify sports videos efficiently.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Video Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Video Classification\\', \\'api_name\\': \\'facebook/timesformer-base-finetuned-ssv2\\', \\'api_call\\': \"TimesformerForVideoClassification.from_pretrained(\\'facebook/timesformer-base-finetuned-ssv2\\')\", \\'api_arguments\\': [\\'images\\', \\'return_tensors\\'], \\'python_environment_requirements\\': [\\'transformers\\', \\'numpy\\', \\'torch\\'], \\'example_code\\': \\'from transformers import AutoImageProcessor, TimesformerForVideoClassification\\\\nimport numpy as np\\\\nimport torch\\\\nvideo = list(np.random.randn(8, 3, 224, 224))\\\\nprocessor = AutoImageProcessor.from_pretrained(facebook/timesformer-base-finetuned-ssv2)\\\\nmodel = TimesformerForVideoClassification.from_pretrained(facebook/timesformer-base-finetuned-ssv2)\\\\ninputs = processor(images=video, return_tensors=pt)\\\\nwith torch.no_grad():\\\\n outputs = model(**inputs)\\\\n logits = outputs.logits\\\\npredicted_class_idx = logits.argmax(-1).item()\\\\nprint(Predicted class:, model.config.id2label[predicted_class_idx])\\', \\'performance\\': {\\'dataset\\': \\'Something Something v2\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'TimeSformer model pre-trained on Something Something v2. It was introduced in the paper TimeSformer: Is Space-Time Attention All You Need for Video Understanding? by Tong et al. and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 285, "text": " We need to create a content filter for images submitted by users in our online community. We want to detect and filter out adult content and offensive images.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'michellejieli/NSFW_text_classifier\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'michellejieli/NSFW_text_classification\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'classifier(I see you\u2019ve set aside this special time to humiliate yourself in public.)\\', \\'performance\\': {\\'dataset\\': \\'Reddit posts\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'DistilBERT is a transformer model that performs sentiment analysis. I fine-tuned the model on Reddit posts with the purpose of classifying not safe for work (NSFW) content, specifically text that is considered inappropriate and unprofessional. The model predicts 2 classes, which are NSFW or safe for work (SFW). The model is a fine-tuned version of DistilBERT. It was fine-tuned on 14317 Reddit posts pulled from the Reddit API.\\'}', metadata={})]", "category": "generic"}
{"question_id": 286, "text": " We are an animal rescue organization, and we are looking for a way to automatically identify if images uploaded to our site contain cats, dogs, or birds.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'abhishek/autotrain-dog-vs-food\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'abhishek/autotrain-dog-vs-food\\')\", \\'api_arguments\\': \\'image_path\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'sasha/dog-food\\', \\'accuracy\\': 0.998}, \\'description\\': \"A pre-trained model for classifying images as either dog or food using Hugging Face\\'s AutoTrain framework.\"}', metadata={})]", "category": "generic"}
{"question_id": 287, "text": " Create a machine learning-based image recognition tool that can identify whether an animal in an image is a cat or a dog.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'abhishek/autotrain-dog-vs-food\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'abhishek/autotrain-dog-vs-food\\')\", \\'api_arguments\\': \\'image_path\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'sasha/dog-food\\', \\'accuracy\\': 0.998}, \\'description\\': \"A pre-trained model for classifying images as either dog or food using Hugging Face\\'s AutoTrain framework.\"}', metadata={})]", "category": "generic"}
{"question_id": 288, "text": " I want to find out whether a given review is positive or negative. I need an API that can help me with sentiment analysis.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'bert-base-multilingual-uncased-sentiment\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'nlptown/bert-base-multilingual-uncased-sentiment\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"result = sentiment_pipeline(\\'I love this product!\\')\", \\'performance\\': {\\'dataset\\': [{\\'language\\': \\'English\\', \\'accuracy\\': {\\'exact\\': \\'67%\\', \\'off-by-1\\': \\'95%\\'}}, {\\'language\\': \\'Dutch\\', \\'accuracy\\': {\\'exact\\': \\'57%\\', \\'off-by-1\\': \\'93%\\'}}, {\\'language\\': \\'German\\', \\'accuracy\\': {\\'exact\\': \\'61%\\', \\'off-by-1\\': \\'94%\\'}}, {\\'language\\': \\'French\\', \\'accuracy\\': {\\'exact\\': \\'59%\\', \\'off-by-1\\': \\'94%\\'}}, {\\'language\\': \\'Italian\\', \\'accuracy\\': {\\'exact\\': \\'59%\\', \\'off-by-1\\': \\'95%\\'}}, {\\'language\\': \\'Spanish\\', \\'accuracy\\': {\\'exact\\': \\'58%\\', \\'off-by-1\\': \\'95%\\'}}]}, \\'description\\': \\'This a bert-base-multilingual-uncased model finetuned for sentiment analysis on product reviews in six languages: English, Dutch, German, French, Spanish and Italian. It predicts the sentiment of the review as a number of stars (between 1 and 5).\\'}', metadata={})]", "category": "generic"}
{"question_id": 289, "text": " I am a traveler and I want to know the best time to visit a Chinese historical site. Take a look at an image and tell me whether it is a good time to visit the site or not.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Zero-Shot Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Zero-Shot Image Classification\\', \\'api_name\\': \\'OFA-Sys/chinese-clip-vit-large-patch14-336px\\', \\'api_call\\': \"ChineseCLIPModel.from_pretrained(\\'OFA-Sys/chinese-clip-vit-large-patch14-336px\\')\", \\'api_arguments\\': {\\'images\\': \\'image\\', \\'text\\': \\'texts\\', \\'return_tensors\\': \\'pt\\', \\'padding\\': \\'True\\'}, \\'python_environment_requirements\\': [\\'PIL\\', \\'requests\\', \\'transformers\\'], \\'example_code\\': \\'from PIL import Image\\\\nimport requests\\\\nfrom transformers import ChineseCLIPProcessor, ChineseCLIPModel\\\\nmodel = ChineseCLIPModel.from_pretrained(OFA-Sys/chinese-clip-vit-large-patch14-336px)\\\\nprocessor = ChineseCLIPProcessor.from_pretrained(OFA-Sys/chinese-clip-vit-large-patch14-336px)\\\\nurl = https://clip-cn-beijing.oss-cn-beijing.aliyuncs.com/pokemon.jpeg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\ntexts = []\\\\ninputs = processor(images=image, return_tensors=pt)\\\\nimage_features = model.get_image_features(**inputs)\\\\nimage_features = image_features / image_features.norm(p=2, dim=-1, keepdim=True)\\\\ninputs = processor(text=texts, padding=True, return_tensors=pt)\\\\ntext_features = model.get_text_features(**inputs)\\\\ntext_features = text_features / text_features.norm(p=2, dim=-1, keepdim=True)\\\\ninputs = processor(text=texts, images=image, return_tensors=pt, padding=True)\\\\noutputs = model(**inputs)\\\\nlogits_per_image = outputs.logits_per_image\\\\nprobs = logits_per_image.softmax(dim=1)\\', \\'performance\\': {\\'dataset\\': {\\'CIFAR10\\': 96.0, \\'CIFAR100\\': 79.75, \\'DTD\\': 51.2, \\'EuroSAT\\': 52.0, \\'FER\\': 55.1, \\'FGVC\\': 26.2, \\'KITTI\\': 49.9, \\'MNIST\\': 79.4, \\'PC\\': 63.5, \\'VOC\\': 84.9}, \\'accuracy\\': \\'various\\'}, \\'description\\': \\'Chinese CLIP is a simple implementation of CLIP on a large-scale dataset of around 200 million Chinese image-text pairs. It uses ViT-L/14@336px as the image encoder and RoBERTa-wwm-base as the text encoder.\\'}', metadata={})]", "category": "generic"}
{"question_id": 290, "text": " Organizing a special event and want to identify the place where street pictures were taken for invitations.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Zero-Shot Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Geolocalization\\', \\'api_name\\': \\'geolocal/StreetCLIP\\', \\'api_call\\': \"CLIPModel.from_pretrained(\\'geolocal/StreetCLIP\\')\", \\'api_arguments\\': {\\'pretrained_model_name_or_path\\': \\'geolocal/StreetCLIP\\'}, \\'python_environment_requirements\\': [\\'transformers\\', \\'PIL\\', \\'requests\\'], \\'example_code\\': \\'from PIL import Image\\\\nimport requests\\\\nfrom transformers import CLIPProcessor, CLIPModel\\\\nmodel = CLIPModel.from_pretrained(geolocal/StreetCLIP)\\\\nprocessor = CLIPProcessor.from_pretrained(geolocal/StreetCLIP)\\\\nurl = https://huggingface.co/geolocal/StreetCLIP/resolve/main/sanfrancisco.jpeg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nchoices = [San Jose, San Diego, Los Angeles, Las Vegas, San Francisco]\\\\ninputs = processor(text=choices, images=image, return_tensors=pt, padding=True)\\\\noutputs = model(**inputs)\\\\nlogits_per_image = outputs.logits_per_image\\\\nprobs = logits_per_image.softmax(dim=1)\\', \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'IM2GPS\\', \\'accuracy\\': {\\'25km\\': 28.3, \\'200km\\': 45.1, \\'750km\\': 74.7, \\'2500km\\': 88.2}}, {\\'name\\': \\'IM2GPS3K\\', \\'accuracy\\': {\\'25km\\': 22.4, \\'200km\\': 37.4, \\'750km\\': 61.3, \\'2500km\\': 80.4}}]}, \\'description\\': \\'StreetCLIP is a robust foundation model for open-domain image geolocalization and other geographic and climate-related tasks. Trained on an original dataset of 1.1 million street-level urban and rural geo-tagged images, it achieves state-of-the-art performance on multiple open-domain image geolocalization benchmarks in zero-shot, outperforming supervised models trained on millions of images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 291, "text": " Our client wants us to monitor consumer sentiment on their e-commerce platform. We must identify low-rated product reviews in six languages: English, Dutch, German, French, Italian, and Spanish.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'bert-base-multilingual-uncased-sentiment\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'nlptown/bert-base-multilingual-uncased-sentiment\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"result = sentiment_pipeline(\\'I love this product!\\')\", \\'performance\\': {\\'dataset\\': [{\\'language\\': \\'English\\', \\'accuracy\\': {\\'exact\\': \\'67%\\', \\'off-by-1\\': \\'95%\\'}}, {\\'language\\': \\'Dutch\\', \\'accuracy\\': {\\'exact\\': \\'57%\\', \\'off-by-1\\': \\'93%\\'}}, {\\'language\\': \\'German\\', \\'accuracy\\': {\\'exact\\': \\'61%\\', \\'off-by-1\\': \\'94%\\'}}, {\\'language\\': \\'French\\', \\'accuracy\\': {\\'exact\\': \\'59%\\', \\'off-by-1\\': \\'94%\\'}}, {\\'language\\': \\'Italian\\', \\'accuracy\\': {\\'exact\\': \\'59%\\', \\'off-by-1\\': \\'95%\\'}}, {\\'language\\': \\'Spanish\\', \\'accuracy\\': {\\'exact\\': \\'58%\\', \\'off-by-1\\': \\'95%\\'}}]}, \\'description\\': \\'This a bert-base-multilingual-uncased model finetuned for sentiment analysis on product reviews in six languages: English, Dutch, German, French, Spanish and Italian. It predicts the sentiment of the review as a number of stars (between 1 and 5).\\'}', metadata={})]", "category": "generic"}
{"question_id": 292, "text": " Write a script for an online forum moderator that will help them to detect gibberish text in a post.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'madhurjindal/autonlp-Gibberish-Detector-492513457\\', \\'api_call\\': \"AutoModelForSequenceClassification.from_pretrained(\\'madhurjindal/autonlp-Gibberish-Detector-492513457\\')\", \\'api_arguments\\': {\\'inputs\\': \\'I love AutoNLP\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'AutoModelForSequenceClassification\\', \\'AutoTokenizer\\': \\'from_pretrained\\'}, \\'example_code\\': \\'from transformers import AutoModelForSequenceClassification, AutoTokenizer\\\\nmodel = AutoModelForSequenceClassification.from_pretrained(madhurjindal/autonlp-Gibberish-Detector-492513457, use_auth_token=True)\\\\ntokenizer = AutoTokenizer.from_pretrained(madhurjindal/autonlp-Gibberish-Detector-492513457, use_auth_token=True)\\\\ninputs = tokenizer(I love AutoNLP, return_tensors=pt)\\\\noutputs = model(**inputs)\\', \\'performance\\': {\\'dataset\\': \\'madhurjindal/autonlp-data-Gibberish-Detector\\', \\'accuracy\\': 0.9735624586913417}, \\'description\\': \\'A multi-class text classification model for detecting gibberish text. Trained using AutoNLP and DistilBERT.\\'}', metadata={})]", "category": "generic"}
{"question_id": 293, "text": " As a restaurant owner, I want to know if the total revenue for last week met our target revenue.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'text2text-generation\\', \\'api_name\\': \\'tuner007/pegasus_summarizer\\', \\'api_call\\': \"PegasusForConditionalGeneration.from_pretrained(\\'tuner007/pegasus_summarizer\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'pip install sentencepiece\\'], \\'example_code\\': \"context = \\\\nIndia wicket-keeper batsman Rishabh Pant has said someone from the crowd threw a ball on pacer Mohammed Siraj while he was fielding in the ongoing third Test against England on Wednesday. Pant revealed the incident made India skipper Virat Kohli upset. I think, somebody threw a ball inside, at Siraj, so he [Kohli] was upset, said Pant in a virtual press conference after the close of the first day\\'s play.You can say whatever you want to chant, but don\\'t throw things at the fielders and all those things. It is not good for cricket, I guess, he added.In the third session of the opening day of the third Test, a section of spectators seemed to have asked Siraj the score of the match to tease the pacer. The India pacer however came with a brilliant reply as he gestured 1-0 (India leading the Test series) towards the crowd.Earlier this month, during the second Test match, there was some bad crowd behaviour on a show as some unruly fans threw champagne corks at India batsman KL Rahul.Kohli also intervened and he was seen gesturing towards the opening batsman to know more about the incident. An over later, the TV visuals showed that many champagne corks were thrown inside the playing field, and the Indian players were visibly left frustrated.Coming back to the game, after bundling out India for 78, openers Rory Burns and Haseeb Hameed ensured that England took the honours on the opening day of the ongoing third Test.At stumps, England\\'s score reads 120/0 and the hosts have extended their lead to 42 runs. For the Three Lions, Burns (52) and Hameed (60) are currently unbeaten at the crease.Talking about the pitch on opening day, Pant said, They took the heavy roller, the wicket was much more settled down, and they batted nicely also, he said. But when we batted, the wicket was slightly soft, and they bowled in good areas, but we could have applied [ourselves] much better.Both England batsmen managed to see off the final session and the hosts concluded the opening day with all ten wickets intact, extending the lead to 42.(ANI)\\\\n\\\\nget_response(context)\", \\'performance\\': {\\'dataset\\': \\'cnn_dailymail\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.604, \\'ROUGE-2\\': 14.64, \\'ROUGE-L\\': 23.884, \\'ROUGE-LSUM\\': 32.902, \\'loss\\': 2.576, \\'gen_len\\': 76.398}}, \\'description\\': \\'PEGASUS fine-tuned for summarization\\'}', metadata={})]", "category": "generic"}
{"question_id": 294, "text": " We are a sports analysis company that processes data from spreadsheets with game statistics. We need to identify the player who has scored the maximum goals in a given match.\\n###Input: \\\"What player scored the most goals?\\\", \\\"Player,Goals\\\\nA,2\\\\nB,3\\\\nC,1\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'text2text-generation\\', \\'api_name\\': \\'tuner007/pegasus_summarizer\\', \\'api_call\\': \"PegasusForConditionalGeneration.from_pretrained(\\'tuner007/pegasus_summarizer\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'pip install sentencepiece\\'], \\'example_code\\': \"context = \\\\nIndia wicket-keeper batsman Rishabh Pant has said someone from the crowd threw a ball on pacer Mohammed Siraj while he was fielding in the ongoing third Test against England on Wednesday. Pant revealed the incident made India skipper Virat Kohli upset. I think, somebody threw a ball inside, at Siraj, so he [Kohli] was upset, said Pant in a virtual press conference after the close of the first day\\'s play.You can say whatever you want to chant, but don\\'t throw things at the fielders and all those things. It is not good for cricket, I guess, he added.In the third session of the opening day of the third Test, a section of spectators seemed to have asked Siraj the score of the match to tease the pacer. The India pacer however came with a brilliant reply as he gestured 1-0 (India leading the Test series) towards the crowd.Earlier this month, during the second Test match, there was some bad crowd behaviour on a show as some unruly fans threw champagne corks at India batsman KL Rahul.Kohli also intervened and he was seen gesturing towards the opening batsman to know more about the incident. An over later, the TV visuals showed that many champagne corks were thrown inside the playing field, and the Indian players were visibly left frustrated.Coming back to the game, after bundling out India for 78, openers Rory Burns and Haseeb Hameed ensured that England took the honours on the opening day of the ongoing third Test.At stumps, England\\'s score reads 120/0 and the hosts have extended their lead to 42 runs. For the Three Lions, Burns (52) and Hameed (60) are currently unbeaten at the crease.Talking about the pitch on opening day, Pant said, They took the heavy roller, the wicket was much more settled down, and they batted nicely also, he said. But when we batted, the wicket was slightly soft, and they bowled in good areas, but we could have applied [ourselves] much better.Both England batsmen managed to see off the final session and the hosts concluded the opening day with all ten wickets intact, extending the lead to 42.(ANI)\\\\n\\\\nget_response(context)\", \\'performance\\': {\\'dataset\\': \\'cnn_dailymail\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.604, \\'ROUGE-2\\': 14.64, \\'ROUGE-L\\': 23.884, \\'ROUGE-LSUM\\': 32.902, \\'loss\\': 2.576, \\'gen_len\\': 76.398}}, \\'description\\': \\'PEGASUS fine-tuned for summarization\\'}', metadata={})]", "category": "generic"}
{"question_id": 295, "text": " A manager in our company requires an aggregated answer of the highest and lowest sales numbers for a given period to evaluate the perfomance.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'text2text-generation\\', \\'api_name\\': \\'tuner007/pegasus_summarizer\\', \\'api_call\\': \"PegasusForConditionalGeneration.from_pretrained(\\'tuner007/pegasus_summarizer\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'pip install sentencepiece\\'], \\'example_code\\': \"context = \\\\nIndia wicket-keeper batsman Rishabh Pant has said someone from the crowd threw a ball on pacer Mohammed Siraj while he was fielding in the ongoing third Test against England on Wednesday. Pant revealed the incident made India skipper Virat Kohli upset. I think, somebody threw a ball inside, at Siraj, so he [Kohli] was upset, said Pant in a virtual press conference after the close of the first day\\'s play.You can say whatever you want to chant, but don\\'t throw things at the fielders and all those things. It is not good for cricket, I guess, he added.In the third session of the opening day of the third Test, a section of spectators seemed to have asked Siraj the score of the match to tease the pacer. The India pacer however came with a brilliant reply as he gestured 1-0 (India leading the Test series) towards the crowd.Earlier this month, during the second Test match, there was some bad crowd behaviour on a show as some unruly fans threw champagne corks at India batsman KL Rahul.Kohli also intervened and he was seen gesturing towards the opening batsman to know more about the incident. An over later, the TV visuals showed that many champagne corks were thrown inside the playing field, and the Indian players were visibly left frustrated.Coming back to the game, after bundling out India for 78, openers Rory Burns and Haseeb Hameed ensured that England took the honours on the opening day of the ongoing third Test.At stumps, England\\'s score reads 120/0 and the hosts have extended their lead to 42 runs. For the Three Lions, Burns (52) and Hameed (60) are currently unbeaten at the crease.Talking about the pitch on opening day, Pant said, They took the heavy roller, the wicket was much more settled down, and they batted nicely also, he said. But when we batted, the wicket was slightly soft, and they bowled in good areas, but we could have applied [ourselves] much better.Both England batsmen managed to see off the final session and the hosts concluded the opening day with all ten wickets intact, extending the lead to 42.(ANI)\\\\n\\\\nget_response(context)\", \\'performance\\': {\\'dataset\\': \\'cnn_dailymail\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.604, \\'ROUGE-2\\': 14.64, \\'ROUGE-L\\': 23.884, \\'ROUGE-LSUM\\': 32.902, \\'loss\\': 2.576, \\'gen_len\\': 76.398}}, \\'description\\': \\'PEGASUS fine-tuned for summarization\\'}', metadata={})]", "category": "generic"}
{"question_id": 296, "text": " A doctor needs to find medical information in a large document. We are building an AI model to help them extract answers to their questions. \\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 297, "text": " We are building a news classification system, and for Spanish articles, we have to categorize them into different sections.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'finiteautomata/beto-sentiment-analysis\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'finiteautomata/beto-sentiment-analysis\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'Hugging Face Transformers library\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'TASS 2020 corpus\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Model trained with TASS 2020 corpus (around ~5k tweets) of several dialects of Spanish. Base model is BETO, a BERT model trained in Spanish. Uses POS, NEG, NEU labels.\\'}', metadata={})]", "category": "generic"}
{"question_id": 298, "text": " In my project, I need to classify news articles into their respective categories, such as technology, sports, and politics, without providing any labeled training data.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'michellejieli/NSFW_text_classifier\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'michellejieli/NSFW_text_classification\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'classifier(I see you\u2019ve set aside this special time to humiliate yourself in public.)\\', \\'performance\\': {\\'dataset\\': \\'Reddit posts\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'DistilBERT is a transformer model that performs sentiment analysis. I fine-tuned the model on Reddit posts with the purpose of classifying not safe for work (NSFW) content, specifically text that is considered inappropriate and unprofessional. The model predicts 2 classes, which are NSFW or safe for work (SFW). The model is a fine-tuned version of DistilBERT. It was fine-tuned on 14317 Reddit posts pulled from the Reddit API.\\'}', metadata={})]", "category": "generic"}
{"question_id": 299, "text": " I have a Catalan language text and I need to publish it as a blog article in Spanish, so I want to translate my Catalan text into Spanish.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text2Text Generation\\', \\'api_name\\': \\'mrm8488/bert2bert_shared-spanish-finetuned-summarization\\', \\'api_call\\': \"AutoModelForSeq2SeqLM.from_pretrained(\\'mrm8488/bert2bert_shared-spanish-finetuned-summarization\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'torch\\', \\'transformers\\'], \\'example_code\\': \\'def generate_summary(text):\\\\n  inputs = tokenizer([text], padding=max_length, truncation=True, max_length=512, return_tensors=pt)\\\\n  input_ids = inputs.input_ids.to(device)\\\\n  attention_mask = inputs.attention_mask.to(device)\\\\n  output = model.generate(input_ids, attention_mask=attention_mask)\\\\n  return tokenizer.decode(output[0], skip_special_tokens=True)\\\\ntext = Your text here...\\\\ngenerate_summary(text)\\', \\'performance\\': {\\'dataset\\': \\'mlsum\\', \\'accuracy\\': {\\'Rouge1\\': 26.24, \\'Rouge2\\': 8.9, \\'RougeL\\': 21.01, \\'RougeLsum\\': 21.02}}, \\'description\\': \\'Spanish BERT2BERT (BETO) fine-tuned on MLSUM ES for summarization\\'}', metadata={})]", "category": "generic"}
{"question_id": 300, "text": " Generate an automatic summarization of an executive's meeting notes without sharing too much details.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'it5-base-news-summarization\\', \\'api_call\\': \"pipeline(\\'summarization\\', model=\\'it5/it5-base-news-summarization\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'newsum(Dal 31 maggio \u00e8 infine partita la piattaforma ITsART, a pi\u00f9 di un anno da quando \u2013 durante il primo lockdown \u2013 il ministro della Cultura Dario Franceschini ne aveva parlato come di \u00abuna sorta di Netflix della cultura\u00bb, pensata per \u00aboffrire a tutto il mondo la cultura italiana a pagamento\u00bb. \u00c8 presto per dare giudizi definitivi sulla piattaforma, e di certo sar\u00e0 difficile farlo anche pi\u00f9 avanti senza numeri precisi. Al momento, l\u2019unica cosa che si pu\u00f2 fare \u00e8 guardare com\u2019\u00e8 fatto il sito, contare quanti contenuti ci sono (circa 700 \u201ctitoli\u201d, tra film, documentari, spettacoli teatrali e musicali e altri eventi) e provare a dare un giudizio sul loro valore e sulla loro variet\u00e0. Intanto, una cosa notata da pi\u00f9 parti \u00e8 che diversi contenuti di ITsART sono a pagamento sulla piattaforma sebbene altrove, per esempio su RaiPlay, siano invece disponibili gratuitamente.)\\', \\'performance\\': {\\'dataset\\': \\'NewsSum-IT\\', \\'accuracy\\': {\\'Rouge1\\': 0.339, \\'Rouge2\\': 0.16, \\'RougeL\\': 0.263}}, \\'description\\': \\'IT5 Base model fine-tuned on news summarization on the Fanpage and Il Post corpora for Italian Language Understanding and Generation.\\'}', metadata={})]", "category": "generic"}
{"question_id": 301, "text": " A customer needs to translate a French email they received into Spanish.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Translation\\', \\'api_name\\': \\'Helsinki-NLP/opus-mt-en-fr\\', \\'api_call\\': \"translate(\\'input_text\\', model=\\'Helsinki-NLP/opus-mt-en-fr\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'opus\\', \\'accuracy\\': {\\'BLEU\\': {\\'newsdiscussdev2015-enfr.en.fr\\': 33.8, \\'newsdiscusstest2015-enfr.en.fr\\': 40.0, \\'newssyscomb2009.en.fr\\': 29.8, \\'news-test2008.en.fr\\': 27.5, \\'newstest2009.en.fr\\': 29.4, \\'newstest2010.en.fr\\': 32.7, \\'newstest2011.en.fr\\': 34.3, \\'newstest2012.en.fr\\': 31.8, \\'newstest2013.en.fr\\': 33.2, \\'Tatoeba.en.fr\\': 50.5}}}, \\'description\\': \\'Helsinki-NLP/opus-mt-en-fr is a translation model that translates English text to French using the Hugging Face Transformers library. It is based on the OPUS dataset and uses a transformer-align architecture with normalization and SentencePiece pre-processing.\\'}', metadata={})]", "category": "generic"}
{"question_id": 302, "text": " I am a writer, and I have written a long article, but I need to come up with a summary for my article to publish.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'pszemraj/long-t5-tglobal-base-16384-book-summary\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'pszemraj/long-t5-tglobal-base-16384-book-summary\\')\", \\'api_arguments\\': [\\'long_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nimport torch\\\\nsummarizer = pipeline(\\\\n summarization,\\\\n pszemraj/long-t5-tglobal-base-16384-book-summary,\\\\n device=0 if torch.cuda.is_available() else -1,\\\\n)\\\\nlong_text = Here is a lot of text I don\\'t want to read. Replace me\\\\nresult = summarizer(long_text)\\\\nprint(result[0][summary_text])\", \\'performance\\': {\\'dataset\\': \\'kmfoda/booksum\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.408, \\'ROUGE-2\\': 6.065, \\'ROUGE-L\\': 16.721, \\'ROUGE-LSUM\\': 33.34}}, \\'description\\': \\'A fine-tuned version of google/long-t5-tglobal-base on the kmfoda/booksum dataset, which can be used to summarize long text and generate SparkNotes-esque summaries of arbitrary topics. The model generalizes reasonably well to academic and narrative text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 303, "text": " Your company has launched a new chatbot service. We need to create conversations with the customers to provide information about our products and answer their queries.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Generation\\', \\'api_name\\': \\'mywateriswet/ShuanBot\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'mywateriswet/ShuanBot\\')\", \\'api_arguments\\': \\'message\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"response = chatbot(\\'What is your name?\\')\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'ShuanBot is a conversational chatbot model based on the GPT-2 architecture. It can be used for generating human-like responses in a chat context.\\'}', metadata={})]", "category": "generic"}
{"question_id": 304, "text": " We are a travel company trying to offer personalized experiences to our customers. We want to implement a conversational AI to help users get information about travel destinations.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Generation\\', \\'api_name\\': \\'mywateriswet/ShuanBot\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'mywateriswet/ShuanBot\\')\", \\'api_arguments\\': \\'message\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"response = chatbot(\\'What is your name?\\')\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'ShuanBot is a conversational chatbot model based on the GPT-2 architecture. It can be used for generating human-like responses in a chat context.\\'}', metadata={})]", "category": "generic"}
{"question_id": 305, "text": " Write a tweet on the topic of \\\"The Future of AI in Education\\\".\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Conversational\\', \\'api_name\\': \\'Pi3141/DialoGPT-medium-elon-3\\', \\'api_call\\': \"pipeline(\\'text-generation\\', model=\\'Pi3141/DialoGPT-medium-elon-3\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'Input a message to start chatting with Pi3141/DialoGPT-medium-elon-3.\\', \\'performance\\': {\\'dataset\\': \\'Twitter tweets by Elon Musk\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'DialoGPT model that talks like Elon Musk, trained on Twitter tweets by Elon Musk. This model will spew meaningless shit about 40% of the time. Trained on 8 epochs. But with a larger dataset this time. The AI can now use more emojis, I think.\\'}', metadata={})]", "category": "generic"}
{"question_id": 306, "text": " Create a poem about nature for a school assignment.\\n###Input: Once upon a time, in a land of greenery and beauty,\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Paraphrasing\\', \\'api_name\\': \\'prithivida/parrot_paraphraser_on_T5\\', \\'api_call\\': \"Parrot(model_tag=\\'prithivida/parrot_paraphraser_on_T5\\', use_gpu=False)\", \\'api_arguments\\': [\\'input_phrase\\', \\'diversity_ranker\\', \\'do_diverse\\', \\'max_return_phrases\\', \\'max_length\\', \\'adequacy_threshold\\', \\'fluency_threshold\\'], \\'python_environment_requirements\\': [\\'torch\\', \\'transformers\\'], \\'example_code\\': \\'from parrot import Parrot\\\\nimport torch\\\\nimport warnings\\\\nwarnings.filterwarnings(ignore)\\\\n\\\\nparrot = Parrot(model_tag=prithivida/parrot_paraphraser_on_T5, use_gpu=False)\\\\nphrases = [Can you recommed some upscale restaurants in Newyork?,\\\\n What are the famous places we should not miss in Russia?\\\\n]\\\\nfor phrase in phrases:\\\\n print(-*100)\\\\n print(Input_phrase: , phrase)\\\\n print(-*100)\\\\n para_phrases = parrot.augment(input_phrase=phrase)\\\\n for para_phrase in para_phrases:\\\\n  print(para_phrase)\\', \\'performance\\': {\\'dataset\\': \\'Not mentioned\\', \\'accuracy\\': \\'Not mentioned\\'}, \\'description\\': \\'Parrot is a paraphrase based utterance augmentation framework purpose built to accelerate training NLU models. It offers knobs to control Adequacy, Fluency, and Diversity as per your needs. It mainly focuses on augmenting texts typed-into or spoken-to conversational interfaces for building robust NLU models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 307, "text": " We want to code a loading spinner to display when our system is undergoing maintenance. Can you generate the code for us?\\n###Input: <noinput>\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'SYSPIN/Marathi_Male_TTS\\', \\'api_call\\': \"api.load(\\'ESPnet/espnet_model_zoo:SYSPIN/Marathi_Male_TTS\\').\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'huggingface_hub\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Marathi Male Text-to-Speech model using ESPnet framework.\\'}', metadata={})]", "category": "generic"}
{"question_id": 308, "text": " I am an English teacher. What kind of creative writing exercise could I give to my students that they could write a story using the words \\\"moon\\\", \\\"rabbit\\\", \\\"forest\\\", and \\\"magic\\\"?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'text2text-generation\\', \\'api_name\\': \\'tuner007/pegasus_summarizer\\', \\'api_call\\': \"PegasusForConditionalGeneration.from_pretrained(\\'tuner007/pegasus_summarizer\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'pip install sentencepiece\\'], \\'example_code\\': \"context = \\\\nIndia wicket-keeper batsman Rishabh Pant has said someone from the crowd threw a ball on pacer Mohammed Siraj while he was fielding in the ongoing third Test against England on Wednesday. Pant revealed the incident made India skipper Virat Kohli upset. I think, somebody threw a ball inside, at Siraj, so he [Kohli] was upset, said Pant in a virtual press conference after the close of the first day\\'s play.You can say whatever you want to chant, but don\\'t throw things at the fielders and all those things. It is not good for cricket, I guess, he added.In the third session of the opening day of the third Test, a section of spectators seemed to have asked Siraj the score of the match to tease the pacer. The India pacer however came with a brilliant reply as he gestured 1-0 (India leading the Test series) towards the crowd.Earlier this month, during the second Test match, there was some bad crowd behaviour on a show as some unruly fans threw champagne corks at India batsman KL Rahul.Kohli also intervened and he was seen gesturing towards the opening batsman to know more about the incident. An over later, the TV visuals showed that many champagne corks were thrown inside the playing field, and the Indian players were visibly left frustrated.Coming back to the game, after bundling out India for 78, openers Rory Burns and Haseeb Hameed ensured that England took the honours on the opening day of the ongoing third Test.At stumps, England\\'s score reads 120/0 and the hosts have extended their lead to 42 runs. For the Three Lions, Burns (52) and Hameed (60) are currently unbeaten at the crease.Talking about the pitch on opening day, Pant said, They took the heavy roller, the wicket was much more settled down, and they batted nicely also, he said. But when we batted, the wicket was slightly soft, and they bowled in good areas, but we could have applied [ourselves] much better.Both England batsmen managed to see off the final session and the hosts concluded the opening day with all ten wickets intact, extending the lead to 42.(ANI)\\\\n\\\\nget_response(context)\", \\'performance\\': {\\'dataset\\': \\'cnn_dailymail\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.604, \\'ROUGE-2\\': 14.64, \\'ROUGE-L\\': 23.884, \\'ROUGE-LSUM\\': 32.902, \\'loss\\': 2.576, \\'gen_len\\': 76.398}}, \\'description\\': \\'PEGASUS fine-tuned for summarization\\'}', metadata={})]", "category": "generic"}
{"question_id": 309, "text": " Our team needs to develop a tool that can translate articles from English to German in real-time.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'sshleifer/tiny-marian-en-de\\', \\'api_call\\': \"pipeline(\\'translation_en_to_de\\', model=\\'sshleifer/tiny-marian-en-de\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A tiny English to German translation model using the Marian framework in Hugging Face Transformers.\\'}', metadata={})]", "category": "generic"}
{"question_id": 310, "text": " Design an efficient search engine that retrieves most relevant articles based on a pool of long text documents.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'pszemraj/long-t5-tglobal-base-16384-book-summary\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'pszemraj/long-t5-tglobal-base-16384-book-summary\\')\", \\'api_arguments\\': [\\'long_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nimport torch\\\\nsummarizer = pipeline(\\\\n summarization,\\\\n pszemraj/long-t5-tglobal-base-16384-book-summary,\\\\n device=0 if torch.cuda.is_available() else -1,\\\\n)\\\\nlong_text = Here is a lot of text I don\\'t want to read. Replace me\\\\nresult = summarizer(long_text)\\\\nprint(result[0][summary_text])\", \\'performance\\': {\\'dataset\\': \\'kmfoda/booksum\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.408, \\'ROUGE-2\\': 6.065, \\'ROUGE-L\\': 16.721, \\'ROUGE-LSUM\\': 33.34}}, \\'description\\': \\'A fine-tuned version of google/long-t5-tglobal-base on the kmfoda/booksum dataset, which can be used to summarize long text and generate SparkNotes-esque summaries of arbitrary topics. The model generalizes reasonably well to academic and narrative text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 311, "text": " Our designer hired me to create an application to automatically translate colors from a design. Usually, our designer likes to work with English, to get the color in Italian.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Translation\\', \\'api_name\\': \\'Helsinki-NLP/opus-mt-it-en\\', \\'api_call\\': \"pipeline(\\'translation_it_to_en\\', model=\\'Helsinki-NLP/opus-mt-it-en\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\', \\'torch\\'], \\'example_code\\': \"from transformers import pipeline\\\\ntranslation = pipeline(\\'translation_it_to_en\\', model=\\'Helsinki-NLP/opus-mt-it-en\\')(\\'Ciao mondo!\\')\", \\'performance\\': {\\'dataset\\': \\'opus\\', \\'accuracy\\': {\\'BLEU\\': {\\'newssyscomb2009.it.en\\': 35.3, \\'newstest2009.it.en\\': 34.0, \\'Tatoeba.it.en\\': 70.9}, \\'chr-F\\': {\\'newssyscomb2009.it.en\\': 0.6, \\'newstest2009.it.en\\': 0.594, \\'Tatoeba.it.en\\': 0.808}}}, \\'description\\': \\'A transformer model for Italian to English translation trained on the OPUS dataset. It can be used for translating Italian text to English.\\'}', metadata={})]", "category": "generic"}
{"question_id": 312, "text": " I am building an AI to create in work report, this model should be able to complete the sentence for the report.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentence Correction\\', \\'api_name\\': \\'flexudy/t5-base-multi-sentence-doctor\\', \\'api_call\\': \"AutoModelWithLMHead.from_pretrained(\\'flexudy/t5-base-multi-sentence-doctor\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'from transformers import AutoTokenizer, AutoModelWithLMHead\\\\ntokenizer = AutoTokenizer.from_pretrained(flexudy/t5-base-multi-sentence-doctor)\\\\nmodel = AutoModelWithLMHead.from_pretrained(flexudy/t5-base-multi-sentence-doctor)\\\\ninput_text = repair_sentence: m a medical doct context: {That is my job I a}{or I save lives} </s>\\\\ninput_ids = tokenizer.encode(input_text, return_tensors=pt)\\\\noutputs = model.generate(input_ids, max_length=32, num_beams=1)\\\\nsentence = tokenizer.decode(outputs[0], skip_special_tokens=True, clean_up_tokenization_spaces=True)\\\\nassert sentence == I am a medical doctor.\\', \\'performance\\': {\\'dataset\\': \\'tatoeba\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'Sentence doctor is a T5 model that attempts to correct the errors or mistakes found in sentences. Model works on English, German and French text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 313, "text": " A language learning application wants to generate fill-in-the-blank questions for learners. We need to mask sentences with a keyword to create these questions.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Fill-Mask\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Masked Language Modeling\\', \\'api_name\\': \\'bert-base-multilingual-cased\\', \\'api_call\\': \"pipeline(\\'fill-mask\\', model=\\'bert-base-multilingual-cased\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"from transformers import pipeline\\\\nunmasker = pipeline(\\'fill-mask\\', model=\\'bert-base-multilingual-cased\\')\\\\nunmasker(Hello I\\'m a [MASK] model.)\", \\'performance\\': {\\'dataset\\': \\'wikipedia\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'BERT multilingual base model (cased) is pretrained on the top 104 languages with the largest Wikipedia using a masked language modeling (MLM) objective. The model is case sensitive and can be used for masked language modeling or next sentence prediction. It is intended to be fine-tuned on a downstream task, such as sequence classification, token classification, or question answering.\\'}', metadata={})]", "category": "generic"}
{"question_id": 314, "text": " Our client has a medical report and we are trying to assist him in finding relevant information.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 315, "text": " I am a data scientist and need to find which among given documents is similar, provide me the example code.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Joblib\\', \\'functionality\\': \\'Multi-class Classification\\', \\'api_name\\': \\'Alexei1/imdb\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'json\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'IMDB\\', \\'accuracy\\': 0.487}, \\'description\\': \\'A tabular classification model trained using AutoTrain for sentiment analysis on the IMDB dataset. The model has a CO2 emission of 0.0186 grams and an accuracy of 0.487.\\'}', metadata={})]", "category": "generic"}
{"question_id": 316, "text": " My company wants to analyze and cluster different movie scripts. I need a suggestion to embed the movie dialogue text into dense vector representations.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Classification\\', \\'api_name\\': \\'lvwerra/distilbert-imdb\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'lvwerra/distilbert-imdb\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\', \\'pytorch\\'], \\'example_code\\': \"classifier(\\'I love this movie!\\')\", \\'performance\\': {\\'dataset\\': \\'imdb\\', \\'accuracy\\': 0.928}, \\'description\\': \\'This model is a fine-tuned version of distilbert-base-uncased on the imdb dataset. It is used for sentiment analysis on movie reviews and achieves an accuracy of 0.928 on the evaluation set.\\'}', metadata={})]", "category": "generic"}
{"question_id": 317, "text": " Analyze how characters in a book are connected and if they share any similarity based on their conversation.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'hyunwoongko/blenderbot-9B\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'hyunwoongko/blenderbot-9B\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'Input a message to start chatting with hyunwoongko/blenderbot-9B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high-performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi-turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 318, "text": " We are a company specializing in building automation systems for homes. We'd like to convert our written rules into an audio format for user interaction purposes.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'mio/Artoria\\', \\'api_call\\': \"pipeline(\\'text-to-speech\\', model=\\'mio/Artoria\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline; tts = pipeline(\\'text-to-speech\\', model=\\'mio/Artoria\\'); tts(\\'s\\')\", \\'performance\\': {\\'dataset\\': \\'fate\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This model was trained by mio using fate recipe in espnet. It is a text-to-speech model that can convert text input into speech output.\\'}', metadata={})]", "category": "generic"}
{"question_id": 319, "text": " We are developing a virtual assistant. One of its major functionalities is to convert written text to speech.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'mio/Artoria\\', \\'api_call\\': \"pipeline(\\'text-to-speech\\', model=\\'mio/Artoria\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline; tts = pipeline(\\'text-to-speech\\', model=\\'mio/Artoria\\'); tts(\\'s\\')\", \\'performance\\': {\\'dataset\\': \\'fate\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This model was trained by mio using fate recipe in espnet. It is a text-to-speech model that can convert text input into speech output.\\'}', metadata={})]", "category": "generic"}
{"question_id": 320, "text": " I have a recorded meeting between multiple participants. Identify when two or more people are speaking simultaneously.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'pyannote.audio\\', \\'functionality\\': \\'Speaker diarization\\', \\'api_name\\': \\'johnislarry/cloned-pyannote-speaker-diarization-endpoint\\', \\'api_call\\': \"Pipeline.from_pretrained(\\'pyannote/speaker-diarization@2.1\\',use_auth_token=\\'ACCESS_TOKEN_GOES_HERE\\')\", \\'api_arguments\\': [\\'num_speakers\\', \\'min_speakers\\', \\'max_speakers\\', \\'segmentation_onset\\'], \\'python_environment_requirements\\': \\'pyannote.audio 2.0\\', \\'example_code\\': {\\'load_pipeline\\': \\'from pyannote.audio import Pipeline\\\\npipeline = Pipeline.from_pretrained(pyannote/speaker-diarization@2022.07)\\', \\'apply_pipeline\\': \\'diarization = pipeline(audio.wav)\\', \\'save_output\\': \\'with open(audio.rttm, w) as rttm:\\\\n  diarization.write_rttm(rttm)\\'}, \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'AISHELL-4\\', \\'accuracy\\': {\\'DER%\\': 14.61, \\'FA%\\': 3.31, \\'Miss%\\': 4.35, \\'Conf%\\': 6.95}}, {\\'name\\': \\'AMI Mix-Headset only_words\\', \\'accuracy\\': {\\'DER%\\': 18.21, \\'FA%\\': 3.28, \\'Miss%\\': 11.07, \\'Conf%\\': 3.87}}, {\\'name\\': \\'AMI Array1-01 only_words\\', \\'accuracy\\': {\\'DER%\\': 29.0, \\'FA%\\': 2.71, \\'Miss%\\': 21.61, \\'Conf%\\': 4.68}}, {\\'name\\': \\'CALLHOME Part2\\', \\'accuracy\\': {\\'DER%\\': 30.24, \\'FA%\\': 3.71, \\'Miss%\\': 16.86, \\'Conf%\\': 9.66}}, {\\'name\\': \\'DIHARD 3 Full\\', \\'accuracy\\': {\\'DER%\\': 20.99, \\'FA%\\': 4.25, \\'Miss%\\': 10.74, \\'Conf%\\': 6.0}}, {\\'name\\': \\'REPERE Phase 2\\', \\'accuracy\\': {\\'DER%\\': 12.62, \\'FA%\\': 1.55, \\'Miss%\\': 3.3, \\'Conf%\\': 7.76}}, {\\'name\\': \\'VoxConverse v0.0.2\\', \\'accuracy\\': {\\'DER%\\': 12.76, \\'FA%\\': 3.45, \\'Miss%\\': 3.85, \\'Conf%\\': 5.46}}]}, \\'description\\': \\'This API provides speaker diarization functionality using the pyannote.audio framework. It is capable of processing audio files and outputting speaker diarization results in RTTM format. The API supports providing the number of speakers, minimum and maximum number of speakers, and adjusting the segmentation onset threshold.\\'}', metadata={})]", "category": "generic"}
{"question_id": 321, "text": " I have a collection of audio recordings from interviews that I need transcribed into text. How can I achieve this using automatic speech recognition?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Speech Recognition\\', \\'api_name\\': \\'jonatasgrosman/wav2vec2-large-xlsr-53-portuguese\\', \\'api_call\\': \"SpeechRecognitionModel(\\'jonatasgrosman/wav2vec2-large-xlsr-53-portuguese\\')\", \\'api_arguments\\': [\\'audio_paths\\'], \\'python_environment_requirements\\': [\\'huggingsound\\', \\'torch\\', \\'librosa\\', \\'datasets\\', \\'transformers\\'], \\'example_code\\': \\'from huggingsound import SpeechRecognitionModel\\\\nmodel = SpeechRecognitionModel(jonatasgrosman/wav2vec2-large-xlsr-53-portuguese)\\\\naudio_paths = [/path/to/file.mp3, /path/to/another_file.wav]\\\\ntranscriptions = model.transcribe(audio_paths)\\', \\'performance\\': {\\'dataset\\': \\'mozilla-foundation/common_voice_6_0\\', \\'accuracy\\': {\\'Test WER\\': 11.31, \\'Test CER\\': 3.74, \\'Test WER (+LM)\\': 9.01, \\'Test CER (+LM)\\': 3.21}}, \\'description\\': \\'Fine-tuned facebook/wav2vec2-large-xlsr-53 on Portuguese using the train and validation splits of Common Voice 6.1. When using this model, make sure that your speech input is sampled at 16kHz.\\'}', metadata={})]", "category": "generic"}
{"question_id": 322, "text": " A Russia-based online lesson website needs an English subtitle for the recorded lessons of the students.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Translation\\', \\'api_name\\': \\'opus-mt-en-ru\\', \\'api_call\\': \"AutoModelForSeq2SeqLM.from_pretrained(\\'Helsinki-NLP/opus-mt-en-ru\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\', \\'torch\\', \\'tensorflow\\'], \\'example_code\\': \"from transformers import pipeline\\\\ntranslation = pipeline(\\'translation_en_to_ru\\', model=\\'Helsinki-NLP/opus-mt-en-ru\\')\\\\ntranslated_text = translation(\\'Hello, how are you?\\')\", \\'performance\\': {\\'dataset\\': \\'newstest2019-enru\\', \\'accuracy\\': \\'27.1\\'}, \\'description\\': \\'Helsinki-NLP/opus-mt-en-ru is a translation model trained on the OPUS dataset, which translates English text to Russian. It is based on the Marian NMT framework and can be used with Hugging Face Transformers.\\'}', metadata={})]", "category": "generic"}
{"question_id": 323, "text": " My lecture has been recorded, and I would like to transcribe the audio to create a transcript of my presentation.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'CTranslate2\\', \\'functionality\\': \\'Automatic Speech Recognition\\', \\'api_name\\': \\'guillaumekln/faster-whisper-large-v2\\', \\'api_call\\': \"WhisperModel(\\'large-v2\\')\", \\'api_arguments\\': [\\'audio.mp3\\'], \\'python_environment_requirements\\': [\\'faster_whisper\\'], \\'example_code\\': \\'from faster_whisper import WhisperModel\\\\nmodel = WhisperModel(large-v2)\\\\nsegments, info = model.transcribe(audio.mp3)\\\\nfor segment in segments:\\\\n print([%.2fs -&gt; %.2fs] %s % (segment.start, segment.end, segment.text))\\', \\'performance\\': {\\'dataset\\': \\'99 languages\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Whisper large-v2 model for CTranslate2. This model can be used in CTranslate2 or projets based on CTranslate2 such as faster-whisper.\\'}', metadata={})]", "category": "generic"}
{"question_id": 324, "text": " Provide the code to enhance a single audio track, possibly containing dialogue, music and background noise, extracted from a video game.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Audio-to-Audio\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Speech Enhancement\\', \\'api_name\\': \\'speechbrain/mtl-mimic-voicebank\\', \\'api_call\\': \"WaveformEnhancement.from_hparams(\\'speechbrain/mtl-mimic-voicebank\\', \\'pretrained_models/mtl-mimic-voicebank\\')\", \\'api_arguments\\': [\\'source\\', \\'savedir\\'], \\'python_environment_requirements\\': \\'pip install speechbrain\\', \\'example_code\\': \"import torchaudio\\\\nfrom speechbrain.pretrained import WaveformEnhancement\\\\nenhance_model = WaveformEnhancement.from_hparams(\\\\n source=speechbrain/mtl-mimic-voicebank,\\\\n savedir=pretrained_models/mtl-mimic-voicebank,\\\\n)\\\\nenhanced = enhance_model.enhance_file(speechbrain/mtl-mimic-voicebank/example.wav)\\\\ntorchaudio.save(\\'enhanced.wav\\', enhanced.unsqueeze(0).cpu(), 16000)\", \\'performance\\': {\\'dataset\\': \\'Voicebank\\', \\'accuracy\\': {\\'Test PESQ\\': 3.05, \\'Test COVL\\': 3.74, \\'Valid WER\\': 2.89, \\'Test WER\\': 2.8}}, \\'description\\': \\'This repository provides all the necessary tools to perform enhancement and\\\\nrobust ASR training (EN) within\\\\nSpeechBrain. For a better experience we encourage you to learn more about\\\\nSpeechBrain. The model performance is:\\\\nRelease\\\\nTest PESQ\\\\nTest COVL\\\\nValid WER\\\\nTest WER\\\\n22-06-21\\\\n3.05\\\\n3.74\\\\n2.89\\\\n2.80\\\\nWorks with SpeechBrain v0.5.12\\'}', metadata={})]", "category": "generic"}
{"question_id": 325, "text": " We're working with a voiceover company, and they're looking for a tool to help them change the voice style of voiceovers while keeping the same content.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'chavinlo/TempoFunk\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'chavinlo/TempoFunk\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Text-to-Video model using Hugging Face Transformers library. Model is capable of generating video content based on the input text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 326, "text": " Develop a listening assistant device for audiobooks that is capable of detecting and reducing noise.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Voice Activity Detection, Speech-to-Noise Ratio, and C50 Room Acoustics Estimation\\', \\'api_name\\': \\'pyannote/brouhaha\\', \\'api_call\\': \"Model.from_pretrained(\\'pyannote/brouhaha\\', use_auth_token=\\'ACCESS_TOKEN_GOES_HERE\\')\", \\'api_arguments\\': [\\'audio.wav\\'], \\'python_environment_requirements\\': [\\'pyannote-audio\\', \\'brouhaha-vad\\'], \\'example_code\\': [\\'from pyannote.audio import Model\\', \\'model = Model.from_pretrained(pyannote/brouhaha, use_auth_token=ACCESS_TOKEN_GOES_HERE)\\', \\'from pyannote.audio import Inference\\', \\'inference = Inference(model)\\', \\'output = inference(audio.wav)\\', \\'for frame, (vad, snr, c50) in output:\\', \\'  t = frame.middle\\', \\'  print(f{t:8.3f} vad={100*vad:.0f}% snr={snr:.0f} c50={c50:.0f})\\'], \\'performance\\': {\\'dataset\\': \\'LibriSpeech, AudioSet, EchoThief, MIT-Acoustical-Reverberation-Scene\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Brouhaha is a joint voice activity detection, speech-to-noise ratio, and C50 room acoustics estimation model. It is based on the PyTorch framework and uses the pyannote.audio library.\\'}', metadata={})]", "category": "generic"}
{"question_id": 327, "text": " The company needs to separate voice from background noise in a recorded podcast episode.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'pyannote.audio\\', \\'functionality\\': \\'Speaker diarization\\', \\'api_name\\': \\'johnislarry/cloned-pyannote-speaker-diarization-endpoint\\', \\'api_call\\': \"Pipeline.from_pretrained(\\'pyannote/speaker-diarization@2.1\\',use_auth_token=\\'ACCESS_TOKEN_GOES_HERE\\')\", \\'api_arguments\\': [\\'num_speakers\\', \\'min_speakers\\', \\'max_speakers\\', \\'segmentation_onset\\'], \\'python_environment_requirements\\': \\'pyannote.audio 2.0\\', \\'example_code\\': {\\'load_pipeline\\': \\'from pyannote.audio import Pipeline\\\\npipeline = Pipeline.from_pretrained(pyannote/speaker-diarization@2022.07)\\', \\'apply_pipeline\\': \\'diarization = pipeline(audio.wav)\\', \\'save_output\\': \\'with open(audio.rttm, w) as rttm:\\\\n  diarization.write_rttm(rttm)\\'}, \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'AISHELL-4\\', \\'accuracy\\': {\\'DER%\\': 14.61, \\'FA%\\': 3.31, \\'Miss%\\': 4.35, \\'Conf%\\': 6.95}}, {\\'name\\': \\'AMI Mix-Headset only_words\\', \\'accuracy\\': {\\'DER%\\': 18.21, \\'FA%\\': 3.28, \\'Miss%\\': 11.07, \\'Conf%\\': 3.87}}, {\\'name\\': \\'AMI Array1-01 only_words\\', \\'accuracy\\': {\\'DER%\\': 29.0, \\'FA%\\': 2.71, \\'Miss%\\': 21.61, \\'Conf%\\': 4.68}}, {\\'name\\': \\'CALLHOME Part2\\', \\'accuracy\\': {\\'DER%\\': 30.24, \\'FA%\\': 3.71, \\'Miss%\\': 16.86, \\'Conf%\\': 9.66}}, {\\'name\\': \\'DIHARD 3 Full\\', \\'accuracy\\': {\\'DER%\\': 20.99, \\'FA%\\': 4.25, \\'Miss%\\': 10.74, \\'Conf%\\': 6.0}}, {\\'name\\': \\'REPERE Phase 2\\', \\'accuracy\\': {\\'DER%\\': 12.62, \\'FA%\\': 1.55, \\'Miss%\\': 3.3, \\'Conf%\\': 7.76}}, {\\'name\\': \\'VoxConverse v0.0.2\\', \\'accuracy\\': {\\'DER%\\': 12.76, \\'FA%\\': 3.45, \\'Miss%\\': 3.85, \\'Conf%\\': 5.46}}]}, \\'description\\': \\'This API provides speaker diarization functionality using the pyannote.audio framework. It is capable of processing audio files and outputting speaker diarization results in RTTM format. The API supports providing the number of speakers, minimum and maximum number of speakers, and adjusting the segmentation onset threshold.\\'}', metadata={})]", "category": "generic"}
{"question_id": 328, "text": " Develop a method to detect which languages are being spoken in international conference calls.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'pyannote.audio\\', \\'functionality\\': \\'Speaker diarization\\', \\'api_name\\': \\'johnislarry/cloned-pyannote-speaker-diarization-endpoint\\', \\'api_call\\': \"Pipeline.from_pretrained(\\'pyannote/speaker-diarization@2.1\\',use_auth_token=\\'ACCESS_TOKEN_GOES_HERE\\')\", \\'api_arguments\\': [\\'num_speakers\\', \\'min_speakers\\', \\'max_speakers\\', \\'segmentation_onset\\'], \\'python_environment_requirements\\': \\'pyannote.audio 2.0\\', \\'example_code\\': {\\'load_pipeline\\': \\'from pyannote.audio import Pipeline\\\\npipeline = Pipeline.from_pretrained(pyannote/speaker-diarization@2022.07)\\', \\'apply_pipeline\\': \\'diarization = pipeline(audio.wav)\\', \\'save_output\\': \\'with open(audio.rttm, w) as rttm:\\\\n  diarization.write_rttm(rttm)\\'}, \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'AISHELL-4\\', \\'accuracy\\': {\\'DER%\\': 14.61, \\'FA%\\': 3.31, \\'Miss%\\': 4.35, \\'Conf%\\': 6.95}}, {\\'name\\': \\'AMI Mix-Headset only_words\\', \\'accuracy\\': {\\'DER%\\': 18.21, \\'FA%\\': 3.28, \\'Miss%\\': 11.07, \\'Conf%\\': 3.87}}, {\\'name\\': \\'AMI Array1-01 only_words\\', \\'accuracy\\': {\\'DER%\\': 29.0, \\'FA%\\': 2.71, \\'Miss%\\': 21.61, \\'Conf%\\': 4.68}}, {\\'name\\': \\'CALLHOME Part2\\', \\'accuracy\\': {\\'DER%\\': 30.24, \\'FA%\\': 3.71, \\'Miss%\\': 16.86, \\'Conf%\\': 9.66}}, {\\'name\\': \\'DIHARD 3 Full\\', \\'accuracy\\': {\\'DER%\\': 20.99, \\'FA%\\': 4.25, \\'Miss%\\': 10.74, \\'Conf%\\': 6.0}}, {\\'name\\': \\'REPERE Phase 2\\', \\'accuracy\\': {\\'DER%\\': 12.62, \\'FA%\\': 1.55, \\'Miss%\\': 3.3, \\'Conf%\\': 7.76}}, {\\'name\\': \\'VoxConverse v0.0.2\\', \\'accuracy\\': {\\'DER%\\': 12.76, \\'FA%\\': 3.45, \\'Miss%\\': 3.85, \\'Conf%\\': 5.46}}]}, \\'description\\': \\'This API provides speaker diarization functionality using the pyannote.audio framework. It is capable of processing audio files and outputting speaker diarization results in RTTM format. The API supports providing the number of speakers, minimum and maximum number of speakers, and adjusting the segmentation onset threshold.\\'}', metadata={})]", "category": "generic"}
{"question_id": 329, "text": " Develop a system that can detect a keyword in a short audio clip. Make sure it recognizes a variety of keywords.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'pyannote.audio\\', \\'functionality\\': \\'Automatic Speech Recognition\\', \\'api_name\\': \\'pyannote/voice-activity-detection\\', \\'api_call\\': \"Pipeline.from_pretrained(\\'pyannote/voice-activity-detection\\')\", \\'api_arguments\\': [\\'audio.wav\\'], \\'python_environment_requirements\\': [\\'pyannote.audio 2.1\\'], \\'example_code\\': \\'from pyannote.audio import Pipeline\\\\npipeline = Pipeline.from_pretrained(pyannote/voice-activity-detection, use_auth_token=ACCESS_TOKEN_GOES_HERE)\\\\noutput = pipeline(audio.wav)\\\\nfor speech in output.get_timeline().support():\\\\n  # active speech between speech.start and speech.end\\', \\'performance\\': {\\'dataset\\': \\'ami\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'A pretrained voice activity detection pipeline that detects active speech in audio files.\\'}', metadata={})]", "category": "generic"}
{"question_id": 330, "text": " I am the director of a movie theater chain, and we are thinking of implementing a chat product to get movie goers to classify movie reviews as positive or negative through AI.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Classification\\', \\'api_name\\': \\'lvwerra/distilbert-imdb\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'lvwerra/distilbert-imdb\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\', \\'pytorch\\'], \\'example_code\\': \"classifier(\\'I love this movie!\\')\", \\'performance\\': {\\'dataset\\': \\'imdb\\', \\'accuracy\\': 0.928}, \\'description\\': \\'This model is a fine-tuned version of distilbert-base-uncased on the imdb dataset. It is used for sentiment analysis on movie reviews and achieves an accuracy of 0.928 on the evaluation set.\\'}', metadata={})]", "category": "generic"}
{"question_id": 331, "text": " You have been provided with the dataset of plants, and your task is to identify the species of plants among Iris Setosa, Iris Versicolor, and Iris Virginica.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Joblib\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'abhishek/autotrain-iris-knn\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'scikit-learn/iris\\', \\'accuracy\\': 0.9}, \\'description\\': \\'A K-Nearest Neighbors (KNN) model trained on the Iris dataset for multi-class classification. The model is trained using AutoTrain and has an accuracy of 0.9.\\'}', metadata={})]", "category": "generic"}
{"question_id": 332, "text": " Our energy company is trying to identify anomalies in the energy consumption data. Could you perform anomaly detection on the time series data?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Keras\\', \\'functionality\\': \\'anomaly-detection\\', \\'api_name\\': \\'keras-io/timeseries-anomaly-detection\\', \\'api_call\\': \"TFAutoModelForSequenceClassification.from_pretrained(\\'keras-io/timeseries-anomaly-detection\\')\", \\'api_arguments\\': {\\'optimizer\\': {\\'name\\': \\'Adam\\', \\'learning_rate\\': 0.001, \\'decay\\': 0.0, \\'beta_1\\': 0.9, \\'beta_2\\': 0.999, \\'epsilon\\': 1e-07, \\'amsgrad\\': False}, \\'training_precision\\': \\'float32\\'}, \\'python_environment_requirements\\': [\\'tensorflow\\', \\'keras\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'Numenta Anomaly Benchmark(NAB)\\', \\'accuracy\\': {\\'Train Loss\\': 0.006, \\'Validation Loss\\': 0.008}}, \\'description\\': \\'This script demonstrates how you can use a reconstruction convolutional autoencoder model to detect anomalies in timeseries data. We will use the Numenta Anomaly Benchmark(NAB) dataset. It provides artifical timeseries data containing labeled anomalous periods of behavior. Data are ordered, timestamped, single-valued metrics.\\'}', metadata={})]", "category": "generic"}
{"question_id": 333, "text": " Our company seeks a solution to reduce the carbon emissions due to our operations. We have collected historical data about our emissions. We are looking for a machine learning model to predict the carbon emissions based on this data.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'kochetkovIT/autotrain-ironhack-49741119788\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'json\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'kochetkovIT/autotrain-data-ironhack\\', \\'accuracy\\': {\\'Loss\\': 2.603, \\'R2\\': 0.013, \\'MSE\\': 6.776, \\'MAE\\': 1.666, \\'RMSLE\\': 0.502}}, \\'description\\': \\'A tabular regression model trained using AutoTrain to predict carbon emissions (in grams) with an R2 score of 0.013.\\'}', metadata={})]", "category": "generic"}
{"question_id": 334, "text": " Create a program that predicts carbon emissions for new data using the given model.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'45473113800\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'samvelkoch/autotrain-data-prknsn-2\\', \\'accuracy\\': {\\'Loss\\': 5.079, \\'R2\\': 0.109, \\'MSE\\': 25.795, \\'MAE\\': 3.78, \\'RMSLE\\': 0.849}}, \\'description\\': \\'A tabular regression model trained with AutoTrain for predicting carbon emissions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 335, "text": " A company has collected data about fish measurements and weights. We need to estimate a fish's weight based on its measurements.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Scikit-learn\\', \\'functionality\\': \\'GradientBoostingRegressor\\', \\'api_name\\': \\'Fish-Weight\\', \\'api_call\\': \"load(\\'path_to_folder/example.pkl\\')\", \\'api_arguments\\': {\\'model_path\\': \\'path_to_folder/example.pkl\\'}, \\'python_environment_requirements\\': {\\'skops.hub_utils\\': \\'download\\', \\'skops.io\\': \\'load\\'}, \\'example_code\\': \"from skops.hub_utils import download\\\\nfrom skops.io import load\\\\ndownload(\\'brendenc/Fish-Weight\\', \\'path_to_folder\\')\\\\nmodel = load(\\'path_to_folder/example.pkl\\')\", \\'performance\\': {\\'dataset\\': \\'Fish dataset\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This is a GradientBoostingRegressor on a fish dataset. This model is intended for educational purposes.\\'}', metadata={})]", "category": "generic"}
{"question_id": 336, "text": " Our company manages a restaurant. We need to analyze customers' tips and predict how much tip a new customer would give based on their total bill, sex, smoker, day, time, and party size. We should use a pre-trained model.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Scikit-learn\\', \\'functionality\\': \\'baseline-trainer\\', \\'api_name\\': \\'merve/tips9y0jvt5q-tip-regression\\', \\'api_call\\': \"pipeline(\\'tabular-regression\\', model=\\'merve/tips9y0jvt5q-tip-regression\\')\", \\'api_arguments\\': \\'N/A\\', \\'python_environment_requirements\\': \\'dabl\\', \\'example_code\\': \\'N/A\\', \\'performance\\': {\\'dataset\\': \\'tips9y0jvt5q\\', \\'accuracy\\': {\\'r2\\': 0.41524, \\'neg_mean_squared_error\\': -1.098792}}, \\'description\\': \\'Baseline Model trained on tips9y0jvt5q to apply regression on tip. The model uses Ridge(alpha=10) and is trained with dabl library as a baseline. For better results, use AutoTrain.\\'}', metadata={})]", "category": "generic"}
{"question_id": 337, "text": " This time, we want to build a player against the user with reinforcement learning. Can you help?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Reinforcement Learning\\', \\'framework\\': \\'ML-Agents\\', \\'functionality\\': \\'SoccerTwos\\', \\'api_name\\': \\'Raiden-1001/poca-Soccerv7\\', \\'api_call\\': \"mlagents-load-from-hf --repo-id=\\'Raiden-1001/poca-Soccerv7.1\\' --local-dir=\\'./downloads\\'\", \\'api_arguments\\': [\\'your_configuration_file_path.yaml\\', \\'run_id\\'], \\'python_environment_requirements\\': [\\'unity-ml-agents\\', \\'deep-reinforcement-learning\\', \\'ML-Agents-SoccerTwos\\'], \\'example_code\\': \\'Step 1: Write your model_id: Raiden-1001/poca-Soccerv7\\\\nStep 2: Select your .nn /.onnx file\\\\nClick on Watch the agent play \ud83d\udc40\\', \\'performance\\': {\\'dataset\\': \\'SoccerTwos\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This is a trained model of a poca agent playing SoccerTwos using the Unity ML-Agents Library.\\'}', metadata={})]", "category": "generic"}
{"question_id": 338, "text": " We are building a robot for indoor navigation. We need to configure it to recognize its surroundings and interact with objects in the environment.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Reinforcement Learning Robotics\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Inference API\\', \\'api_name\\': \\'Antheia/Hanna\\', \\'api_call\\': \"pipeline(\\'robotics\\', model=\\'Antheia/Hanna\\')\", \\'api_arguments\\': \\'model\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'openai/webgpt_comparisons\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Antheia/Hanna is a reinforcement learning model for robotics tasks, trained on the openai/webgpt_comparisons dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 340, "text": " We are building a medical knowledge-based management system. We need to process and extract features from biomedical entity names.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Token Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Named Entity Recognition\\', \\'api_name\\': \\'d4data/biomedical-ner-all\\', \\'api_call\\': \"AutoModelForTokenClassification.from_pretrained(\\'d4data/biomedical-ner-all\\')\", \\'api_arguments\\': {\\'model\\': \\'AutoModelForTokenClassification.from_pretrained(d4data/biomedical-ner-all)\\', \\'tokenizer\\': \\'AutoTokenizer.from_pretrained(d4data/biomedical-ner-all)\\', \\'aggregation_strategy\\': \\'simple\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'from transformers import pipeline, AutoTokenizer, AutoModelForTokenClassification\\'}, \\'example_code\\': \\'pipe(The patient reported no recurrence of palpitations at follow-up 6 months after the ablation.)\\', \\'performance\\': {\\'dataset\\': \\'Maccrobat\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'An English Named Entity Recognition model, trained on Maccrobat to recognize the bio-medical entities (107 entities) from a given text corpus (case reports etc.). This model was built on top of distilbert-base-uncased.\\'}', metadata={})]", "category": "generic"}
{"question_id": 341, "text": " A product is built that analyzes book reviews in order to determine how similar two examples from multiple books are to each other.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'pszemraj/long-t5-tglobal-base-16384-book-summary\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'pszemraj/long-t5-tglobal-base-16384-book-summary\\')\", \\'api_arguments\\': [\\'long_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nimport torch\\\\nsummarizer = pipeline(\\\\n summarization,\\\\n pszemraj/long-t5-tglobal-base-16384-book-summary,\\\\n device=0 if torch.cuda.is_available() else -1,\\\\n)\\\\nlong_text = Here is a lot of text I don\\'t want to read. Replace me\\\\nresult = summarizer(long_text)\\\\nprint(result[0][summary_text])\", \\'performance\\': {\\'dataset\\': \\'kmfoda/booksum\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.408, \\'ROUGE-2\\': 6.065, \\'ROUGE-L\\': 16.721, \\'ROUGE-LSUM\\': 33.34}}, \\'description\\': \\'A fine-tuned version of google/long-t5-tglobal-base on the kmfoda/booksum dataset, which can be used to summarize long text and generate SparkNotes-esque summaries of arbitrary topics. The model generalizes reasonably well to academic and narrative text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 342, "text": " Write a welcome email to a new employee joining the company.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'text2text-generation\\', \\'api_name\\': \\'tuner007/pegasus_summarizer\\', \\'api_call\\': \"PegasusForConditionalGeneration.from_pretrained(\\'tuner007/pegasus_summarizer\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'pip install sentencepiece\\'], \\'example_code\\': \"context = \\\\nIndia wicket-keeper batsman Rishabh Pant has said someone from the crowd threw a ball on pacer Mohammed Siraj while he was fielding in the ongoing third Test against England on Wednesday. Pant revealed the incident made India skipper Virat Kohli upset. I think, somebody threw a ball inside, at Siraj, so he [Kohli] was upset, said Pant in a virtual press conference after the close of the first day\\'s play.You can say whatever you want to chant, but don\\'t throw things at the fielders and all those things. It is not good for cricket, I guess, he added.In the third session of the opening day of the third Test, a section of spectators seemed to have asked Siraj the score of the match to tease the pacer. The India pacer however came with a brilliant reply as he gestured 1-0 (India leading the Test series) towards the crowd.Earlier this month, during the second Test match, there was some bad crowd behaviour on a show as some unruly fans threw champagne corks at India batsman KL Rahul.Kohli also intervened and he was seen gesturing towards the opening batsman to know more about the incident. An over later, the TV visuals showed that many champagne corks were thrown inside the playing field, and the Indian players were visibly left frustrated.Coming back to the game, after bundling out India for 78, openers Rory Burns and Haseeb Hameed ensured that England took the honours on the opening day of the ongoing third Test.At stumps, England\\'s score reads 120/0 and the hosts have extended their lead to 42 runs. For the Three Lions, Burns (52) and Hameed (60) are currently unbeaten at the crease.Talking about the pitch on opening day, Pant said, They took the heavy roller, the wicket was much more settled down, and they batted nicely also, he said. But when we batted, the wicket was slightly soft, and they bowled in good areas, but we could have applied [ourselves] much better.Both England batsmen managed to see off the final session and the hosts concluded the opening day with all ten wickets intact, extending the lead to 42.(ANI)\\\\n\\\\nget_response(context)\", \\'performance\\': {\\'dataset\\': \\'cnn_dailymail\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.604, \\'ROUGE-2\\': 14.64, \\'ROUGE-L\\': 23.884, \\'ROUGE-LSUM\\': 32.902, \\'loss\\': 2.576, \\'gen_len\\': 76.398}}, \\'description\\': \\'PEGASUS fine-tuned for summarization\\'}', metadata={})]", "category": "generic"}
{"question_id": 343, "text": " We need to extract useful features from Russian text for downstream tasks.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Translation\\', \\'api_name\\': \\'opus-mt-en-ru\\', \\'api_call\\': \"AutoModelForSeq2SeqLM.from_pretrained(\\'Helsinki-NLP/opus-mt-en-ru\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\', \\'torch\\', \\'tensorflow\\'], \\'example_code\\': \"from transformers import pipeline\\\\ntranslation = pipeline(\\'translation_en_to_ru\\', model=\\'Helsinki-NLP/opus-mt-en-ru\\')\\\\ntranslated_text = translation(\\'Hello, how are you?\\')\", \\'performance\\': {\\'dataset\\': \\'newstest2019-enru\\', \\'accuracy\\': \\'27.1\\'}, \\'description\\': \\'Helsinki-NLP/opus-mt-en-ru is a translation model trained on the OPUS dataset, which translates English text to Russian. It is based on the Marian NMT framework and can be used with Hugging Face Transformers.\\'}', metadata={})]", "category": "generic"}
{"question_id": 344, "text": " We are working on a project to detect patterns and correlations in a dataset that contains a mix of code segments and comments. Find a suitable model to complete this task.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'kochetkovIT/autotrain-ironhack-49741119788\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'json\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'kochetkovIT/autotrain-data-ironhack\\', \\'accuracy\\': {\\'Loss\\': 2.603, \\'R2\\': 0.013, \\'MSE\\': 6.776, \\'MAE\\': 1.666, \\'RMSLE\\': 0.502}}, \\'description\\': \\'A tabular regression model trained using AutoTrain to predict carbon emissions (in grams) with an R2 score of 0.013.\\'}', metadata={})]", "category": "generic"}
{"question_id": 345, "text": " The marketing department wants to use AI-generated images for their next social media campaign. They want a high-resolution image of a vintage sports car racing through a desert landscape during sunset. \\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'abhishek/autotrain-dog-vs-food\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'abhishek/autotrain-dog-vs-food\\')\", \\'api_arguments\\': \\'image_path\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'sasha/dog-food\\', \\'accuracy\\': 0.998}, \\'description\\': \"A pre-trained model for classifying images as either dog or food using Hugging Face\\'s AutoTrain framework.\"}', metadata={})]", "category": "generic"}
{"question_id": 346, "text": " My company is working on a children's storybook. I need to generate images based on the text descriptions of scenes in the story.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'prompthero/openjourney-v4\\', \\'api_call\\': \"pipeline(\\'text-to-image\\', model=\\'prompthero/openjourney-v4\\')\", \\'api_arguments\\': {\\'text\\': \\'string\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"generate_image(\\'your text here\\')\", \\'performance\\': {\\'dataset\\': \\'Midjourney v4 images\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Openjourney v4 is trained on +124k Midjourney v4 images by PromptHero. It is used for generating images based on text inputs.\\'}', metadata={})]", "category": "generic"}
{"question_id": 347, "text": " Design a text-based interface where you can input text and get an image description from the given text.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'prompthero/openjourney-v4\\', \\'api_call\\': \"pipeline(\\'text-to-image\\', model=\\'prompthero/openjourney-v4\\')\", \\'api_arguments\\': {\\'text\\': \\'string\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"generate_image(\\'your text here\\')\", \\'performance\\': {\\'dataset\\': \\'Midjourney v4 images\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Openjourney v4 is trained on +124k Midjourney v4 images by PromptHero. It is used for generating images based on text inputs.\\'}', metadata={})]", "category": "generic"}
{"question_id": 348, "text": " I am designing an e-commerce website, and I want a program to generate images based on written descriptions to be used as mock product images.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'Lykon/DreamShaper\\', \\'api_call\\': \"pipeline(\\'text-to-image\\', model=Lykon/DreamShaper)\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers, torch\\', \\'example_code\\': \\'https://huggingface.co/spaces/Lykon/DreamShaper-webui\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Dream Shaper is a text-to-image model that generates artistic images based on the given input text. Read more about this model here: https://civitai.com/models/4384/dreamshaper\\'}', metadata={})]", "category": "generic"}
{"question_id": 349, "text": " In our company, we have to generate descriptive captions for photographs related to the products.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Image-to-Text\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Image Captioning\\', \\'api_name\\': \\'blip-image-captioning-large\\', \\'api_call\\': \\'BlipForConditionalGeneration.from_pretrained(Salesforce/blip-image-captioning-large)\\', \\'api_arguments\\': {\\'raw_image\\': \\'Image\\', \\'text\\': \\'Optional Text\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'BlipProcessor, BlipForConditionalGeneration\\', \\'PIL\\': \\'Image\\', \\'requests\\': \\'requests\\'}, \\'example_code\\': {\\'import_requests\\': \\'import requests\\', \\'import_PIL\\': \\'from PIL import Image\\', \\'import_transformers\\': \\'from transformers import BlipProcessor, BlipForConditionalGeneration\\', \\'load_processor\\': \\'processor = BlipProcessor.from_pretrained(Salesforce/blip-image-captioning-large)\\', \\'load_model\\': \\'model = BlipForConditionalGeneration.from_pretrained(Salesforce/blip-image-captioning-large)\\', \\'load_image\\': \"img_url = \\'https://storage.googleapis.com/sfr-vision-language-research/BLIP/demo.jpg\\'\\\\nraw_image = Image.open(requests.get(img_url, stream=True).raw).convert(\\'RGB\\')\", \\'conditional_captioning\\': \\'text = a photography of\\\\ninputs = processor(raw_image, text, return_tensors=pt)\\\\nout = model.generate(**inputs)\\\\nprint(processor.decode(out[0], skip_special_tokens=True))\\', \\'unconditional_captioning\\': \\'inputs = processor(raw_image, return_tensors=pt)\\\\nout = model.generate(**inputs)\\\\nprint(processor.decode(out[0], skip_special_tokens=True))\\'}, \\'performance\\': {\\'dataset\\': \\'COCO\\', \\'accuracy\\': {\\'image-text retrieval\\': \\'+2.7% recall@1\\', \\'image captioning\\': \\'+2.8% CIDEr\\', \\'VQA\\': \\'+1.6% VQA score\\'}}, \\'description\\': \\'BLIP is a Vision-Language Pre-training (VLP) framework that achieves state-of-the-art results on a wide range of vision-language tasks, such as image-text retrieval, image captioning, and VQA. It effectively utilizes noisy web data by bootstrapping the captions, where a captioner generates synthetic captions and a filter removes the noisy ones.\\'}', metadata={})]", "category": "generic"}
{"question_id": 350, "text": " We are building a visual tour guide application for smartphones. The app should be able to identify landmarks and provide information about them. I want to use the BLIP-2 model for this task.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Zero-Shot Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Geolocalization\\', \\'api_name\\': \\'geolocal/StreetCLIP\\', \\'api_call\\': \"CLIPModel.from_pretrained(\\'geolocal/StreetCLIP\\')\", \\'api_arguments\\': {\\'pretrained_model_name_or_path\\': \\'geolocal/StreetCLIP\\'}, \\'python_environment_requirements\\': [\\'transformers\\', \\'PIL\\', \\'requests\\'], \\'example_code\\': \\'from PIL import Image\\\\nimport requests\\\\nfrom transformers import CLIPProcessor, CLIPModel\\\\nmodel = CLIPModel.from_pretrained(geolocal/StreetCLIP)\\\\nprocessor = CLIPProcessor.from_pretrained(geolocal/StreetCLIP)\\\\nurl = https://huggingface.co/geolocal/StreetCLIP/resolve/main/sanfrancisco.jpeg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nchoices = [San Jose, San Diego, Los Angeles, Las Vegas, San Francisco]\\\\ninputs = processor(text=choices, images=image, return_tensors=pt, padding=True)\\\\noutputs = model(**inputs)\\\\nlogits_per_image = outputs.logits_per_image\\\\nprobs = logits_per_image.softmax(dim=1)\\', \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'IM2GPS\\', \\'accuracy\\': {\\'25km\\': 28.3, \\'200km\\': 45.1, \\'750km\\': 74.7, \\'2500km\\': 88.2}}, {\\'name\\': \\'IM2GPS3K\\', \\'accuracy\\': {\\'25km\\': 22.4, \\'200km\\': 37.4, \\'750km\\': 61.3, \\'2500km\\': 80.4}}]}, \\'description\\': \\'StreetCLIP is a robust foundation model for open-domain image geolocalization and other geographic and climate-related tasks. Trained on an original dataset of 1.1 million street-level urban and rural geo-tagged images, it achieves state-of-the-art performance on multiple open-domain image geolocalization benchmarks in zero-shot, outperforming supervised models trained on millions of images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 351, "text": " I need a tool to analyze a chart image and generate a summary of the information contained within the chart.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'google/pegasus-xsum\\', \\'api_call\\': \"pipeline(\\'summarization\\', model=\\'google/pegasus-xsum\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'xsum\\', \\'accuracy\\': {\\'ROUGE-1\\': 46.862, \\'ROUGE-2\\': 24.453, \\'ROUGE-L\\': 39.055, \\'ROUGE-LSUM\\': 39.099}}, {\\'name\\': \\'cnn_dailymail\\', \\'accuracy\\': {\\'ROUGE-1\\': 22.206, \\'ROUGE-2\\': 7.67, \\'ROUGE-L\\': 15.405, \\'ROUGE-LSUM\\': 19.218}}, {\\'name\\': \\'samsum\\', \\'accuracy\\': {\\'ROUGE-1\\': 21.81, \\'ROUGE-2\\': 4.253, \\'ROUGE-L\\': 17.447, \\'ROUGE-LSUM\\': 18.891}}]}, \\'description\\': \\'PEGASUS is a pre-trained model for abstractive summarization, developed by Google. It is based on the Transformer architecture and trained on both C4 and HugeNews datasets. The model is designed to extract gap sentences and generate summaries by stochastically sampling important sentences.\\'}', metadata={})]", "category": "generic"}
{"question_id": 352, "text": " Generate a video for a cooking show intro with the text \\\"Chef John's Culinary Adventures.\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'chavinlo/TempoFunk\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'chavinlo/TempoFunk\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Text-to-Video model using Hugging Face Transformers library. Model is capable of generating video content based on the input text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 353, "text": " We are doing a creative commercial. We require a video that represents a person walking along a beach.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'chavinlo/TempoFunk\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'chavinlo/TempoFunk\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Text-to-Video model using Hugging Face Transformers library. Model is capable of generating video content based on the input text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 354, "text": " Help me to build a self-learning assistant capable of answering questions about an image.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Image-to-Text\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'promptcap-coco-vqa\\', \\'api_call\\': \"PromptCap(\\'vqascore/promptcap-coco-vqa\\')\", \\'api_arguments\\': {\\'prompt\\': \\'string\\', \\'image\\': \\'string\\'}, \\'python_environment_requirements\\': \\'pip install promptcap\\', \\'example_code\\': [\\'import torch\\', \\'from promptcap import PromptCap\\', \\'model = PromptCap(vqascore/promptcap-coco-vqa)\\', \\'if torch.cuda.is_available():\\', \\'  model.cuda()\\', \\'prompt = please describe this image according to the given question: what piece of clothing is this boy putting on?\\', \\'image = glove_boy.jpeg\\', \\'print(model.caption(prompt, image))\\'], \\'performance\\': {\\'dataset\\': {\\'coco\\': {\\'accuracy\\': \\'150 CIDEr\\'}, \\'OK-VQA\\': {\\'accuracy\\': \\'60.4%\\'}, \\'A-OKVQA\\': {\\'accuracy\\': \\'59.6%\\'}}}, \\'description\\': \\'PromptCap is a captioning model that can be controlled by natural language instruction. The instruction may contain a question that the user is interested in. It achieves SOTA performance on COCO captioning (150 CIDEr) and knowledge-based VQA tasks when paired with GPT-3 (60.4% on OK-VQA and 59.6% on A-OKVQA).\\'}', metadata={})]", "category": "generic"}
{"question_id": 355, "text": " We have an online journal application that requires users to upload documents in which we need to automatically answer the questions related to the uploaded document.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 356, "text": " An educational company is building a document question answering system for their students. They want the system to answer questions about the contents of textbooks and study guides quickly and accurately.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 357, "text": " Our company focuses on predicting molecular properties. We need the best model to tackle this problem for our project. We heard about Graphormer and want to start with a pretrained model.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Graph Machine Learning\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'graphormer-base-pcqm4mv1\\', \\'api_call\\': \"AutoModel.from_pretrained(\\'graphormer-base-pcqm4mv1\\')\", \\'api_arguments\\': [\\'model_name\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'See the Graph Classification with Transformers tutorial\\', \\'performance\\': {\\'dataset\\': \\'PCQM4M-LSC\\', \\'accuracy\\': \\'1st place on the KDD CUP 2021 (quantum prediction track)\\'}, \\'description\\': \\'The Graphormer is a graph Transformer model, pretrained on PCQM4M-LSC, and which got 1st place on the KDD CUP 2021 (quantum prediction track). Developed by Microsoft, this model should be used for graph classification tasks or graph representation tasks; the most likely associated task is molecule modeling. It can either be used as such, or finetuned on downstream tasks.\\'}', metadata={})]", "category": "generic"}
{"question_id": 358, "text": " We have a company that processes loan applications. Give us the answer for a question, based on a document.\\n###Input: Our company policy restricts the loan applicant's eligibility to the citizens of United States. The applicant needs to be 18 years old or above and their monthly salary should at least be $4,000. FetchTypeOfYear: 2019. Can anyone with a monthly salary of $3,000 apply?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 359, "text": " Our company is setting up an online shop for selling computer parts. We need a system that can identify the components in images uploaded by users.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Zero-Shot Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Zero-Shot Image Classification\\', \\'api_name\\': \\'clip-vit-base-patch32-ko\\', \\'api_call\\': \"pipeline(\\'zero-shot-image-classification\\', model=\\'Bingsu/clip-vit-base-patch32-ko\\')\", \\'api_arguments\\': {\\'images\\': \\'url\\', \\'candidate_labels\\': \\'Array of strings\\', \\'hypothesis_template\\': \\'String\\'}, \\'python_environment_requirements\\': [\\'transformers\\', \\'torch\\', \\'PIL\\', \\'requests\\'], \\'example_code\\': \"from transformers import pipeline\\\\nrepo = \\'Bingsu/clip-vit-base-patch32-ko\\'\\\\npipe = pipeline(\\'zero-shot-image-classification\\', model=repo)\\\\nurl = \\'http://images.cocodataset.org/val2017/000000039769.jpg\\'\\\\nresult = pipe(images=url, candidate_labels=[], hypothesis_template=\\'{}\\')\\\\nresult\", \\'performance\\': {\\'dataset\\': \\'AIHUB\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Korean CLIP model trained by Making Monolingual Sentence Embeddings Multilingual using Knowledge Distillation. It is a zero-shot image classification model that can be used to classify images without any training data.\\'}', metadata={})]", "category": "generic"}
{"question_id": 360, "text": " I want to classify the images of houseplants to find out their type, like whether it's a cactus, fern, or succulent.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'abhishek/autotrain-dog-vs-food\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'abhishek/autotrain-dog-vs-food\\')\", \\'api_arguments\\': \\'image_path\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'sasha/dog-food\\', \\'accuracy\\': 0.998}, \\'description\\': \"A pre-trained model for classifying images as either dog or food using Hugging Face\\'s AutoTrain framework.\"}', metadata={})]", "category": "generic"}
{"question_id": 361, "text": " I have few picture from my album, and I would like to find out whether it is a hotdog or not.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'julien-c/hotdog-not-hotdog\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'julien-c/hotdog-not-hotdog\\')\", \\'api_arguments\\': \\'image\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': 0.825}, \\'description\\': \\'A model that classifies images as hotdog or not hotdog.\\'}', metadata={})]", "category": "generic"}
{"question_id": 362, "text": " Our client is building an application that converts scanned tables into structured data. We need to detect the rows and columns in a given table image.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'microsoft/table-transformer-structure-recognition\\', \\'api_call\\': \"pipeline(\\'object-detection\\', model=\\'microsoft/table-transformer-structure-recognition\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'PubTables1M\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Table Transformer (DETR) model trained on PubTables1M for detecting the structure (like rows, columns) in tables.\\'}', metadata={})]", "category": "generic"}
{"question_id": 363, "text": " I am hosting a vegan food event and I want to know if any of the dishes contains meat from the picture I have taken.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Image-to-Image\\', \\'api_name\\': \\'GreeneryScenery/SheepsControlV5\\', \\'api_call\\': \"pipeline(\\'image-to-image\\', model=\\'GreeneryScenery/SheepsControlV5\\')\", \\'api_arguments\\': {\\'input_image\\': \\'path/to/image/file\\'}, \\'python_environment_requirements\\': {\\'huggingface_hub\\': \\'>=0.0.17\\', \\'transformers\\': \\'>=4.13.0\\', \\'torch\\': \\'>=1.10.0\\'}, \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'poloclub/diffusiondb\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'SheepsControlV5 is an image-to-image model trained on the poloclub/diffusiondb dataset. It is designed for transforming input images into a different style or representation.\\'}', metadata={})]", "category": "generic"}
{"question_id": 364, "text": " In a kitchen, as a robot chef, we need to identify different objects such as fruits and dishes.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'abhishek/autotrain-dog-vs-food\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'abhishek/autotrain-dog-vs-food\\')\", \\'api_arguments\\': \\'image_path\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'sasha/dog-food\\', \\'accuracy\\': 0.998}, \\'description\\': \"A pre-trained model for classifying images as either dog or food using Hugging Face\\'s AutoTrain framework.\"}', metadata={})]", "category": "generic"}
{"question_id": 365, "text": " We need to segment clothes in photos for our online fashion store.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Segmentation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Segmentation\\', \\'api_name\\': \\'mattmdjaga/segformer_b2_clothes\\', \\'api_call\\': \"SegformerForSemanticSegmentation.from_pretrained(\\'mattmdjaga/segformer_b2_clothes\\')\", \\'api_arguments\\': [\\'image\\'], \\'python_environment_requirements\\': [\\'transformers\\', \\'PIL\\', \\'requests\\', \\'matplotlib\\', \\'torch\\'], \\'example_code\\': \"from transformers import AutoFeatureExtractor, SegformerForSemanticSegmentation\\\\nfrom PIL import Image\\\\nimport requests\\\\nimport matplotlib.pyplot as plt\\\\nimport torch.nn as nn\\\\nextractor = AutoFeatureExtractor.from_pretrained(\\'mattmdjaga/segformer_b2_clothes\\')\\\\nmodel = SegformerForSemanticSegmentation.from_pretrained(\\'mattmdjaga/segformer_b2_clothes\\')\\\\nurl = \\'https://plus.unsplash.com/premium_photo-1673210886161-bfcc40f54d1f?ixlib=rb-4.0.3&amp;ixid=MnwxMjA3fDB8MHxzZWFyY2h8MXx8cGVyc29uJTIwc3RhbmRpbmd8ZW58MHx8MHx8&amp;w=1000&amp;q=80\\'\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\ninputs = extractor(images=image, return_tensors=\\'pt\\')\\\\noutputs = model(**inputs)\\\\nlogits = outputs.logits.cpu()\\\\nupsampled_logits = nn.functional.interpolate(logits, size=image.size[::-1], mode=\\'bilinear\\', align_corners=False)\\\\npred_seg = upsampled_logits.argmax(dim=1)[0]\\\\nplt.imshow(pred_seg)\", \\'performance\\': {\\'dataset\\': \\'mattmdjaga/human_parsing_dataset\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'SegFormer model fine-tuned on ATR dataset for clothes segmentation.\\'}', metadata={})]", "category": "generic"}
{"question_id": 366, "text": " Automate the process of creating a normal map from a given object image.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Normal Map Estimation\\', \\'api_name\\': \\'lllyasviel/sd-controlnet-normal\\', \\'api_call\\': \"ControlNetModel.from_pretrained(\\'lllyasviel/sd-controlnet-normal\\')\", \\'api_arguments\\': [\\'image\\', \\'num_inference_steps\\'], \\'python_environment_requirements\\': [\\'diffusers\\', \\'transformers\\', \\'accelerate\\'], \\'example_code\\': \"from PIL import Image\\\\nfrom transformers import pipeline\\\\nimport numpy as np\\\\nimport cv2\\\\nfrom diffusers import StableDiffusionControlNetPipeline, ControlNetModel, UniPCMultistepScheduler\\\\nimport torch\\\\nfrom diffusers.utils import load_image\\\\nimage = load_image(https://huggingface.co/lllyasviel/sd-controlnet-normal/resolve/main/images/toy.png).convert(RGB)\\\\ndepth_estimator = pipeline(depth-estimation, model =Intel/dpt-hybrid-midas )\\\\nimage = depth_estimator(image)[\\'predicted_depth\\'][0]\\\\nimage = image.numpy()\\\\nimage_depth = image.copy()\\\\nimage_depth -= np.min(image_depth)\\\\nimage_depth /= np.max(image_depth)\\\\nbg_threhold = 0.4\\\\nx = cv2.Sobel(image, cv2.CV_32F, 1, 0, ksize=3)\\\\nx[image_depth &lt; bg_threhold] = 0\\\\ny = cv2.Sobel(image, cv2.CV_32F, 0, 1, ksize=3)\\\\ny[image_depth &lt; bg_threhold] = 0\\\\nz = np.ones_like(x) * np.pi * 2.0\\\\nimage = np.stack([x, y, z], axis=2)\\\\nimage /= np.sum(image ** 2.0, axis=2, keepdims=True) ** 0.5\\\\nimage = (image * 127.5 + 127.5).clip(0, 255).astype(np.uint8)\\\\nimage = Image.fromarray(image)\\\\ncontrolnet = ControlNetModel.from_pretrained(\\\\n fusing/stable-diffusion-v1-5-controlnet-normal, torch_dtype=torch.float16\\\\n)\\\\npipe = StableDiffusionControlNetPipeline.from_pretrained(\\\\n runwayml/stable-diffusion-v1-5, controlnet=controlnet, safety_checker=None, torch_dtype=torch.float16\\\\n)\\\\npipe.scheduler = UniPCMultistepScheduler.from_config(pipe.scheduler.config)\\\\npipe.enable_xformers_memory_efficient_attention()\\\\npipe.enable_model_cpu_offload()\\\\nimage = pipe(cute toy, image, num_inference_steps=20).images[0]\\\\nimage.save(\\'images/toy_normal_out.png\\')\", \\'performance\\': {\\'dataset\\': \\'DIODE\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'ControlNet is a neural network structure to control diffusion models by adding extra conditions. This checkpoint corresponds to the ControlNet conditioned on Normal Map Estimation. It can be used in combination with Stable Diffusion.\\'}', metadata={})]", "category": "generic"}
{"question_id": 367, "text": " We have a picture of a landscape, and we'd like to add a building and a river in this picture.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Image-to-Image\\', \\'api_name\\': \\'GreeneryScenery/SheepsControlV3\\', \\'api_call\\': \"pipeline(\\'image-to-image\\', model=\\'GreeneryScenery/SheepsControlV3\\')\", \\'api_arguments\\': {\\'image\\': \\'Path to image file\\', \\'text_guidance\\': \\'Optional text guidance for the model\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'latest\\', \\'torch\\': \\'latest\\'}, \\'example_code\\': [\\'from transformers import pipeline\\', \"model = pipeline(\\'image-to-image\\', model=\\'GreeneryScenery/SheepsControlV3\\')\", \"result = model({\\'image\\': \\'path/to/image.jpg\\', \\'text_guidance\\': \\'Optional text guidance\\'})\"], \\'performance\\': {\\'dataset\\': \\'GreeneryScenery/SheepsControlV3\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'GreeneryScenery/SheepsControlV3 is a model for image-to-image tasks. It can be used to generate images based on the input image and optional text guidance. The model has some limitations, such as the conditioning image not affecting the output image much. Improvements can be made by training for more epochs, using better prompts, and preprocessing the data.\\'}', metadata={})]", "category": "generic"}
{"question_id": 368, "text": " Our client, an architecture firm, needs a rendering of a \\\"luxury living room with a fireplace\\\" as a visualization for a prestigious project.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'impira/layoutlm-invoices\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=\\'impira/layoutlm-invoices\\')\", \\'api_arguments\\': \\'question, context\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"qa_pipeline(question=\\'your question\\', context=\\'your document context\\')\", \\'performance\\': {\\'dataset\\': \\'proprietary dataset of invoices, SQuAD2.0, and DocVQA\\', \\'accuracy\\': \\'not provided\\'}, \\'description\\': \\'This is a fine-tuned version of the multi-modal LayoutLM model for the task of question answering on invoices and other documents. It has been fine-tuned on a proprietary dataset of invoices as well as both SQuAD2.0 and DocVQA for general comprehension. Unlike other QA models, which can only extract consecutive tokens (because they predict the start and end of a sequence), this model can predict longer-range, non-consecutive sequences with an additional classifier head.\\'}', metadata={})]", "category": "generic"}
{"question_id": 369, "text": " We would like to restore an old and blurry image of our grandparents for an anniversary present.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image-to-Image\\', \\'framework\\': \\'Keras\\', \\'functionality\\': \\'Image Deblurring\\', \\'api_name\\': \\'google/maxim-s3-deblurring-gopro\\', \\'api_call\\': \"from_pretrained_keras(\\'google/maxim-s3-deblurring-gopro\\')\", \\'api_arguments\\': [\\'image\\'], \\'python_environment_requirements\\': [\\'huggingface_hub\\', \\'PIL\\', \\'tensorflow\\', \\'numpy\\', \\'requests\\'], \\'example_code\\': \\'from huggingface_hub import from_pretrained_keras\\\\nfrom PIL import Image\\\\nimport tensorflow as tf\\\\nimport numpy as np\\\\nimport requests\\\\nurl = https://github.com/sayakpaul/maxim-tf/raw/main/images/Deblurring/input/1fromGOPR0950.png\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nimage = np.array(image)\\\\nimage = tf.convert_to_tensor(image)\\\\nimage = tf.image.resize(image, (256, 256))\\\\nmodel = from_pretrained_keras(google/maxim-s3-deblurring-gopro)\\\\npredictions = model.predict(tf.expand_dims(image, 0))\\', \\'performance\\': {\\'dataset\\': \\'GoPro\\', \\'accuracy\\': {\\'PSNR\\': 32.86, \\'SSIM\\': 0.961}}, \\'description\\': \\'MAXIM model pre-trained for image deblurring. It was introduced in the paper MAXIM: Multi-Axis MLP for Image Processing by Zhengzhong Tu, Hossein Talebi, Han Zhang, Feng Yang, Peyman Milanfar, Alan Bovik, Yinxiao Li and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 370, "text": " A creative artist wants to generate a painting from a given input text titled \\\"A head full of roses,\\\" utilizing the image provided in the API description.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Image-to-Text\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Captioning\\', \\'api_name\\': \\'microsoft/git-base\\', \\'api_call\\': \"pipeline(\\'image-to-text\\', model=\\'microsoft/git-base\\')\", \\'api_arguments\\': \\'image\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'git_base(image)\\', \\'performance\\': {\\'dataset\\': [\\'COCO\\', \\'Conceptual Captions (CC3M)\\', \\'SBU\\', \\'Visual Genome (VG)\\', \\'Conceptual Captions (CC12M)\\', \\'ALT200M\\'], \\'accuracy\\': \\'Refer to the paper for evaluation results\\'}, \\'description\\': \"GIT (short for GenerativeImage2Text) model, base-sized version. It was introduced in the paper GIT: A Generative Image-to-text Transformer for Vision and Language by Wang et al. and first released in this repository. The model is trained using \\'teacher forcing\\' on a lot of (image, text) pairs. The goal for the model is simply to predict the next text token, giving the image tokens and previous text tokens. This allows the model to be used for tasks like image and video captioning, visual question answering (VQA) on images and videos, and even image classification (by simply conditioning the model on the image and asking it to generate a class for it in text).\"}', metadata={})]", "category": "generic"}
{"question_id": 371, "text": " I would like to generate a high-quality image of a church using unconditional image generation.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Unconditional Image Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Unconditional Image Generation\\', \\'api_name\\': \\'google/ncsnpp-church-256\\', \\'api_call\\': \"DiffusionPipeline.from_pretrained(\\'google/ncsnpp-church-256\\')\", \\'api_arguments\\': \\'model_id\\', \\'python_environment_requirements\\': \\'diffusers\\', \\'example_code\\': \\'!pip install diffusers\\\\nfrom diffusers import DiffusionPipeline\\\\nmodel_id = google/ncsnpp-church-256\\\\nsde_ve = DiffusionPipeline.from_pretrained(model_id)\\\\nimage = sde_ve()[sample]\\\\nimage[0].save(sde_ve_generated_image.png)\\', \\'performance\\': {\\'dataset\\': \\'CIFAR-10\\', \\'accuracy\\': {\\'Inception_score\\': 9.89, \\'FID\\': 2.2, \\'likelihood\\': 2.99}}, \\'description\\': \\'Score-Based Generative Modeling through Stochastic Differential Equations (SDE) for unconditional image generation. This model achieves record-breaking performance on CIFAR-10 and can generate high fidelity images of size 1024 x 1024.\\'}', metadata={})]", "category": "generic"}
{"question_id": 372, "text": " Recently, my company is asked to produce new portraits for the entire team. Therefore, I am looking for an AI which can generate human faces from scratch.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Conversational\\', \\'api_name\\': \\'facebook/blenderbot-3B\\', \\'api_call\\': \"BlenderbotForConditionalGeneration.from_pretrained(\\'facebook/blenderbot-3B\\')\", \\'api_arguments\\': [\\'model_name\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'Input a message to start chatting with facebook/blenderbot-3B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'BlenderBot-3B is a large-scale neural model designed for open-domain chatbot applications. It is trained on the blended_skill_talk dataset and can engage in multi-turn conversations, providing engaging talking points, asking and answering questions, and displaying knowledge, empathy, and personality. The model is available through the Hugging Face Transformers library.\\'}', metadata={})]", "category": "generic"}
{"question_id": 373, "text": " Our team wants to make a cat-themed video game. We would like to design various cartoon cat characters through AI-generated images.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'saltacc/anime-ai-detect\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'saltacc/anime-ai-detect\\')\", \\'api_arguments\\': [\\'image\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'aibooru and imageboard sites\\', \\'accuracy\\': \\'96%\\'}, \\'description\\': \\'A BEiT classifier to see if anime art was made by an AI or a human.\\'}', metadata={})]", "category": "generic"}
{"question_id": 374, "text": " Can you teach me to build a program that will analyze video and describe what's happening in natural language?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'ImRma/Brucelee\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'ImRma/Brucelee\\')\", \\'api_arguments\\': [\\'your_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Hugging Face model for converting Persian and English text into video.\\'}', metadata={})]", "category": "generic"}
{"question_id": 375, "text": " In a security context, the company is starting a project to detect any signs of violence in video streams from its CCTV cameras. It's essential to analyze the video content for any violent behaviors.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Video Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'videomae-base-finetuned-RealLifeViolenceSituations-subset\\', \\'api_call\\': \"AutoModelForVideoClassification.from_pretrained(\\'dangle124/videomae-base-finetuned-RealLifeViolenceSituations-subset\\')\", \\'api_arguments\\': {\\'model_name\\': \\'dangle124/videomae-base-finetuned-RealLifeViolenceSituations-subset\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'4.27.2\\', \\'pytorch\\': \\'1.13.1\\', \\'datasets\\': \\'2.10.1\\', \\'tokenizers\\': \\'0.13.2\\'}, \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'unknown\\', \\'accuracy\\': 0.9533}, \\'description\\': \\'This model is a fine-tuned version of MCG-NJU/videomae-base on an unknown dataset. It is trained for video classification task, specifically for RealLifeViolenceSituations.\\'}', metadata={})]", "category": "generic"}
{"question_id": 376, "text": " Create a sports highlight generator that takes video clips as input and identifies the category of the sports activity happening in the video.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Zero-Shot Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Zero-Shot Image Classification\\', \\'api_name\\': \\'patrickjohncyh/fashion-clip\\', \\'api_call\\': \"CLIPModel.from_pretrained(\\'patrickjohncyh/fashion-clip\\')\", \\'api_arguments\\': {\\'image\\': \\'File\\', \\'class_names\\': \\'String (comma-separated)\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import CLIPProcessor, CLIPModel; model = CLIPModel.from_pretrained(\\'patrickjohncyh/fashion-clip\\'); processor = CLIPProcessor.from_pretrained(\\'patrickjohncyh/fashion-clip\\'); inputs = processor(text=\\'blue shoes\\', images=image, return_tensors=\\'pt\\', padding=True); logits_per_image = model(**inputs).logits_per_image; probs = logits_per_image.softmax(dim=-1).tolist()[0]\", \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'FMNIST\\', \\'accuracy\\': 0.83}, {\\'name\\': \\'KAGL\\', \\'accuracy\\': 0.73}, {\\'name\\': \\'DEEP\\', \\'accuracy\\': 0.62}]}, \\'description\\': \\'FashionCLIP is a CLIP-based model developed to produce general product representations for fashion concepts. Leveraging the pre-trained checkpoint (ViT-B/32) released by OpenAI, it is trained on a large, high-quality novel fashion dataset to study whether domain specific fine-tuning of CLIP-like models is sufficient to produce product representations that are zero-shot transferable to entirely new datasets and tasks.\\'}', metadata={})]", "category": "generic"}
{"question_id": 377, "text": " I am a movie director and I need to detect the genre of a movie based on its actions.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Joblib\\', \\'functionality\\': \\'Multi-class Classification\\', \\'api_name\\': \\'Alexei1/imdb\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'json\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'IMDB\\', \\'accuracy\\': 0.487}, \\'description\\': \\'A tabular classification model trained using AutoTrain for sentiment analysis on the IMDB dataset. The model has a CO2 emission of 0.0186 grams and an accuracy of 0.487.\\'}', metadata={})]", "category": "generic"}
{"question_id": 378, "text": " Bob is designing an app for his city. He needs to quickly identify if the provided image has a bike or a car. Design a model that recognizes a bike or car and provides classification output for the given image.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'abhishek/autotrain-dog-vs-food\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'abhishek/autotrain-dog-vs-food\\')\", \\'api_arguments\\': \\'image_path\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'sasha/dog-food\\', \\'accuracy\\': 0.998}, \\'description\\': \"A pre-trained model for classifying images as either dog or food using Hugging Face\\'s AutoTrain framework.\"}', metadata={})]", "category": "generic"}
{"question_id": 379, "text": " I need to analyze the sentiment of tweets to find out which products are getting positive customer feedback.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'cardiffnlp/twitter-roberta-base-sentiment-latest\\', \\'api_call\\': \"pipeline(sentiment-analysis, model=AutoModel.from_pretrained(\\'cardiffnlp/twitter-roberta-base-sentiment-latest\\'), tokenizer=AutoTokenizer.from_pretrained(\\'cardiffnlp/twitter-roberta-base-sentiment-latest\\'))\", \\'api_arguments\\': {\\'model\\': \\'model_path\\', \\'tokenizer\\': \\'model_path\\'}, \\'python_environment_requirements\\': [\\'transformers\\', \\'numpy\\', \\'scipy\\'], \\'example_code\\': \\'from transformers import pipeline\\\\nsentiment_task = pipeline(sentiment-analysis, model=model_path, tokenizer=model_path)\\\\nsentiment_task(Covid cases are increasing fast!)\\', \\'performance\\': {\\'dataset\\': \\'tweet_eval\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This is a RoBERTa-base model trained on ~124M tweets from January 2018 to December 2021, and finetuned for sentiment analysis with the TweetEval benchmark. The model is suitable for English.\\'}', metadata={})]", "category": "generic"}
{"question_id": 380, "text": " We are building a content moderation system. Our clients upload the content, it can be generated by human or AI. We want to have a filtering API to advise on the original text if it is generated by GPT-2.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Generation\\', \\'api_name\\': \\'gpt2\\', \\'api_call\\': \"pipeline(\\'text-generation\\', model=\\'gpt2\\')\", \\'api_arguments\\': [\\'text\\', \\'max_length\\', \\'num_return_sequences\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline, set_seed\\\\ngenerator = pipeline(\\'text-generation\\', model=\\'gpt2\\')\\\\nset_seed(42)\\\\ngenerator(Hello, I\\'m a language model,, max_length=30, num_return_sequences=5)\", \\'performance\\': {\\'dataset\\': {\\'LAMBADA\\': {\\'accuracy\\': \\'35.13\\'}, \\'CBT-CN\\': {\\'accuracy\\': \\'45.99\\'}, \\'CBT-NE\\': {\\'accuracy\\': \\'87.65\\'}, \\'WikiText2\\': {\\'accuracy\\': \\'83.4\\'}, \\'PTB\\': {\\'accuracy\\': \\'29.41\\'}, \\'enwiki8\\': {\\'accuracy\\': \\'65.85\\'}, \\'text8\\': {\\'accuracy\\': \\'1.16\\'}, \\'WikiText103\\': {\\'accuracy\\': \\'1.17\\'}, \\'1BW\\': {\\'accuracy\\': \\'37.50\\'}}}, \\'description\\': \\'GPT-2 is a transformers model pretrained on a very large corpus of English data in a self-supervised fashion. This means it was pretrained on the raw texts only, with no humans labelling them in any way (which is why it can use lots of publicly available data) with an automatic process to generate inputs and labels from those texts. More precisely, it was trained to guess the next word in sentences.\\'}', metadata={})]", "category": "generic"}
{"question_id": 381, "text": " I have a list of comments related to stocks. Check the overall sentiment towards each stock.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Sentiment Inferencing for stock-related comments\\', \\'api_name\\': \\'zhayunduo/roberta-base-stocktwits-finetuned\\', \\'api_call\\': \"RobertaForSequenceClassification.from_pretrained(\\'zhayunduo/roberta-base-stocktwits-finetuned\\')\", \\'api_arguments\\': {\\'model\\': \\'RobertaForSequenceClassification\\', \\'tokenizer\\': \\'RobertaTokenizer\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import RobertaForSequenceClassification, RobertaTokenizer\\\\nfrom transformers import pipeline\\\\nimport pandas as pd\\\\nimport emoji\\\\ntokenizer_loaded = RobertaTokenizer.from_pretrained(\\'zhayunduo/roberta-base-stocktwits-finetuned\\')\\\\nmodel_loaded = RobertaForSequenceClassification.from_pretrained(\\'zhayunduo/roberta-base-stocktwits-finetuned\\')\\\\nnlp = pipeline(text-classification, model=model_loaded, tokenizer=tokenizer_loaded)\\\\nsentences = pd.Series([\\'just buy\\',\\'just sell it\\',\\'entity rocket to the sky!\\',\\'go down\\',\\'even though it is going up, I still think it will not keep this trend in the near future\\'])\\\\nsentences = list(sentences)\\\\nresults = nlp(sentences)\\\\nprint(results)\", \\'performance\\': {\\'dataset\\': \\'stocktwits\\', \\'accuracy\\': 0.9343}, \\'description\\': \"This model is fine-tuned with roberta-base model on 3,200,000 comments from stocktwits, with the user-labeled tags \\'Bullish\\' or \\'Bearish\\'.\"}', metadata={})]", "category": "generic"}
{"question_id": 382, "text": " I'd like to give a potted plant to my friend. Write me a short and simple plant care instruction.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Paraphrasing\\', \\'api_name\\': \\'prithivida/parrot_paraphraser_on_T5\\', \\'api_call\\': \"Parrot(model_tag=\\'prithivida/parrot_paraphraser_on_T5\\', use_gpu=False)\", \\'api_arguments\\': [\\'input_phrase\\', \\'diversity_ranker\\', \\'do_diverse\\', \\'max_return_phrases\\', \\'max_length\\', \\'adequacy_threshold\\', \\'fluency_threshold\\'], \\'python_environment_requirements\\': [\\'torch\\', \\'transformers\\'], \\'example_code\\': \\'from parrot import Parrot\\\\nimport torch\\\\nimport warnings\\\\nwarnings.filterwarnings(ignore)\\\\n\\\\nparrot = Parrot(model_tag=prithivida/parrot_paraphraser_on_T5, use_gpu=False)\\\\nphrases = [Can you recommed some upscale restaurants in Newyork?,\\\\n What are the famous places we should not miss in Russia?\\\\n]\\\\nfor phrase in phrases:\\\\n print(-*100)\\\\n print(Input_phrase: , phrase)\\\\n print(-*100)\\\\n para_phrases = parrot.augment(input_phrase=phrase)\\\\n for para_phrase in para_phrases:\\\\n  print(para_phrase)\\', \\'performance\\': {\\'dataset\\': \\'Not mentioned\\', \\'accuracy\\': \\'Not mentioned\\'}, \\'description\\': \\'Parrot is a paraphrase based utterance augmentation framework purpose built to accelerate training NLU models. It offers knobs to control Adequacy, Fluency, and Diversity as per your needs. It mainly focuses on augmenting texts typed-into or spoken-to conversational interfaces for building robust NLU models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 383, "text": " I need to extract biomedical entities from a given set of case reports to make it easier for researchers to analyze the data.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Token Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Named Entity Recognition\\', \\'api_name\\': \\'d4data/biomedical-ner-all\\', \\'api_call\\': \"AutoModelForTokenClassification.from_pretrained(\\'d4data/biomedical-ner-all\\')\", \\'api_arguments\\': {\\'model\\': \\'AutoModelForTokenClassification.from_pretrained(d4data/biomedical-ner-all)\\', \\'tokenizer\\': \\'AutoTokenizer.from_pretrained(d4data/biomedical-ner-all)\\', \\'aggregation_strategy\\': \\'simple\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'from transformers import pipeline, AutoTokenizer, AutoModelForTokenClassification\\'}, \\'example_code\\': \\'pipe(The patient reported no recurrence of palpitations at follow-up 6 months after the ablation.)\\', \\'performance\\': {\\'dataset\\': \\'Maccrobat\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'An English Named Entity Recognition model, trained on Maccrobat to recognize the bio-medical entities (107 entities) from a given text corpus (case reports etc.). This model was built on top of distilbert-base-uncased.\\'}', metadata={})]", "category": "generic"}
{"question_id": 384, "text": " I need an AI feature capable of extracting names of people, organizations, and locations from various news articles in different languages.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'finiteautomata/beto-sentiment-analysis\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'finiteautomata/beto-sentiment-analysis\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'Hugging Face Transformers library\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'TASS 2020 corpus\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Model trained with TASS 2020 corpus (around ~5k tweets) of several dialects of Spanish. Base model is BETO, a BERT model trained in Spanish. Uses POS, NEG, NEU labels.\\'}', metadata={})]", "category": "generic"}
{"question_id": 385, "text": " We are a consulting firm, and we want to easily identify company names from texts.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Text Classification\\', \\'api_name\\': \\'joeddav/distilbert-base-uncased-go-emotions-student\\', \\'api_call\\': \"pipeline(\\'text-classification\\', model=\\'joeddav/distilbert-base-uncased-go-emotions-student\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': [\\'transformers\\', \\'torch\\', \\'tensorflow\\'], \\'example_code\\': \"from transformers import pipeline\\\\nnlp = pipeline(\\'text-classification\\', model=\\'joeddav/distilbert-base-uncased-go-emotions-student\\')\\\\nresult = nlp(\\'I am so happy today!\\')\", \\'performance\\': {\\'dataset\\': \\'go_emotions\\'}, \\'description\\': \\'This model is distilled from the zero-shot classification pipeline on the unlabeled GoEmotions dataset. It is primarily intended as a demo of how an expensive NLI-based zero-shot model can be distilled to a more efficient student, allowing a classifier to be trained with only unlabeled data.\\'}', metadata={})]", "category": "generic"}
{"question_id": 386, "text": " Our company is making a chatbot that needs to extract information from a paragraph. Get the named entities in the paragraph.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Generation\\', \\'api_name\\': \\'mywateriswet/ShuanBot\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'mywateriswet/ShuanBot\\')\", \\'api_arguments\\': \\'message\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"response = chatbot(\\'What is your name?\\')\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'ShuanBot is a conversational chatbot model based on the GPT-2 architecture. It can be used for generating human-like responses in a chat context.\\'}', metadata={})]", "category": "generic"}
{"question_id": 387, "text": " A journalist is looking for historical Olympic host cities and wants to find the year when Beijing hosted the games.\\n###Input: {\\\"table\\\": {\\n    \\\"year\\\": [1896, 1900, 1904, 2004, 2008, 2012],\\n    \\\"city\\\": [\\\"Athens\\\", \\\"Paris\\\", \\\"St. Louis\\\", \\\"Athens\\\", \\\"Beijing\\\", \\\"London\\\"]},\\n    \\\"query\\\": \\\"Select the year when Beijing hosted the Olympic games\\\"}\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Table Question Answering\\', \\'framework\\': \\'PyTorch Transformers\\', \\'functionality\\': \\'Table-based QA\\', \\'api_name\\': \\'neulab/omnitab-large-1024shot\\', \\'api_call\\': \"AutoModelForSeq2SeqLM.from_pretrained(\\'neulab/omnitab-large-1024shot\\')\", \\'api_arguments\\': {\\'table\\': \\'pd.DataFrame.from_dict(data)\\', \\'query\\': \\'str\\'}, \\'python_environment_requirements\\': [\\'transformers\\', \\'pandas\\'], \\'example_code\\': \\'from transformers import AutoTokenizer, AutoModelForSeq2SeqLM\\\\nimport pandas as pd\\\\ntokenizer = AutoTokenizer.from_pretrained(neulab/omnitab-large-1024shot)\\\\nmodel = AutoModelForSeq2SeqLM.from_pretrained(neulab/omnitab-large-1024shot)\\\\ndata = {\\\\n year: [1896, 1900, 1904, 2004, 2008, 2012],\\\\n city: [athens, paris, st. louis, athens, beijing, london]\\\\n}\\\\ntable = pd.DataFrame.from_dict(data)\\\\nquery = In which year did beijing host the Olympic Games?\\\\nencoding = tokenizer(table=table, query=query, return_tensors=pt)\\\\noutputs = model.generate(**encoding)\\\\nprint(tokenizer.batch_decode(outputs, skip_special_tokens=True))\\', \\'performance\\': {\\'dataset\\': \\'wikitablequestions\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'OmniTab is a table-based QA model proposed in OmniTab: Pretraining with Natural and Synthetic Data for Few-shot Table-based Question Answering. neulab/omnitab-large-1024shot (based on BART architecture) is initialized with microsoft/tapex-large and continuously pretrained on natural and synthetic data (SQL2NL model trained in the 1024-shot setting).\\'}', metadata={})]", "category": "generic"}
{"question_id": 388, "text": " I am a Korean exchange student in the United States. I am building a Korean Table Question Answering assistant to help me with my studies.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Table Question Answering\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'dsba-lab/koreapas-finetuned-korwikitq\\', \\'api_call\\': \"pipeline(\\'table-question-answering\\', model=\\'dsba-lab/koreapas-finetuned-korwikitq\\')\", \\'api_arguments\\': {}, \\'python_environment_requirements\\': {\\'transformers\\': \\'>=4.0.0\\'}, \\'example_code\\': \"from transformers import pipeline; table_qa = pipeline(\\'table-question-answering\\', model=\\'dsba-lab/koreapas-finetuned-korwikitq\\'); table_qa(table=table, query=\\'\uc9c8\ubb38\\')\", \\'performance\\': {\\'dataset\\': \\'korwikitq\\', \\'accuracy\\': None}, \\'description\\': \\'A Korean Table Question Answering model finetuned on the korwikitq dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 389, "text": " Create an API that processes large data sets of financial transactions and can deliver information on the number of transactions and their monetary value, based on a date range.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'text2text-generation\\', \\'api_name\\': \\'financial-summarization-pegasus\\', \\'api_call\\': \"PegasusForConditionalGeneration.from_pretrained(\\'human-centered-summarization/financial-summarization-pegasus\\')\", \\'api_arguments\\': [\\'model_name\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'from transformers import PegasusTokenizer, PegasusForConditionalGeneration, TFPegasusForConditionalGeneration\\\\nmodel_name = human-centered-summarization/financial-summarization-pegasus\\\\ntokenizer = PegasusTokenizer.from_pretrained(model_name)\\\\nmodel = PegasusForConditionalGeneration.from_pretrained(model_name)\\\\ntext_to_summarize = National Commercial Bank (NCB), Saudi Arabia\u2019s largest lender by assets, agreed to buy rival Samba Financial Group for $15 billion in the biggest banking takeover this year.NCB will pay 28.45 riyals ($7.58) for each Samba share, according to a statement on Sunday, valuing it at about 55.7 billion riyals. NCB will offer 0.739 new shares for each Samba share, at the lower end of the 0.736-0.787 ratio the banks set when they signed an initial framework agreement in June.The offer is a 3.5% premium to Samba\u2019s Oct. 8 closing price of 27.50 riyals and about 24% higher than the level the shares traded at before the talks were made public. Bloomberg News first reported the merger discussions.The new bank will have total assets of more than $220 billion, creating the Gulf region\u2019s third-largest lender. The entity\u2019s $46 billion market capitalization nearly matches that of Qatar National Bank QPSC, which is still the Middle East\u2019s biggest lender with about $268 billion of assets.\\\\ninput_ids = tokenizer(text_to_summarize, return_tensors=pt).input_ids\\\\noutput = model.generate(input_ids, max_length=32, num_beams=5, early_stopping=True)\\\\nprint(tokenizer.decode(output[0], skip_special_tokens=True))\\', \\'performance\\': {\\'dataset\\': \\'xsum\\', \\'accuracy\\': {\\'ROUGE-1\\': 35.206, \\'ROUGE-2\\': 16.569, \\'ROUGE-L\\': 30.128, \\'ROUGE-LSUM\\': 30.171}}, \\'description\\': \\'This model was fine-tuned on a novel financial news dataset, which consists of 2K articles from Bloomberg, on topics such as stock, markets, currencies, rate and cryptocurrencies. It is based on the PEGASUS model and in particular PEGASUS fine-tuned on the Extreme Summarization (XSum) dataset: google/pegasus-xsum model. PEGASUS was originally proposed by Jingqing Zhang, Yao Zhao, Mohammad Saleh and Peter J. Liu in PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization.\\'}', metadata={})]", "category": "generic"}
{"question_id": 390, "text": " A product manager needs an explanation related to model conversion. They are confused why it is important. Can you please help them by providing an answer?\\n###Input: \\n{\\n 'question': 'Why is model conversion important?',\\n 'context': 'The option to convert models between FARM and transformers gives freedom to the user and let people easily switch between frameworks.'\\n}\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'deepset/tinyroberta-squad2\\', \\'api_call\\': \"AutoModelForQuestionAnswering.from_pretrained(\\'deepset/tinyroberta-squad2\\')\", \\'api_arguments\\': {\\'model_name_or_path\\': \\'deepset/tinyroberta-squad2\\', \\'question\\': \\'Why is model conversion important?\\', \\'context\\': \\'The option to convert models between FARM and transformers gives freedom to the user and let people easily switch between frameworks.\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import AutoModelForQuestionAnswering, AutoTokenizer, pipeline\\\\nmodel_name = deepset/tinyroberta-squad2\\\\nnlp = pipeline(\\'question-answering\\', model=model_name, tokenizer=model_name)\\\\nQA_input = {\\\\n \\'question\\': \\'Why is model conversion important?\\',\\\\n \\'context\\': \\'The option to convert models between FARM and transformers gives freedom to the user and let people easily switch between frameworks.\\'\\\\n}\\\\nres = nlp(QA_input)\", \\'performance\\': {\\'dataset\\': \\'squad_v2\\', \\'accuracy\\': {\\'exact\\': 78.69114798281817, \\'f1\\': 81.9198998536977}}, \\'description\\': \\'This is the distilled version of the deepset/roberta-base-squad2 model. This model has a comparable prediction quality and runs at twice the speed of the base model.\\'}', metadata={})]", "category": "generic"}
{"question_id": 391, "text": " We are building a memory game where a description is displayed for a few seconds and later on, questions on what is shown comes up for the user to remember.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Visual Question Answering\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Visual Question Answering\\', \\'api_name\\': \\'temp_vilt_vqa\\', \\'api_call\\': \"pipeline(\\'visual-question-answering\\', model=\\'Bingsu/temp_vilt_vqa\\', tokenizer=\\'Bingsu/temp_vilt_vqa\\')\", \\'api_arguments\\': {\\'model\\': \\'Bingsu/temp_vilt_vqa\\', \\'tokenizer\\': \\'Bingsu/temp_vilt_vqa\\'}, \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A visual question answering model for answering questions related to images using the Hugging Face Transformers library.\\'}', metadata={})]", "category": "generic"}
{"question_id": 392, "text": " We are a medical research team working on a COVID-19 project. We need assistance in question answering related to the pandemic and related research papers.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'deepset/roberta-base-squad2-covid\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=RobertaForQuestionAnswering.from_pretrained(\\'deepset/roberta-base-squad2-covid\\'), tokenizer=RobertaTokenizer.from_pretrained(\\'deepset/roberta-base-squad2-covid\\'))\", \\'api_arguments\\': {\\'model_name\\': \\'deepset/roberta-base-squad2-covid\\', \\'tokenizer\\': \\'deepset/roberta-base-squad2-covid\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': {\\'QA_input\\': {\\'question\\': \\'Why is model conversion important?\\', \\'context\\': \\'The option to convert models between FARM and transformers gives freedom to the user and let people easily switch between frameworks.\\'}, \\'res\\': \\'nlp(QA_input)\\'}, \\'performance\\': {\\'dataset\\': \\'squad_v2\\', \\'accuracy\\': {\\'XVAL_EM\\': 0.17890995260663506, \\'XVAL_f1\\': 0.49925444207319924, \\'XVAL_top_3_recall\\': 0.8021327014218009}}, \\'description\\': \\'This model is a Roberta-based model fine-tuned on SQuAD-style CORD-19 annotations for the task of extractive question answering in the context of COVID-19. It can be used with the Hugging Face Transformers library for question answering tasks.\\'}', metadata={})]", "category": "generic"}
{"question_id": 393, "text": " The customer is writing a book about the solar system and just needs the answer to a question: How long does it take for Mars to orbit the sun?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Audio-to-Audio\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Asteroid\\', \\'api_name\\': \\'ConvTasNet_Libri2Mix_sepclean_16k\\', \\'api_call\\': \"Asteroid(\\'JorisCos/ConvTasNet_Libri2Mix_sepclean_16k\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'asteroid\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'Libri2Mix\\', \\'accuracy\\': {\\'si_sdr\\': 15.243671356901526, \\'si_sdr_imp\\': 15.243034178473609, \\'sdr\\': 15.668108919568112, \\'sdr_imp\\': 15.578229918028036, \\'sir\\': 25.295100756629957, \\'sir_imp\\': 25.205219921301754, \\'sar\\': 16.307682590197313, \\'sar_imp\\': -51.64989963759405, \\'stoi\\': 0.9394951175291422, \\'stoi_imp\\': 0.22640192740016568}}, \\'description\\': \\'This model was trained by Joris Cosentino using the librimix recipe in Asteroid. It was trained on the sep_clean task of the Libri2Mix dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 394, "text": " I am designing a quiz app that caters to blind users. The app should interpret the images and answer questions based on the image contents.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Visual Question Answering\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Visual Question Answering\\', \\'api_name\\': \\'blip-vqa-base\\', \\'api_call\\': \"BlipForQuestionAnswering.from_pretrained(\\'Salesforce/blip-vqa-base\\')\", \\'api_arguments\\': {\\'raw_image\\': \\'Image\\', \\'question\\': \\'String\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'BlipProcessor, BlipForQuestionAnswering\\', \\'PIL\\': \\'Image\\', \\'requests\\': \\'requests\\'}, \\'example_code\\': \"import requests\\\\nfrom PIL import Image\\\\nfrom transformers import BlipProcessor, BlipForQuestionAnswering\\\\nprocessor = BlipProcessor.from_pretrained(Salesforce/blip-vqa-base)\\\\nmodel = BlipForQuestionAnswering.from_pretrained(Salesforce/blip-vqa-base)\\\\nimg_url = \\'https://storage.googleapis.com/sfr-vision-language-research/BLIP/demo.jpg\\'\\\\nraw_image = Image.open(requests.get(img_url, stream=True).raw).convert(\\'RGB\\')\\\\nquestion = how many dogs are in the picture?\\\\ninputs = processor(raw_image, question, return_tensors=pt)\\\\nout = model.generate(**inputs)\\\\nprint(processor.decode(out[0], skip_special_tokens=True))\", \\'performance\\': {\\'dataset\\': \\'VQA\\', \\'accuracy\\': \\'+1.6% in VQA score\\'}, \\'description\\': \\'BLIP is a Vision-Language Pre-training (VLP) framework that transfers flexibly to both vision-language understanding and generation tasks. It effectively utilizes noisy web data by bootstrapping the captions, where a captioner generates synthetic captions and a filter removes the noisy ones. This model is trained on visual question answering with a base architecture (using ViT base backbone).\\'}', metadata={})]", "category": "generic"}
{"question_id": 395, "text": " The company develops a digital assistant that can answer questions about software products. Implement a feature to provide answers to user questions.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'distilbert-base-uncased-distilled-squad\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=\\'distilbert-base-uncased-distilled-squad\\')\", \\'api_arguments\\': [\\'question\\', \\'context\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nquestion_answerer = pipeline(question-answering, model=\\'distilbert-base-uncased-distilled-squad\\')\\\\ncontext = r\\\\n... Extractive Question Answering is the task of extracting an answer from a text given a question. An example of a\\\\n... question answering dataset is the SQuAD dataset, which is entirely based on that task. If you would like to fine-tune\\\\n... a model on a SQuAD task, you may leverage the examples/pytorch/question-answering/run_squad.py script.\\\\n... \\\\nresult = question_answerer(question=What is a good example of a question answering dataset?, context=context)\\\\nprint(\\\\n... fAnswer: \\'{result[\\'answer\\']}\\', score: {round(result[\\'score\\'], 4)}, start: {result[\\'start\\']}, end: {result[\\'end\\']}\\\\n...)\", \\'performance\\': {\\'dataset\\': \\'SQuAD v1.1\\', \\'accuracy\\': \\'86.9 F1 score\\'}, \\'description\\': \"DistilBERT base uncased distilled SQuAD is a fine-tuned version of DistilBERT-base-uncased for the task of question answering. It has 40% less parameters than bert-base-uncased, runs 60% faster while preserving over 95% of BERT\\'s performances as measured on the GLUE language understanding benchmark.\"}', metadata={})]", "category": "generic"}
{"question_id": 396, "text": " I'm talking to a new person online. Can this API help keep me safe by filtering out any inappropriate messages they send me?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'michellejieli/NSFW_text_classifier\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'michellejieli/NSFW_text_classification\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'classifier(I see you\u2019ve set aside this special time to humiliate yourself in public.)\\', \\'performance\\': {\\'dataset\\': \\'Reddit posts\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'DistilBERT is a transformer model that performs sentiment analysis. I fine-tuned the model on Reddit posts with the purpose of classifying not safe for work (NSFW) content, specifically text that is considered inappropriate and unprofessional. The model predicts 2 classes, which are NSFW or safe for work (SFW). The model is a fine-tuned version of DistilBERT. It was fine-tuned on 14317 Reddit posts pulled from the Reddit API.\\'}', metadata={})]", "category": "generic"}
{"question_id": 397, "text": " I want to build a personal assistant to check the logical relationship between two sentences, especially when I have a lot of texts in English.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Paraphrasing\\', \\'api_name\\': \\'prithivida/parrot_paraphraser_on_T5\\', \\'api_call\\': \"Parrot(model_tag=\\'prithivida/parrot_paraphraser_on_T5\\', use_gpu=False)\", \\'api_arguments\\': [\\'input_phrase\\', \\'diversity_ranker\\', \\'do_diverse\\', \\'max_return_phrases\\', \\'max_length\\', \\'adequacy_threshold\\', \\'fluency_threshold\\'], \\'python_environment_requirements\\': [\\'torch\\', \\'transformers\\'], \\'example_code\\': \\'from parrot import Parrot\\\\nimport torch\\\\nimport warnings\\\\nwarnings.filterwarnings(ignore)\\\\n\\\\nparrot = Parrot(model_tag=prithivida/parrot_paraphraser_on_T5, use_gpu=False)\\\\nphrases = [Can you recommed some upscale restaurants in Newyork?,\\\\n What are the famous places we should not miss in Russia?\\\\n]\\\\nfor phrase in phrases:\\\\n print(-*100)\\\\n print(Input_phrase: , phrase)\\\\n print(-*100)\\\\n para_phrases = parrot.augment(input_phrase=phrase)\\\\n for para_phrase in para_phrases:\\\\n  print(para_phrase)\\', \\'performance\\': {\\'dataset\\': \\'Not mentioned\\', \\'accuracy\\': \\'Not mentioned\\'}, \\'description\\': \\'Parrot is a paraphrase based utterance augmentation framework purpose built to accelerate training NLU models. It offers knobs to control Adequacy, Fluency, and Diversity as per your needs. It mainly focuses on augmenting texts typed-into or spoken-to conversational interfaces for building robust NLU models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 398, "text": " I am reading a book and, whenever I engage with an interesting topic, I write a short sentence summarizing that topic. I would like to have an assistant that, given a sentence, automatically classifies that topic among 'technology', 'literature', and 'science'.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'pszemraj/long-t5-tglobal-base-16384-book-summary\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'pszemraj/long-t5-tglobal-base-16384-book-summary\\')\", \\'api_arguments\\': [\\'long_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nimport torch\\\\nsummarizer = pipeline(\\\\n summarization,\\\\n pszemraj/long-t5-tglobal-base-16384-book-summary,\\\\n device=0 if torch.cuda.is_available() else -1,\\\\n)\\\\nlong_text = Here is a lot of text I don\\'t want to read. Replace me\\\\nresult = summarizer(long_text)\\\\nprint(result[0][summary_text])\", \\'performance\\': {\\'dataset\\': \\'kmfoda/booksum\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.408, \\'ROUGE-2\\': 6.065, \\'ROUGE-L\\': 16.721, \\'ROUGE-LSUM\\': 33.34}}, \\'description\\': \\'A fine-tuned version of google/long-t5-tglobal-base on the kmfoda/booksum dataset, which can be used to summarize long text and generate SparkNotes-esque summaries of arbitrary topics. The model generalizes reasonably well to academic and narrative text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 399, "text": " Determine if one Russian sentence logically contradicts the information provided by another Russian sentence.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'Kirili4ik/mbart_ruDialogSum\\', \\'api_call\\': \"MBartForConditionalGeneration.from_pretrained(\\'Kirili4ik/mbart_ruDialogSum\\')\", \\'api_arguments\\': {\\'model_name\\': \\'Kirili4ik/mbart_ruDialogSum\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'from transformers import MBartTokenizer, MBartForConditionalGeneration\\\\nmodel_name = Kirili4ik/mbart_ruDialogSum\\\\ntokenizer = AutoTokenizer.from_pretrained(model_name)\\\\nmodel = MBartForConditionalGeneration.from_pretrained(model_name)\\\\nmodel.eval()\\\\narticle_text = ...\\\\ninput_ids = tokenizer(\\\\n [article_text],\\\\n max_length=600,\\\\n padding=max_length,\\\\n truncation=True,\\\\n return_tensors=pt,\\\\n)[input_ids]\\\\noutput_ids = model.generate(\\\\n input_ids=input_ids,\\\\n top_k=0,\\\\n num_beams=3,\\\\n no_repeat_ngram_size=3\\\\n)[0]\\\\nsummary = tokenizer.decode(output_ids, skip_special_tokens=True)\\\\nprint(summary)\\', \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'SAMSum Corpus (translated to Russian)\\', \\'accuracy\\': {\\'Validation ROGUE-1\\': 34.5, \\'Validation ROGUE-L\\': 33, \\'Test ROGUE-1\\': 31, \\'Test ROGUE-L\\': 28}}]}, \\'description\\': \\'MBart for Russian summarization fine-tuned for dialogues summarization. This model was firstly fine-tuned by Ilya Gusev on Gazeta dataset. We have fine tuned that model on SamSum dataset translated to Russian using GoogleTranslateAPI. Moreover! We have implemented a ! telegram bot @summarization_bot ! with the inference of this model. Add it to the chat and get summaries instead of dozens spam messages!\\'}', metadata={})]", "category": "generic"}
{"question_id": 400, "text": " The assignment is to develop a software for translating sentences from Spanish to English at a near-human performance.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Translation\\', \\'api_name\\': \\'Helsinki-NLP/opus-mt-es-en\\', \\'api_call\\': \"pipeline(\\'translation_es_to_en\\', model=\\'Helsinki-NLP/opus-mt-es-en\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\ntranslation = pipeline(\\'translation_es_to_en\\', model=\\'Helsinki-NLP/opus-mt-es-en\\')(\\'Hola, \u00bfc\u00f3mo est\u00e1s?\\')\", \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'newssyscomb2009-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 30.6, \\'chr-F\\': 0.57}}, {\\'name\\': \\'news-test2008-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 27.9, \\'chr-F\\': 0.553}}, {\\'name\\': \\'newstest2009-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 30.4, \\'chr-F\\': 0.572}}, {\\'name\\': \\'newstest2010-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 36.1, \\'chr-F\\': 0.614}}, {\\'name\\': \\'newstest2011-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 34.2, \\'chr-F\\': 0.599}}, {\\'name\\': \\'newstest2012-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 37.9, \\'chr-F\\': 0.624}}, {\\'name\\': \\'newstest2013-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 35.3, \\'chr-F\\': 0.609}}, {\\'name\\': \\'Tatoeba-test.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 59.6, \\'chr-F\\': 0.739}}]}, \\'description\\': \\'Helsinki-NLP/opus-mt-es-en is a machine translation model trained to translate from Spanish to English using the Hugging Face Transformers library. The model is based on the Marian framework and was trained on the OPUS dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 401, "text": " I am an author. I have a new book that needs a summary. I want the summary to resemble the style of SparkNotes.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'pszemraj/long-t5-tglobal-base-16384-book-summary\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'pszemraj/long-t5-tglobal-base-16384-book-summary\\')\", \\'api_arguments\\': [\\'long_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nimport torch\\\\nsummarizer = pipeline(\\\\n summarization,\\\\n pszemraj/long-t5-tglobal-base-16384-book-summary,\\\\n device=0 if torch.cuda.is_available() else -1,\\\\n)\\\\nlong_text = Here is a lot of text I don\\'t want to read. Replace me\\\\nresult = summarizer(long_text)\\\\nprint(result[0][summary_text])\", \\'performance\\': {\\'dataset\\': \\'kmfoda/booksum\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.408, \\'ROUGE-2\\': 6.065, \\'ROUGE-L\\': 16.721, \\'ROUGE-LSUM\\': 33.34}}, \\'description\\': \\'A fine-tuned version of google/long-t5-tglobal-base on the kmfoda/booksum dataset, which can be used to summarize long text and generate SparkNotes-esque summaries of arbitrary topics. The model generalizes reasonably well to academic and narrative text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 402, "text": " An editor wants to summarize his article in French.\\n###Input: \\\"Selon un rapport r\\u00e9cent, les constructeurs automobiles pr\\u00e9voient d'acc\\u00e9l\\u00e9rer la production de voitures \\u00e9lectriques et de r\\u00e9duire la production de voitures \\u00e0 moteur \\u00e0 combustion interne. Les constructeurs pr\\u00e9voient de consacrer davantage de ressources \\u00e0 la recherche et au d\\u00e9veloppement de technologies pour am\\u00e9liorer l'efficacit\\u00e9 des batteries, r\\u00e9duire les co\\u00fbts de production et \\u00e9tendre le r\\u00e9seau de recharge. Les gouvernements du monde entier continuent de soutenir l'adoption de la mobilit\\u00e9 \\u00e9lectrique \\u00e0 travers des incitations financi\\u00e8res et des objectifs ambitieux de r\\u00e9duction des \\u00e9missions de CO2.\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Abstractive Text Summarization\\', \\'api_name\\': \\'plguillou/t5-base-fr-sum-cnndm\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'plguillou/t5-base-fr-sum-cnndm\\')\", \\'api_arguments\\': {\\'input_text\\': \\'summarize: ARTICLE\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'from transformers import T5Tokenizer, T5ForConditionalGeneration\\'}, \\'example_code\\': \\'tokenizer = T5Tokenizer.from_pretrained(plguillou/t5-base-fr-sum-cnndm)\\\\nmodel = T5ForConditionalGeneration.from_pretrained(plguillou/t5-base-fr-sum-cnndm)\\', \\'performance\\': {\\'dataset\\': \\'cnn_dailymail\\', \\'ROUGE-1\\': 44.5252, \\'ROUGE-2\\': 22.652, \\'ROUGE-L\\': 29.8866}, \\'description\\': \\'This model is a T5 Transformers model (JDBN/t5-base-fr-qg-fquad) that was fine-tuned in French for abstractive text summarization.\\'}', metadata={})]", "category": "generic"}
{"question_id": 403, "text": " Design a chatbot for a school website, which can answer queries related to admissions, classes, teachers, and extracurriculars.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Generation\\', \\'api_name\\': \\'mywateriswet/ShuanBot\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'mywateriswet/ShuanBot\\')\", \\'api_arguments\\': \\'message\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"response = chatbot(\\'What is your name?\\')\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'ShuanBot is a conversational chatbot model based on the GPT-2 architecture. It can be used for generating human-like responses in a chat context.\\'}', metadata={})]", "category": "generic"}
{"question_id": 404, "text": " We have a long article, and we want to generate a summary to understand the main points of the article.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'pszemraj/long-t5-tglobal-base-16384-book-summary\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'pszemraj/long-t5-tglobal-base-16384-book-summary\\')\", \\'api_arguments\\': [\\'long_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nimport torch\\\\nsummarizer = pipeline(\\\\n summarization,\\\\n pszemraj/long-t5-tglobal-base-16384-book-summary,\\\\n device=0 if torch.cuda.is_available() else -1,\\\\n)\\\\nlong_text = Here is a lot of text I don\\'t want to read. Replace me\\\\nresult = summarizer(long_text)\\\\nprint(result[0][summary_text])\", \\'performance\\': {\\'dataset\\': \\'kmfoda/booksum\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.408, \\'ROUGE-2\\': 6.065, \\'ROUGE-L\\': 16.721, \\'ROUGE-LSUM\\': 33.34}}, \\'description\\': \\'A fine-tuned version of google/long-t5-tglobal-base on the kmfoda/booksum dataset, which can be used to summarize long text and generate SparkNotes-esque summaries of arbitrary topics. The model generalizes reasonably well to academic and narrative text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 405, "text": " We need to prepare some sample conversations featuring frequently asked questions for helping customers with our products.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'hyunwoongko/blenderbot-9B\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'hyunwoongko/blenderbot-9B\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'Input a message to start chatting with hyunwoongko/blenderbot-9B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high-performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi-turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 406, "text": " Can you give me advice on how to choose the best video game for me?\\n###Input: {\\\"instruction\\\": \\\"what is the best way to choose a video game?\\\", \\\"knowledge\\\": \\\"Some factors to consider when choosing a video game are personal preferences, genre, graphics, gameplay, storyline, platform, and reviews from other players or gaming websites.\\\", \\\"dialog\\\": [\\\"What type of video games do you prefer playing?\\\", \\\"I enjoy action-adventure games and a decent storyline.\\\"]}\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 407, "text": " Write a script to translate the following French sentence into English: \\\"Je t\\u2019aime.\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Translation, Summarization, Question Answering, Sentiment Analysis\\', \\'api_name\\': \\'t5-3b\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'t5-3b\\')\", \\'api_arguments\\': \\'input_text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"input_text = \\'translate English to French: The quick brown fox jumps over the lazy dog\\'; inputs = tokenizer.encode(input_text, return_tensors=\\'pt\\'); outputs = model.generate(inputs); translated_text = tokenizer.decode(outputs[0])\", \\'performance\\': {\\'dataset\\': \\'c4\\', \\'accuracy\\': \\'See research paper, Table 14\\'}, \\'description\\': \\'T5-3B is a Text-To-Text Transfer Transformer (T5) model with 3 billion parameters. It is designed for various NLP tasks such as translation, summarization, question answering, and sentiment analysis. The model is pre-trained on the Colossal Clean Crawled Corpus (C4) and fine-tuned on multiple supervised and unsupervised tasks.\\'}', metadata={})]", "category": "generic"}
{"question_id": 408, "text": " Your company is developing a chatbot and requires accurate summaries of lengthy dialogues without losing context.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'hyunwoongko/blenderbot-9B\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'hyunwoongko/blenderbot-9B\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'Input a message to start chatting with hyunwoongko/blenderbot-9B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high-performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi-turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 409, "text": " I want a system that can translate an article from English to French automatically.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Translation\\', \\'api_name\\': \\'Helsinki-NLP/opus-mt-en-fr\\', \\'api_call\\': \"translate(\\'input_text\\', model=\\'Helsinki-NLP/opus-mt-en-fr\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'opus\\', \\'accuracy\\': {\\'BLEU\\': {\\'newsdiscussdev2015-enfr.en.fr\\': 33.8, \\'newsdiscusstest2015-enfr.en.fr\\': 40.0, \\'newssyscomb2009.en.fr\\': 29.8, \\'news-test2008.en.fr\\': 27.5, \\'newstest2009.en.fr\\': 29.4, \\'newstest2010.en.fr\\': 32.7, \\'newstest2011.en.fr\\': 34.3, \\'newstest2012.en.fr\\': 31.8, \\'newstest2013.en.fr\\': 33.2, \\'Tatoeba.en.fr\\': 50.5}}}, \\'description\\': \\'Helsinki-NLP/opus-mt-en-fr is a translation model that translates English text to French using the Hugging Face Transformers library. It is based on the OPUS dataset and uses a transformer-align architecture with normalization and SentencePiece pre-processing.\\'}', metadata={})]", "category": "generic"}
{"question_id": 410, "text": " We want our System to generate possible user queries for a document provided as a text input.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Text2Text Generation\\', \\'api_name\\': \\'castorini/doc2query-t5-base-msmarco\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'castorini/doc2query-t5-base-msmarco\\')\", \\'api_arguments\\': \\'text, max_length\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'MS MARCO\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'A T5 model trained on the MS MARCO dataset for generating queries from documents.\\'}', metadata={})]", "category": "generic"}
{"question_id": 411, "text": " Our client wants to create marketing slogans. Help them by completing this slogan \\\"Customer satisfaction is our top <mask>.\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Fill-Mask\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Fill-Mask\\', \\'api_name\\': \\'camembert-base\\', \\'api_call\\': \"pipeline(\\'fill-mask\\', model=\\'camembert-base\\', tokenizer=\\'camembert-base\\')\", \\'api_arguments\\': [\\'model\\', \\'tokenizer\\'], \\'python_environment_requirements\\': [\\'transformers\\', \\'torch\\'], \\'example_code\\': \"from transformers import pipeline; camembert_fill_mask = pipeline(\\'fill-mask\\', model=\\'camembert-base\\', tokenizer=\\'camembert-base\\'); results = camembert_fill_mask(\\'Le camembert est <mask> :)\\')\", \\'performance\\': {\\'dataset\\': \\'oscar\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'CamemBERT is a state-of-the-art language model for French based on the RoBERTa model. It is available on Hugging Face in 6 different versions with varying number of parameters, amount of pretraining data, and pretraining data source domains. It can be used for Fill-Mask tasks.\\'}', metadata={})]", "category": "generic"}
{"question_id": 412, "text": " I need a French language model to help me fill in the blanks in given sentences.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Audio-to-Audio\\', \\'framework\\': \\'Fairseq\\', \\'functionality\\': \\'speech-to-speech-translation\\', \\'api_name\\': \\'facebook/textless_sm_en_fr\\', \\'api_call\\': \"load_model_ensemble_and_task_from_hf_hub(\\'facebook/textless_sm_en_fr\\')\", \\'api_arguments\\': [\\'input_file\\'], \\'python_environment_requirements\\': [\\'huggingface_hub\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'This model is a speech-to-speech translation model trained by Facebook. It is designed for translating English speech to French speech.\\'}', metadata={})]", "category": "generic"}
{"question_id": 413, "text": " We would like to build a semantic text search system that can find similar documents in a repository based on a given description.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Sentence Similarity\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'shibing624/text2vec-base-chinese\\', \\'api_call\\': \"SentenceModel(\\'shibing624/text2vec-base-chinese\\')\", \\'api_arguments\\': [\\'sentences\\'], \\'python_environment_requirements\\': [\\'text2vec\\', \\'transformers\\'], \\'example_code\\': \"from text2vec import SentenceModel\\\\nsentences = [\\'\\', \\'\\']\\\\nmodel = SentenceModel(\\'shibing624/text2vec-base-chinese\\')\\\\nembeddings = model.encode(sentences)\\\\nprint(embeddings)\", \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'ATEC\\', \\'accuracy\\': \\'31.93\\'}, {\\'name\\': \\'BQ\\', \\'accuracy\\': \\'42.67\\'}, {\\'name\\': \\'LCQMC\\', \\'accuracy\\': \\'70.16\\'}, {\\'name\\': \\'PAWSX\\', \\'accuracy\\': \\'17.21\\'}, {\\'name\\': \\'STS-B\\', \\'accuracy\\': \\'79.30\\'}]}, \\'description\\': \\'This is a CoSENT(Cosine Sentence) model: shibing624/text2vec-base-chinese. It maps sentences to a 768 dimensional dense vector space and can be used for tasks like sentence embeddings, text matching or semantic search.\\'}', metadata={})]", "category": "generic"}
{"question_id": 415, "text": " I want to create a playlist of similar songs based on lyrics. How can I compare the sentences and find similar ones?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Sentence Similarity\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Sentence Transformers\\', \\'api_name\\': \\'sentence-transformers/multi-qa-mpnet-base-dot-v1\\', \\'api_call\\': \"SentenceTransformer(\\'sentence-transformers/multi-qa-mpnet-base-dot-v1\\')\", \\'api_arguments\\': [\\'query\\', \\'docs\\'], \\'python_environment_requirements\\': \\'pip install -U sentence-transformers\\', \\'example_code\\': \"from sentence_transformers import SentenceTransformer, util\\\\nquery = How many people live in London?\\\\ndocs = [Around 9 Million people live in London, London is known for its financial district]\\\\nmodel = SentenceTransformer(\\'sentence-transformers/multi-qa-mpnet-base-dot-v1\\')\\\\nquery_emb = model.encode(query)\\\\ndoc_emb = model.encode(docs)\\\\nscores = util.dot_score(query_emb, doc_emb)[0].cpu().tolist()\\\\ndoc_score_pairs = list(zip(docs, scores))\\\\ndoc_score_pairs = sorted(doc_score_pairs, key=lambda x: x[1], reverse=True)\\\\nfor doc, score in doc_score_pairs:\\\\n print(score, doc)\", \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'WikiAnswers\\', \\'accuracy\\': 77427422}, {\\'name\\': \\'PAQ\\', \\'accuracy\\': 64371441}, {\\'name\\': \\'Stack Exchange\\', \\'accuracy\\': 25316456}]}, \\'description\\': \\'This is a sentence-transformers model that maps sentences & paragraphs to a 768 dimensional dense vector space and was designed for semantic search. It has been trained on 215M (question, answer) pairs from diverse sources.\\'}', metadata={})]", "category": "generic"}
{"question_id": 416, "text": " I\\u2019m putting together a dating site where users can submit questions they'd like the matching algorithm to ask. I want to suggest questions like the ones they have already submitted. Can you provide me with a model to do that?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'deepset/roberta-base-squad2-covid\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=RobertaForQuestionAnswering.from_pretrained(\\'deepset/roberta-base-squad2-covid\\'), tokenizer=RobertaTokenizer.from_pretrained(\\'deepset/roberta-base-squad2-covid\\'))\", \\'api_arguments\\': {\\'model_name\\': \\'deepset/roberta-base-squad2-covid\\', \\'tokenizer\\': \\'deepset/roberta-base-squad2-covid\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': {\\'QA_input\\': {\\'question\\': \\'Why is model conversion important?\\', \\'context\\': \\'The option to convert models between FARM and transformers gives freedom to the user and let people easily switch between frameworks.\\'}, \\'res\\': \\'nlp(QA_input)\\'}, \\'performance\\': {\\'dataset\\': \\'squad_v2\\', \\'accuracy\\': {\\'XVAL_EM\\': 0.17890995260663506, \\'XVAL_f1\\': 0.49925444207319924, \\'XVAL_top_3_recall\\': 0.8021327014218009}}, \\'description\\': \\'This model is a Roberta-based model fine-tuned on SQuAD-style CORD-19 annotations for the task of extractive question answering in the context of COVID-19. It can be used with the Hugging Face Transformers library for question answering tasks.\\'}', metadata={})]", "category": "generic"}
{"question_id": 417, "text": " I'm working on a phonebot, and I need the bot to be able to read a sensitive warning message to the users.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Generation\\', \\'api_name\\': \\'mywateriswet/ShuanBot\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'mywateriswet/ShuanBot\\')\", \\'api_arguments\\': \\'message\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"response = chatbot(\\'What is your name?\\')\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'ShuanBot is a conversational chatbot model based on the GPT-2 architecture. It can be used for generating human-like responses in a chat context.\\'}', metadata={})]", "category": "generic"}
{"question_id": 418, "text": " Our customer wants to have this Chinese text \\\"<Chinese_text>\\\" read aloud as an audio file in female voice.\\n###Input: \\\"<Chinese_text>\\\": \\\"\\u4f60\\u597d\\uff0c\\u6b22\\u8fce\\u6765\\u5230\\u6570\\u5b57\\u4e16\\u754c\\u3002\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'kan-bayashi_csmsc_tts_train_tacotron2_raw_phn_pypinyin_g2p_phone_train.loss.best\\', \\'api_call\\': \"Text2Speech.from_pretrained(\\'espnet/kan-bayashi_csmsc_tts_train_tacotron2_raw_phn_pypinyin_g2p_phone_train.loss.best\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'torch\\', \\'espnet_model_zoo\\'], \\'example_code\\': \\'import soundfile\\\\nfrom espnet2.bin.tts_inference import Text2Speech\\\\ntext2speech = Text2Speech.from_pretrained(espnet/kan-bayashi_csmsc_tts_train_tacotron2_raw_phn_pypinyin_g2p_phone_train.loss.best)\\\\ntext = \u6625\u6c5f\u6f6e\u6c34\u8fde\u6d77\u5e73\uff0c\u6d77\u4e0a\u660e\u6708\u5171\u6f6e\u751f\\\\nspeech = text2speech(text)[wav]\\\\nsoundfile.write(out.wav, speech.numpy(), text2speech.fs, PCM_16)\\', \\'performance\\': {\\'dataset\\': \\'csmsc\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'A pre-trained Text-to-Speech model for Chinese language using ESPnet framework. It can be used to convert text input into speech output in Chinese.\\'}', metadata={})]", "category": "generic"}
{"question_id": 419, "text": " Develop a text-to-speech model for our mobile app to read news articles for our users.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'lakahaga/novel_reading_tts\\', \\'api_call\\': \"AutoModelForTTS.from_pretrained(\\'lakahaga/novel_reading_tts\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"inputs = processor(text, return_tensors=\\'pt\\'); generated_audio = model.generate(**inputs);\", \\'performance\\': {\\'dataset\\': \\'novelspeech\\', \\'accuracy\\': None}, \\'description\\': \\'This model was trained by lakahaga using novelspeech recipe in espnet. It is designed for Korean text-to-speech tasks.\\'}', metadata={})]", "category": "generic"}
{"question_id": 420, "text": " You are an Indian in Telugu pandit teaching kids how to pronounce conventional holy prayers.. Teach me mantras in Telugu synthesized by human like voice pronunciation.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'SYSPIN/Telugu_Male_TTS\\', \\'api_call\\': \"pipeline(\\'text-to-speech\\', model=\\'SYSPIN/Telugu_Male_TTS\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Telugu Male Text-to-Speech model using the ESPnet framework, provided by Hugging Face.\\'}', metadata={})]", "category": "generic"}
{"question_id": 421, "text": " We are a French company creating an interactive voice response system for our call centers, and we need to convert text to speech for our telephony prompts.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'Fairseq\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'facebook/tts_transformer-fr-cv7_css10\\', \\'api_call\\': \"load_model_ensemble_and_task_from_hf_hub(\\'facebook/tts_transformer-fr-cv7_css10\\')\", \\'api_arguments\\': {\\'text\\': \\'string\\'}, \\'python_environment_requirements\\': [\\'fairseq\\'], \\'example_code\\': \\'from fairseq.checkpoint_utils import load_model_ensemble_and_task_from_hf_hub\\\\nfrom fairseq.models.text_to_speech.hub_interface import TTSHubInterface\\\\nimport IPython.display as ipd\\\\nmodels, cfg, task = load_model_ensemble_and_task_from_hf_hub(\\\\n facebook/tts_transformer-fr-cv7_css10,\\\\n arg_overrides={vocoder: hifigan, fp16: False}\\\\n)\\\\nmodel = models[0]\\\\nTTSHubInterface.update_cfg_with_data_cfg(cfg, task.data_cfg)\\\\ngenerator = task.build_generator(model, cfg)\\\\ntext = Bonjour, ceci est un test.\\\\nsample = TTSHubInterface.get_model_input(task, text)\\\\nwav, rate = TTSHubInterface.get_prediction(task, model, generator, sample)\\\\nipd.Audio(wav, rate=rate)\\', \\'performance\\': {\\'dataset\\': \\'common_voice\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'Transformer text-to-speech model from fairseq S^2. French, single-speaker male voice. Pre-trained on Common Voice v7, fine-tuned on CSS10.\\'}', metadata={})]", "category": "generic"}
{"question_id": 422, "text": " Create Japanese audio from the following text: \\\"\\u3053\\u3093\\u306b\\u3061\\u306f\\u3001\\u79c1\\u305f\\u3061\\u306f\\u3042\\u306a\\u305f\\u306e\\u52a9\\u3051\\u304c\\u5fc5\\u8981\\u3067\\u3059\\u3002\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'kan-bayashi_jvs_tts_finetune_jvs001_jsut_vits_raw_phn_jaconv_pyopenjta-truncated-178804\\', \\'api_call\\': \"AutoModelForCausalLM.from_pretrained(\\'espnet/kan-bayashi_jvs_tts_finetune_jvs001_jsut_vits_raw_phn_jaconv_pyopenjta-truncated-178804\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Japanese text-to-speech model trained using the ESPnet framework. It is designed to convert text input into natural-sounding speech.\\'}', metadata={})]", "category": "generic"}
{"question_id": 423, "text": " We need to create automated transcripts from recorded podcasts that include punctuation for better readability.\\n \n Use this API documentation for reference: [Document(page_content=\"{'domain': 'Natural Language Processing Token Classification', 'framework': 'Transformers', 'functionality': 'punctuation prediction', 'api_name': 'oliverguhr/fullstop-punctuation-multilang-large', 'api_call': 'PunctuationModel()', 'api_arguments': ['text'], 'python_environment_requirements': ['pip install deepmultilingualpunctuation'], 'example_code': 'from deepmultilingualpunctuation import PunctuationModel\\\\nmodel = PunctuationModel()\\\\ntext = My name is Clara and I live in Berkeley California Ist das eine Frage Frau M\u00fcller\\\\nresult = model.restore_punctuation(text)\\\\nprint(result)', 'performance': {'dataset': 'wmt/europarl', 'EN_accuracy': 0.775, 'DE_accuracy': 0.814, 'FR_accuracy': 0.782, 'IT_accuracy': 0.762}, 'description': 'This model predicts the punctuation of English, Italian, French and German texts. It was developed to restore the punctuation of transcribed spoken language and trained on the Europarl Dataset provided by the SEPP-NLG Shared Task. The model restores the following punctuation markers: ., ,, ?, -, :.'}\", metadata={})]", "category": "generic"}
{"question_id": 424, "text": " Convert the audio file of a phone interview to text for further analysis.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Audio-to-Audio\\', \\'framework\\': \\'Fairseq\\', \\'functionality\\': \\'speech-to-speech-translation\\', \\'api_name\\': \\'facebook/textless_sm_ro_en\\', \\'api_call\\': \"pipeline(\\'audio-to-audio\\', model=\\'facebook/textless_sm_ro_en\\')\", \\'api_arguments\\': \\'audio file or recording\\', \\'python_environment_requirements\\': \\'fairseq, huggingface_hub\\', \\'example_code\\': \\'https://huggingface.co/facebook/textless_sm_cs_en\\', \\'performance\\': {\\'dataset\\': \\'unknown\\', \\'accuracy\\': \\'unknown\\'}, \\'description\\': \\'A speech-to-speech translation model for Romanian to English developed by Facebook AI\\'}', metadata={})]", "category": "generic"}
{"question_id": 426, "text": " I want to use speech enhancement for an audio call of a coworker.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'SYSPIN/Telugu_Male_TTS\\', \\'api_call\\': \"pipeline(\\'text-to-speech\\', model=\\'SYSPIN/Telugu_Male_TTS\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Telugu Male Text-to-Speech model using the ESPnet framework, provided by Hugging Face.\\'}', metadata={})]", "category": "generic"}
{"question_id": 427, "text": " How do I separate the speakers from an audio file using the pre-trained ConvTasNet_Libri2Mix_sepclean_8k model from Hugging Face?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Audio-to-Audio\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Asteroid\\', \\'api_name\\': \\'JorisCos/ConvTasNet_Libri2Mix_sepnoisy_16k\\', \\'api_call\\': \"BaseModel.from_pretrained(\\'JorisCos/ConvTasNet_Libri2Mix_sepnoisy_16k\\')\", \\'api_arguments\\': \\'audio\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'Libri2Mix\\', \\'accuracy\\': {\\'si_sdr\\': 10.617130949793383, \\'si_sdr_imp\\': 12.551811412989263, \\'sdr\\': 11.231867464482065, \\'sdr_imp\\': 13.059765009747343, \\'sir\\': 24.461138352988346, \\'sir_imp\\': 24.371856452307703, \\'sar\\': 11.5649982725426, \\'sar_imp\\': 4.662525705768228, \\'stoi\\': 0.8701085138712695, \\'stoi_imp\\': 0.2245418019822898}}, \\'description\\': \\'This model was trained by Joris Cosentino using the librimix recipe in Asteroid. It was trained on the sep_noisy task of the Libri2Mix dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 428, "text": " A travel app we work on can translate language of a guide in real-time. We are targeting Spanish-speaking tourists.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'CTranslate2\\', \\'functionality\\': \\'Automatic Speech Recognition\\', \\'api_name\\': \\'guillaumekln/faster-whisper-large-v2\\', \\'api_call\\': \"WhisperModel(\\'large-v2\\')\", \\'api_arguments\\': [\\'audio.mp3\\'], \\'python_environment_requirements\\': [\\'faster_whisper\\'], \\'example_code\\': \\'from faster_whisper import WhisperModel\\\\nmodel = WhisperModel(large-v2)\\\\nsegments, info = model.transcribe(audio.mp3)\\\\nfor segment in segments:\\\\n print([%.2fs -&gt; %.2fs] %s % (segment.start, segment.end, segment.text))\\', \\'performance\\': {\\'dataset\\': \\'99 languages\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Whisper large-v2 model for CTranslate2. This model can be used in CTranslate2 or projets based on CTranslate2 such as faster-whisper.\\'}', metadata={})]", "category": "generic"}
{"question_id": 429, "text": " Design an audio description system for a user who needs help translating spoken English audio to spoken Hokkien audio.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'Fairseq\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'unit_hifigan_HK_layer12.km2500_frame_TAT-TTS\\', \\'api_call\\': \"load_model_ensemble_and_task_from_hf_hub(\\'facebook/unit_hifigan_HK_layer12.km2500_frame_TAT-TT\\')\", \\'api_arguments\\': {\\'unit\\': \\'Text input for the TTS model\\'}, \\'python_environment_requirements\\': [\\'fairseq\\', \\'huggingface_hub\\', \\'torchaudio\\'], \\'example_code\\': \"import json\\\\nimport os\\\\nfrom pathlib import Path\\\\nimport IPython.display as ipd\\\\nfrom fairseq import hub_utils\\\\nfrom fairseq.checkpoint_utils import load_model_ensemble_and_task_from_hf_hub\\\\nfrom fairseq.models.speech_to_text.hub_interface import S2THubInterface\\\\nfrom fairseq.models.text_to_speech import CodeHiFiGANVocoder\\\\nfrom fairseq.models.text_to_speech.hub_interface import VocoderHubInterface\\\\nfrom huggingface_hub import snapshot_download\\\\nimport torchaudio\\\\ncache_dir = os.getenv(HUGGINGFACE_HUB_CACHE)\\\\nlibrary_name = fairseq\\\\ncache_dir = (\\\\n cache_dir or (Path.home() / .cache / library_name).as_posix()\\\\n)\\\\ncache_dir = snapshot_download(\\\\n ffacebook/unit_hifigan_HK_layer12.km2500_frame_TAT-TTS, cache_dir=cache_dir, library_name=library_name\\\\n)\\\\nx = hub_utils.from_pretrained(\\\\n cache_dir,\\\\n model.pt,\\\\n .,\\\\n archive_map=CodeHiFiGANVocoder.hub_models(),\\\\n config_yaml=config.json,\\\\n fp16=False,\\\\n is_vocoder=True,\\\\n)\\\\nwith open(f{x[\\'args\\'][\\'data\\']}/config.json) as f:\\\\n vocoder_cfg = json.load(f)\\\\nassert (\\\\n len(x[args][model_path]) == 1\\\\n), Too many vocoder models in the input\\\\nvocoder = CodeHiFiGANVocoder(x[args][model_path][0], vocoder_cfg)\\\\ntts_model = VocoderHubInterface(vocoder_cfg, vocoder)\\\\ntts_sample = tts_model.get_model_input(unit)\\\\nwav, sr = tts_model.get_prediction(tts_sample)\\\\nipd.Audio(wav, rate=sr)\", \\'performance\\': {\\'dataset\\': \\'TAT-TTS\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Hokkien unit HiFiGAN based vocoder from fairseq. Trained with TAT-TTS data with 4 speakers in Taiwanese Hokkien accent.\\'}', metadata={})]", "category": "generic"}
{"question_id": 430, "text": " Implement a method to perform speech-to-speech translation between Hokkien and English using the xm_transformer_s2ut_hk-en model.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Audio-to-Audio\\', \\'framework\\': \\'Fairseq\\', \\'functionality\\': \\'speech-to-speech-translation\\', \\'api_name\\': \\'xm_transformer_s2ut_en-hk\\', \\'api_call\\': \"load_model_ensemble_and_task_from_hf_hub(\\'facebook/xm_transformer_s2ut_en-hk\\')\", \\'api_arguments\\': {\\'arg_overrides\\': {\\'config_yaml\\': \\'config.yaml\\', \\'task\\': \\'speech_to_text\\'}, \\'cache_dir\\': \\'cache_dir\\'}, \\'python_environment_requirements\\': [\\'fairseq\\', \\'huggingface_hub\\', \\'torchaudio\\'], \\'example_code\\': {\\'import_modules\\': [\\'import json\\', \\'import os\\', \\'from pathlib import Path\\', \\'import IPython.display as ipd\\', \\'from fairseq import hub_utils\\', \\'from fairseq.checkpoint_utils import load_model_ensemble_and_task_from_hf_hub\\', \\'from fairseq.models.speech_to_text.hub_interface import S2THubInterface\\', \\'from fairseq.models.text_to_speech import CodeHiFiGANVocoder\\', \\'from fairseq.models.text_to_speech.hub_interface import VocoderHubInterface\\', \\'from huggingface_hub import snapshot_download\\', \\'import torchaudio\\'], \\'load_model\\': [\"cache_dir = os.getenv(\\'HUGGINGFACE_HUB_CACHE\\')\", \"models, cfg, task = load_model_ensemble_and_task_from_hf_hub(\\'facebook/xm_transformer_s2ut_en-hk\\', arg_overrides={\\'config_yaml\\': \\'config.yaml\\', \\'task\\': \\'speech_to_text\\'}, cache_dir=cache_dir)\", \\'model = models[0].cpu()\\', \"cfg[\\'task\\'].cpu = True\"], \\'generate_prediction\\': [\\'generator = task.build_generator([model], cfg)\\', \"audio, _ = torchaudio.load(\\'/path/to/an/audio/file\\')\", \\'sample = S2THubInterface.get_model_input(task, audio)\\', \\'unit = S2THubInterface.get_prediction(task, model, generator, sample)\\'], \\'speech_synthesis\\': [\"library_name = \\'fairseq\\'\", \"cache_dir = (cache_dir or (Path.home() / \\'.cache\\' / library_name).as_posix())\", \"cache_dir = snapshot_download(\\'facebook/unit_hifigan_HK_layer12.km2500_frame_TAT-TTS\\', cache_dir=cache_dir, library_name=library_name)\", \"x = hub_utils.from_pretrained(cache_dir, \\'model.pt\\', \\'.\\', archive_map=CodeHiFiGANVocoder.hub_models(), config_yaml=\\'config.json\\', fp16=False, is_vocoder=True)\", \"with open(f\\'{x[\\'args\\'][\\'data\\']}/config.json\\') as f:\", \\'  vocoder_cfg = json.load(f)\\', \"assert (len(x[\\'args\\'][\\'model_path\\']) == 1), \\'Too many vocoder models in the input\\'\", \"vocoder = CodeHiFiGANVocoder(x[\\'args\\'][\\'model_path\\'][0], vocoder_cfg)\", \\'tts_model = VocoderHubInterface(vocoder_cfg, vocoder)\\', \\'tts_sample = tts_model.get_model_input(unit)\\', \\'wav, sr = tts_model.get_prediction(tts_sample)\\', \\'ipd.Audio(wav, rate=sr)\\']}, \\'performance\\': {\\'dataset\\': \\'MuST-C\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'Speech-to-speech translation model with single-pass decoder (S2UT) from fairseq: English-Hokkien. Trained with supervised data in TED domain, and weakly supervised data in TED and Audiobook domain.\\'}', metadata={})]", "category": "generic"}
{"question_id": 431, "text": " A company wants to analyze the sentiment of the customer feedback in their Spanish-speaking call center. Generate a script for this task.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'finiteautomata/beto-sentiment-analysis\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'finiteautomata/beto-sentiment-analysis\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'Hugging Face Transformers library\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'TASS 2020 corpus\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Model trained with TASS 2020 corpus (around ~5k tweets) of several dialects of Spanish. Base model is BETO, a BERT model trained in Spanish. Uses POS, NEG, NEU labels.\\'}', metadata={})]", "category": "generic"}
{"question_id": 432, "text": " Our company is working on an AI-powered language learning app for German. We need to build a feature that classifies emotions in German speech.\\n \n Use this API documentation for reference: [Document(page_content=\"{'domain': 'Natural Language Processing Text Classification', 'framework': 'Hugging Face Transformers', 'functionality': 'German Sentiment Classification', 'api_name': 'oliverguhr/german-sentiment-bert', 'api_call': 'SentimentModel()', 'api_arguments': ['texts'], 'python_environment_requirements': 'pip install germansentiment', 'example_code': ['from germansentiment import SentimentModel', 'model = SentimentModel()', 'texts = [', ' Mit keinem guten Ergebniss,Das ist gar nicht mal so gut,', ' Total awesome!,nicht so schlecht wie erwartet,', ' Der Test verlief positiv.,Sie fhrt ein grnes Auto.]', 'result = model.predict_sentiment(texts)', 'print(result)'], 'performance': {'dataset': ['holidaycheck', 'scare', 'filmstarts', 'germeval', 'PotTS', 'emotions', 'sb10k', 'Leipzig Wikipedia Corpus 2016', 'all'], 'accuracy': [0.9568, 0.9418, 0.9021, 0.7536, 0.678, 0.9649, 0.7376, 0.9967, 0.9639]}, 'description': 'This model was trained for sentiment classification of German language texts. The model uses the Googles Bert architecture and was trained on 1.834 million German-language samples. The training data contains texts from various domains like Twitter, Facebook and movie, app and hotel reviews.'}\", metadata={})]", "category": "generic"}
{"question_id": 433, "text": " We are a voice assistant service, and we need to verify the speaker identity for enhanced security when users access the system.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Voice Activity Detection\\', \\'api_name\\': \\'d4data/Indian-voice-cloning\\', \\'api_call\\': \"pipeline(\\'voice-activity-detection\\', model=\\'d4data/Indian-voice-cloning\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A model for detecting voice activity in Indian languages.\\'}', metadata={})]", "category": "generic"}
{"question_id": 434, "text": " I run a call center and I need a system that can identify the person on the other end of the line by analyzing their voice.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Voice Activity Detection\\', \\'api_name\\': \\'d4data/Indian-voice-cloning\\', \\'api_call\\': \"pipeline(\\'voice-activity-detection\\', model=\\'d4data/Indian-voice-cloning\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A model for detecting voice activity in Indian languages.\\'}', metadata={})]", "category": "generic"}
{"question_id": 435, "text": " A transcription service wants to identify which speaker said which words in an audio file. Help them with this task.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'CTranslate2\\', \\'functionality\\': \\'Automatic Speech Recognition\\', \\'api_name\\': \\'guillaumekln/faster-whisper-large-v2\\', \\'api_call\\': \"WhisperModel(\\'large-v2\\')\", \\'api_arguments\\': [\\'audio.mp3\\'], \\'python_environment_requirements\\': [\\'faster_whisper\\'], \\'example_code\\': \\'from faster_whisper import WhisperModel\\\\nmodel = WhisperModel(large-v2)\\\\nsegments, info = model.transcribe(audio.mp3)\\\\nfor segment in segments:\\\\n print([%.2fs -&gt; %.2fs] %s % (segment.start, segment.end, segment.text))\\', \\'performance\\': {\\'dataset\\': \\'99 languages\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Whisper large-v2 model for CTranslate2. This model can be used in CTranslate2 or projets based on CTranslate2 such as faster-whisper.\\'}', metadata={})]", "category": "generic"}
{"question_id": 436, "text": " I am running a chain of wine stores and I want to categorize wines and recommend them based on their quality.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Scikit-learn\\', \\'functionality\\': \\'Wine Quality classification\\', \\'api_name\\': \\'julien-c/wine-quality\\', \\'api_call\\': \"joblib.load(cached_download(hf_hub_url(\\'julien-c/wine-quality\\', \\'winequality-red.csv\\')))\", \\'api_arguments\\': [\\'X\\'], \\'python_environment_requirements\\': [\\'huggingface_hub\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \\'from huggingface_hub import hf_hub_url, cached_download\\\\nimport joblib\\\\nimport pandas as pd\\\\nREPO_ID = julien-c/wine-quality\\\\nFILENAME = sklearn_model.joblib\\\\nmodel = joblib.load(cached_download(\\\\n hf_hub_url(REPO_ID, FILENAME)\\\\n))\\\\ndata_file = cached_download(\\\\n hf_hub_url(REPO_ID, winequality-red.csv)\\\\n)\\\\nwinedf = pd.read_csv(data_file, sep=;)\\\\nX = winedf.drop([quality], axis=1)\\\\nY = winedf[quality]\\\\nprint(X[:3])\\\\nlabels = model.predict(X[:3])\\\\nmodel.score(X, Y)\\', \\'performance\\': {\\'dataset\\': \\'julien-c/wine-quality\\', \\'accuracy\\': 0.6616635397123202}, \\'description\\': \\'A Simple Example of Scikit-learn Pipeline for Wine Quality classification. Inspired by https://towardsdatascience.com/a-simple-example-of-pipeline-in-machine-learning-with-scikit-learn-e726ffbb6976 by Saptashwa Bhattacharyya.\\'}', metadata={})]", "category": "generic"}
{"question_id": 437, "text": " I own an e-commerce platform and I need to predict which customers will make a purchase based on their browsing behavior.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Keras\\', \\'functionality\\': \\'Binary Classification\\', \\'api_name\\': \\'TF_Decision_Trees\\', \\'api_call\\': \\'TF_Decision_Trees(input_features, target)\\', \\'api_arguments\\': [\\'input_features\\', \\'target\\'], \\'python_environment_requirements\\': [\\'tensorflow >= 7.0\\'], \\'example_code\\': \\'https://github.com/tdubon/TF-GB-Forest/blob/c0cf4c7e3e29d819b996cfe4eecc1f2728115e52/TFDecisionTrees_Final.ipynb\\', \\'performance\\': {\\'dataset\\': \\'Census-Income Data Set\\', \\'accuracy\\': 96.57}, \\'description\\': \"Use TensorFlow\\'s Gradient Boosted Trees model in binary classification of structured data. Build a decision forests model by specifying the input feature usage. Implement a custom Binary Target encoder as a Keras Preprocessing layer to encode the categorical features with respect to their target value co-occurrences, and then use the encoded features to build a decision forests model. The model is trained on the US Census Income Dataset containing approximately 300k instances with 41 numerical and categorical variables. The task is to determine whether a person makes over 50k a year.\"}', metadata={})]", "category": "generic"}
{"question_id": 438, "text": " Predict the income category of a person, based on their demographic information. The model should use Tensorflow decision trees to make binary classifications.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Keras\\', \\'functionality\\': \\'Binary Classification\\', \\'api_name\\': \\'TF_Decision_Trees\\', \\'api_call\\': \\'TF_Decision_Trees(input_features, target)\\', \\'api_arguments\\': [\\'input_features\\', \\'target\\'], \\'python_environment_requirements\\': [\\'tensorflow >= 7.0\\'], \\'example_code\\': \\'https://github.com/tdubon/TF-GB-Forest/blob/c0cf4c7e3e29d819b996cfe4eecc1f2728115e52/TFDecisionTrees_Final.ipynb\\', \\'performance\\': {\\'dataset\\': \\'Census-Income Data Set\\', \\'accuracy\\': 96.57}, \\'description\\': \"Use TensorFlow\\'s Gradient Boosted Trees model in binary classification of structured data. Build a decision forests model by specifying the input feature usage. Implement a custom Binary Target encoder as a Keras Preprocessing layer to encode the categorical features with respect to their target value co-occurrences, and then use the encoded features to build a decision forests model. The model is trained on the US Census Income Dataset containing approximately 300k instances with 41 numerical and categorical variables. The task is to determine whether a person makes over 50k a year.\"}', metadata={})]", "category": "generic"}
{"question_id": 439, "text": " A travel agency needs a system to predict whether a client's vacation will be successful based on their chosen destination, accommodation, and travel style. We want to provide suggestions for clients who may need additional support.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'prompthero/openjourney-v4\\', \\'api_call\\': \"pipeline(\\'text-to-image\\', model=\\'prompthero/openjourney-v4\\')\", \\'api_arguments\\': {\\'text\\': \\'string\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"generate_image(\\'your text here\\')\", \\'performance\\': {\\'dataset\\': \\'Midjourney v4 images\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Openjourney v4 is trained on +124k Midjourney v4 images by PromptHero. It is used for generating images based on text inputs.\\'}', metadata={})]", "category": "generic"}
{"question_id": 440, "text": " A new city planning company wants to estimate carbon emissions for different types of buildings to provide sustainable living solutions. We need a model to classify the carbon emissions.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Joblib\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'tejas23/autotrain-amx2-1702259725\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': {\\'accuracy\\': 0.827}}, \\'description\\': \\'Multi-class Classification Model for Carbon Emissions\\'}', metadata={})]", "category": "generic"}
{"question_id": 441, "text": " We want to evaluate the carbon footprint of a construction project based on the tabular data of material consumption.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'45473113800\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'samvelkoch/autotrain-data-prknsn-2\\', \\'accuracy\\': {\\'Loss\\': 5.079, \\'R2\\': 0.109, \\'MSE\\': 25.795, \\'MAE\\': 3.78, \\'RMSLE\\': 0.849}}, \\'description\\': \\'A tabular regression model trained with AutoTrain for predicting carbon emissions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 442, "text": " Our company is an environmental consultancy firm. Determine the carbon emissions of different facilities based on the provided data.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Joblib\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'jwan2021/autotrain-jwan-autotrain1-1768961489\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'json\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'Validation Metrics\\', \\'accuracy\\': 0.983}, \\'description\\': \\'Binary Classification model for Carbon Emissions prediction\\'}', metadata={})]", "category": "generic"}
{"question_id": 443, "text": " I need assistance in predicting carbon emissions of a city based on historical data. Use the dataset provided to predict future carbon emissions.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'kochetkovIT/autotrain-ironhack-49741119788\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'json\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'kochetkovIT/autotrain-data-ironhack\\', \\'accuracy\\': {\\'Loss\\': 2.603, \\'R2\\': 0.013, \\'MSE\\': 6.776, \\'MAE\\': 1.666, \\'RMSLE\\': 0.502}}, \\'description\\': \\'A tabular regression model trained using AutoTrain to predict carbon emissions (in grams) with an R2 score of 0.013.\\'}', metadata={})]", "category": "generic"}
{"question_id": 444, "text": " We have some clients asking for estimating CO2 emissions based on their historic data, which is in a CSV file.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Joblib\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'pachi107/autotrain-in-class-test-1780161764\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data.csv\\'], \\'python_environment_requirements\\': [\\'json\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(data.csv)\\\\ndata = data[features]\\\\ndata.columns = [feat_ + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'Validation Metrics\\', \\'accuracy\\': 0.974}, \\'description\\': \\'A binary classification model for predicting CO2 emissions based on tabular data. Trained using AutoTrain with a model ID of 1780161764.\\'}', metadata={})]", "category": "generic"}
{"question_id": 445, "text": " Determine the electricity consumption of a residential area based on historical data.\\n \n Use this API documentation for reference: [Document(page_content=\"{'domain': 'Tabular Tabular Regression', 'framework': 'Scikit-learn', 'functionality': 'skops', 'api_name': 'rajistics/MAPIE-TS-Electricity', 'api_call': 'RandomForestRegressor(max_depth=10, n_estimators=50, random_state=59)', 'api_arguments': {'bootstrap': 'True', 'ccp_alpha': '0.0', 'criterion': 'squared_error', 'max_depth': '10', 'max_features': '1.0', 'max_leaf_nodes': '', 'max_samples': '', 'min_impurity_decrease': '0.0', 'min_samples_leaf': '1', 'min_samples_split': '2', 'min_weight_fraction_leaf': '0.0', 'n_estimators': '50', 'n_jobs': '', 'oob_score': 'False', 'random_state': '59', 'verbose': '0', 'warm_start': 'False'}, 'python_environment_requirements': '', 'example_code': '', 'performance': {'dataset': '', 'accuracy': ''}, 'description': 'A RandomForestRegressor model for electricity consumption prediction.'}\", metadata={})]", "category": "generic"}
{"question_id": 446, "text": " A soccer simulation company wants to use a reinforcement learning agent that can play SoccerTwos effectively.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Reinforcement Learning\\', \\'framework\\': \\'Unity ML-Agents Library\\', \\'functionality\\': \\'Train and play SoccerTwos\\', \\'api_name\\': \\'poca-SoccerTwosv2\\', \\'api_call\\': \"mlagents-load-from-hf --repo-id=\\'Raiden-1001/poca-SoccerTwosv2\\' --local-dir=\\'./downloads\\'\", \\'api_arguments\\': [\\'your_configuration_file_path.yaml\\', \\'run_id\\'], \\'python_environment_requirements\\': [\\'ml-agents\\'], \\'example_code\\': \\'mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume\\', \\'performance\\': {\\'dataset\\': \\'SoccerTwos\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'A trained model of a poca agent playing SoccerTwos using the Unity ML-Agents Library.\\'}', metadata={})]", "category": "generic"}
{"question_id": 447, "text": " Our company is developing a game, and they want to include an AI agent to play the game. We need to evaluate its performance on the CartPole-v1 environment for consistency.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Reinforcement Learning\\', \\'framework\\': \\'Stable-Baselines3\\', \\'functionality\\': \\'CartPole-v1\\', \\'api_name\\': \\'dqn-CartPole-v1\\', \\'api_call\\': \"load_from_hub(repo_id=\\'sb3/dqn-CartPole-v1\\',filename=\\'{MODEL FILENAME}.zip\\',)\", \\'api_arguments\\': [\\'algo\\', \\'env\\', \\'logs\\'], \\'python_environment_requirements\\': [\\'rl_zoo3\\', \\'stable-baselines3\\', \\'stable-baselines3-contrib\\'], \\'example_code\\': \\'python train.py --algo dqn --env CartPole-v1 -f logs/\\', \\'performance\\': {\\'dataset\\': \\'CartPole-v1\\', \\'accuracy\\': \\'500.00 +/- 0.00\\'}, \\'description\\': \\'This is a trained model of a DQN agent playing CartPole-v1 using the stable-baselines3 library and the RL Zoo. The RL Zoo is a training framework for Stable Baselines3 reinforcement learning agents, with hyperparameter optimization and pre-trained agents included.\\'}', metadata={})]", "category": "generic"}
{"question_id": 448, "text": " We need help from a AI to continuously develop a virtual environment for our production robots.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'hyunwoongko/blenderbot-9B\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'hyunwoongko/blenderbot-9B\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'Input a message to start chatting with hyunwoongko/blenderbot-9B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high-performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi-turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 449, "text": " We are a news agency that wants to extract useful features from Korean news articles for a content recommendation service.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'text2text-generation\\', \\'api_name\\': \\'google/pegasus-newsroom\\', \\'api_call\\': \"pipeline(\\'summarization\\', model=\\'google/pegasus-newsroom\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'This model can be loaded on the Inference API on-demand.\\', \\'performance\\': {\\'dataset\\': \\'newsroom\\', \\'accuracy\\': \\'45.98/34.20/42.18\\'}, \\'description\\': \\'PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization by Jingqing Zhang, Yao Zhao, Mohammad Saleh, and Peter J. Liu. The model is trained on both C4 and HugeNews datasets and is designed for summarization tasks.\\'}', metadata={})]", "category": "generic"}
{"question_id": 450, "text": " We want to automatically generate hashtags for the provided image URL to improve social media post performance.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Image-to-Text\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Image Captioning\\', \\'api_name\\': \\'blip-image-captioning-large\\', \\'api_call\\': \\'BlipForConditionalGeneration.from_pretrained(Salesforce/blip-image-captioning-large)\\', \\'api_arguments\\': {\\'raw_image\\': \\'Image\\', \\'text\\': \\'Optional Text\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'BlipProcessor, BlipForConditionalGeneration\\', \\'PIL\\': \\'Image\\', \\'requests\\': \\'requests\\'}, \\'example_code\\': {\\'import_requests\\': \\'import requests\\', \\'import_PIL\\': \\'from PIL import Image\\', \\'import_transformers\\': \\'from transformers import BlipProcessor, BlipForConditionalGeneration\\', \\'load_processor\\': \\'processor = BlipProcessor.from_pretrained(Salesforce/blip-image-captioning-large)\\', \\'load_model\\': \\'model = BlipForConditionalGeneration.from_pretrained(Salesforce/blip-image-captioning-large)\\', \\'load_image\\': \"img_url = \\'https://storage.googleapis.com/sfr-vision-language-research/BLIP/demo.jpg\\'\\\\nraw_image = Image.open(requests.get(img_url, stream=True).raw).convert(\\'RGB\\')\", \\'conditional_captioning\\': \\'text = a photography of\\\\ninputs = processor(raw_image, text, return_tensors=pt)\\\\nout = model.generate(**inputs)\\\\nprint(processor.decode(out[0], skip_special_tokens=True))\\', \\'unconditional_captioning\\': \\'inputs = processor(raw_image, return_tensors=pt)\\\\nout = model.generate(**inputs)\\\\nprint(processor.decode(out[0], skip_special_tokens=True))\\'}, \\'performance\\': {\\'dataset\\': \\'COCO\\', \\'accuracy\\': {\\'image-text retrieval\\': \\'+2.7% recall@1\\', \\'image captioning\\': \\'+2.8% CIDEr\\', \\'VQA\\': \\'+1.6% VQA score\\'}}, \\'description\\': \\'BLIP is a Vision-Language Pre-training (VLP) framework that achieves state-of-the-art results on a wide range of vision-language tasks, such as image-text retrieval, image captioning, and VQA. It effectively utilizes noisy web data by bootstrapping the captions, where a captioner generates synthetic captions and a filter removes the noisy ones.\\'}', metadata={})]", "category": "generic"}
{"question_id": 451, "text": " I am building a recommendation engine for TV shows. Can you evaluate the TV shows using a BERT-based model trained on sentence embedding to find the most similar TV shows based on description?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Classification\\', \\'api_name\\': \\'lvwerra/distilbert-imdb\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'lvwerra/distilbert-imdb\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\', \\'pytorch\\'], \\'example_code\\': \"classifier(\\'I love this movie!\\')\", \\'performance\\': {\\'dataset\\': \\'imdb\\', \\'accuracy\\': 0.928}, \\'description\\': \\'This model is a fine-tuned version of distilbert-base-uncased on the imdb dataset. It is used for sentiment analysis on movie reviews and achieves an accuracy of 0.928 on the evaluation set.\\'}', metadata={})]", "category": "generic"}
{"question_id": 452, "text": " I want to create a solution that can answer questions related to an image of my pet dogs.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'julien-c/hotdog-not-hotdog\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'julien-c/hotdog-not-hotdog\\')\", \\'api_arguments\\': \\'image\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': 0.825}, \\'description\\': \\'A model that classifies images as hotdog or not hotdog.\\'}', metadata={})]", "category": "generic"}
{"question_id": 453, "text": " Design an explanation on how to use Pix2Struct to analyze and generate text based on visuals such as graphs and charts.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Visual Question Answering\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'google/pix2struct-chartqa-base\\', \\'api_call\\': \"Pix2StructForConditionalGeneration.from_pretrained(\\'google/pix2struct-chartqa-base\\')\", \\'api_arguments\\': [\\'t5x_checkpoint_path\\', \\'pytorch_dump_path\\', \\'use-large\\'], \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'python convert_pix2struct_checkpoint_to_pytorch.py --t5x_checkpoint_path PATH_TO_T5X_CHECKPOINTS --pytorch_dump_path PATH_TO_SAVE\\', \\'performance\\': {\\'dataset\\': \\'ChartQA\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Pix2Struct is an image encoder - text decoder model that is trained on image-text pairs for various tasks, including image captionning and visual question answering. The model is pretrained by learning to parse masked screenshots of web pages into simplified HTML. It can achieve state-of-the-art results in six out of nine tasks across four domains: documents, illustrations, user interfaces, and natural images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 454, "text": " Design a product that allows users to read street signs in a foreign language.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'PyTorch Transformers\\', \\'functionality\\': \\'text2text-generation\\', \\'api_name\\': \\'facebook/nllb-200-distilled-600M\\', \\'api_call\\': \"pipeline(\\'translation_xx_to_yy\\', model=\\'facebook/nllb-200-distilled-600M\\')\", \\'api_arguments\\': [\\'model\\', \\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\', \\'torch\\'], \\'example_code\\': \"from transformers import pipeline; translator = pipeline(\\'translation_xx_to_yy\\', model=\\'facebook/nllb-200-distilled-600M\\'); translator(\\'Hello World\\')\", \\'performance\\': {\\'dataset\\': \\'Flores-200\\', \\'accuracy\\': \\'BLEU, spBLEU, chrF++\\'}, \\'description\\': \\'NLLB-200 is a machine translation model primarily intended for research in machine translation, especially for low-resource languages. It allows for single sentence translation among 200 languages. The model was trained on general domain text data and is not intended to be used with domain specific texts, such as medical domain or legal domain. The model is not intended to be used for document translation.\\'}', metadata={})]", "category": "generic"}
{"question_id": 455, "text": " I have a text question about an image, and I would like to receive an appropriate answer.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'prompthero/openjourney-v4\\', \\'api_call\\': \"pipeline(\\'text-to-image\\', model=\\'prompthero/openjourney-v4\\')\", \\'api_arguments\\': {\\'text\\': \\'string\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"generate_image(\\'your text here\\')\", \\'performance\\': {\\'dataset\\': \\'Midjourney v4 images\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Openjourney v4 is trained on +124k Midjourney v4 images by PromptHero. It is used for generating images based on text inputs.\\'}', metadata={})]", "category": "generic"}
{"question_id": 456, "text": " Create software that extracts answers from input documents when given a set of questions.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'distilbert-base-uncased-distilled-squad\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=\\'distilbert-base-uncased-distilled-squad\\')\", \\'api_arguments\\': [\\'question\\', \\'context\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nquestion_answerer = pipeline(question-answering, model=\\'distilbert-base-uncased-distilled-squad\\')\\\\ncontext = r\\\\n... Extractive Question Answering is the task of extracting an answer from a text given a question. An example of a\\\\n... question answering dataset is the SQuAD dataset, which is entirely based on that task. If you would like to fine-tune\\\\n... a model on a SQuAD task, you may leverage the examples/pytorch/question-answering/run_squad.py script.\\\\n... \\\\nresult = question_answerer(question=What is a good example of a question answering dataset?, context=context)\\\\nprint(\\\\n... fAnswer: \\'{result[\\'answer\\']}\\', score: {round(result[\\'score\\'], 4)}, start: {result[\\'start\\']}, end: {result[\\'end\\']}\\\\n...)\", \\'performance\\': {\\'dataset\\': \\'SQuAD v1.1\\', \\'accuracy\\': \\'86.9 F1 score\\'}, \\'description\\': \"DistilBERT base uncased distilled SQuAD is a fine-tuned version of DistilBERT-base-uncased for the task of question answering. It has 40% less parameters than bert-base-uncased, runs 60% faster while preserving over 95% of BERT\\'s performances as measured on the GLUE language understanding benchmark.\"}', metadata={})]", "category": "generic"}
{"question_id": 457, "text": " We received a scanned document with a lot of information. We need to go through it and find answers to specific questions. Create a program to extract information from the document.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 458, "text": " In a pharmaceutical company, we are working on developing a new drug, and I want you to help classify the molecular structures of my dataset.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'sentiment_analysis_generic_dataset\\', \\'api_call\\': \"pipeline(\\'text-classification\\', model=\\'Seethal/sentiment_analysis_generic_dataset\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"sentiment_analysis(\\'I love this product!\\')\", \\'performance\\': {\\'dataset\\': \\'generic_dataset\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'This is a fine-tuned downstream version of the bert-base-uncased model for sentiment analysis, this model is not intended for further downstream fine-tuning for any other tasks. This model is trained on a classified dataset for text classification.\\'}', metadata={})]", "category": "generic"}
{"question_id": 459, "text": " Calculate the distance between the objects in an image for an autonomous vehicle driving in a parking lot.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Object Detection\\', \\'api_name\\': \\'keremberke/yolov8m-forklift-detection\\', \\'api_call\\': \"YOLO(\\'keremberke/yolov8m-forklift-detection\\')\", \\'api_arguments\\': {\\'image\\': \\'URL or local path to the image\\'}, \\'python_environment_requirements\\': [\\'ultralyticsplus==0.0.23\\', \\'ultralytics==8.0.21\\'], \\'example_code\\': [\\'from ultralyticsplus import YOLO, render_result\\', \"model = YOLO(\\'keremberke/yolov8m-forklift-detection\\')\", \"model.overrides[\\'conf\\'] = 0.25\", \"model.overrides[\\'iou\\'] = 0.45\", \"model.overrides[\\'agnostic_nms\\'] = False\", \"model.overrides[\\'max_det\\'] = 1000\", \"image = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\", \\'results = model.predict(image)\\', \\'print(results[0].boxes)\\', \\'render = render_result(model=model, image=image, result=results[0])\\', \\'render.show()\\'], \\'performance\\': {\\'dataset\\': \\'forklift-object-detection\\', \\'accuracy\\': 0.846}, \\'description\\': \\'A YOLOv8 model for detecting forklifts and persons in images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 460, "text": " We want a technology to be developed in the field of autonomous vehicles. Provide an example to estimate the depth of the environment using monocular images.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Depth Estimation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Depth Estimation\\', \\'api_name\\': \\'glpn-kitti\\', \\'api_call\\': \"GLPNForDepthEstimation.from_pretrained(\\'vinvino02/glpn-kitti\\')\", \\'api_arguments\\': \\'images, return_tensors\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'from transformers import GLPNFeatureExtractor, GLPNForDepthEstimation\\\\nimport torch\\\\nimport numpy as np\\\\nfrom PIL import Image\\\\nimport requests\\\\nurl = http://images.cocodataset.org/val2017/000000039769.jpg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nfeature_extractor = GLPNFeatureExtractor.from_pretrained(vinvino02/glpn-kitti)\\\\nmodel = GLPNForDepthEstimation.from_pretrained(vinvino02/glpn-kitti)\\\\ninputs = feature_extractor(images=image, return_tensors=pt)\\\\nwith torch.no_grad():\\\\n outputs = model(**inputs)\\\\n predicted_depth = outputs.predicted_depth\\\\nprediction = torch.nn.functional.interpolate(\\\\n predicted_depth.unsqueeze(1),\\\\n size=image.size[::-1],\\\\n mode=bicubic,\\\\n align_corners=False,\\\\n)\\\\noutput = prediction.squeeze().cpu().numpy()\\\\nformatted = (output * 255 / np.max(output)).astype(uint8)\\\\ndepth = Image.fromarray(formatted)\\', \\'performance\\': {\\'dataset\\': \\'KITTI\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Global-Local Path Networks (GLPN) model trained on KITTI for monocular depth estimation. It was introduced in the paper Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth by Kim et al. and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 461, "text": " As a construction company, we want to have a system to estimate the depth of images taken from different construction sites.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Depth Estimation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Depth Estimation\\', \\'api_name\\': \\'glpn-kitti\\', \\'api_call\\': \"GLPNForDepthEstimation.from_pretrained(\\'vinvino02/glpn-kitti\\')\", \\'api_arguments\\': \\'images, return_tensors\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'from transformers import GLPNFeatureExtractor, GLPNForDepthEstimation\\\\nimport torch\\\\nimport numpy as np\\\\nfrom PIL import Image\\\\nimport requests\\\\nurl = http://images.cocodataset.org/val2017/000000039769.jpg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nfeature_extractor = GLPNFeatureExtractor.from_pretrained(vinvino02/glpn-kitti)\\\\nmodel = GLPNForDepthEstimation.from_pretrained(vinvino02/glpn-kitti)\\\\ninputs = feature_extractor(images=image, return_tensors=pt)\\\\nwith torch.no_grad():\\\\n outputs = model(**inputs)\\\\n predicted_depth = outputs.predicted_depth\\\\nprediction = torch.nn.functional.interpolate(\\\\n predicted_depth.unsqueeze(1),\\\\n size=image.size[::-1],\\\\n mode=bicubic,\\\\n align_corners=False,\\\\n)\\\\noutput = prediction.squeeze().cpu().numpy()\\\\nformatted = (output * 255 / np.max(output)).astype(uint8)\\\\ndepth = Image.fromarray(formatted)\\', \\'performance\\': {\\'dataset\\': \\'KITTI\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Global-Local Path Networks (GLPN) model trained on KITTI for monocular depth estimation. It was introduced in the paper Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth by Kim et al. and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 462, "text": " Our team has built a Japanese language learning app for students. We want to make sure that the user-submitted image is an anime art created by humans and not AI-generated.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Image-to-Text\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'kha-white/manga-ocr-base\\', \\'api_call\\': \"pipeline(\\'ocr\\', model=\\'kha-white/manga-ocr-base\\')\", \\'api_arguments\\': \\'image\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'manga109s\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Optical character recognition for Japanese text, with the main focus being Japanese manga. It uses Vision Encoder Decoder framework. Manga OCR can be used as a general purpose printed Japanese OCR, but its main goal was to provide a high quality text recognition, robust against various scenarios specific to manga: both vertical and horizontal text, text with furigana, text overlaid on images, wide variety of fonts and font styles, and low quality images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 463, "text": " We need to classify the type of an image for an inventory.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'abhishek/autotrain-dog-vs-food\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'abhishek/autotrain-dog-vs-food\\')\", \\'api_arguments\\': \\'image_path\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'sasha/dog-food\\', \\'accuracy\\': 0.998}, \\'description\\': \"A pre-trained model for classifying images as either dog or food using Hugging Face\\'s AutoTrain framework.\"}', metadata={})]", "category": "generic"}
{"question_id": 464, "text": " Imagine we run an e-commerce platform and we need to tag new products with relevant categories based on their product images.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'abhishek/autotrain-dog-vs-food\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'abhishek/autotrain-dog-vs-food\\')\", \\'api_arguments\\': \\'image_path\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'sasha/dog-food\\', \\'accuracy\\': 0.998}, \\'description\\': \"A pre-trained model for classifying images as either dog or food using Hugging Face\\'s AutoTrain framework.\"}', metadata={})]", "category": "generic"}
{"question_id": 465, "text": " We are building an AI-based camera system to ensure safety on a construction site. Detect workers wearing hard hats in a given image.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Object Detection\\', \\'api_name\\': \\'keremberke/yolov8m-hard-hat-detection\\', \\'api_call\\': \"YOLO(\\'keremberke/yolov8m-hard-hat-detection\\')\", \\'api_arguments\\': {\\'image\\': \\'URL or local path to the image\\'}, \\'python_environment_requirements\\': [\\'ultralyticsplus==0.0.24\\', \\'ultralytics==8.0.23\\'], \\'example_code\\': \"from ultralyticsplus import YOLO, render_result\\\\n\\\\nmodel = YOLO(\\'keremberke/yolov8m-hard-hat-detection\\')\\\\n\\\\nmodel.overrides[\\'conf\\'] = 0.25\\\\nmodel.overrides[\\'iou\\'] = 0.45\\\\nmodel.overrides[\\'agnostic_nms\\'] = False\\\\nmodel.overrides[\\'max_det\\'] = 1000\\\\n\\\\nimage = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\\\\n\\\\nresults = model.predict(image)\\\\n\\\\nprint(results[0].boxes)\\\\nrender = render_result(model=model, image=image, result=results[0])\\\\nrender.show()\", \\'performance\\': {\\'dataset\\': \\'hard-hat-detection\\', \\'accuracy\\': 0.811}, \\'description\\': \"A YOLOv8 model for detecting hard hats in images. The model can distinguish between \\'Hardhat\\' and \\'NO-Hardhat\\' classes. It can be used to ensure safety compliance in construction sites or other industrial environments where hard hats are required.\"}', metadata={})]", "category": "generic"}
{"question_id": 466, "text": " My client needs help with security. They have a surveillance camera set up and we need to identify if any unknown person enters their property.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Object Detection\\', \\'api_name\\': \\'keremberke/yolov8s-csgo-player-detection\\', \\'api_call\\': \"YOLO(\\'keremberke/yolov8s-csgo-player-detection\\')\", \\'api_arguments\\': {\\'image\\': \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'}, \\'python_environment_requirements\\': [\\'ultralyticsplus==0.0.23\\', \\'ultralytics==8.0.21\\'], \\'example_code\\': \"from ultralyticsplus import YOLO, render_result\\\\nmodel = YOLO(\\'keremberke/yolov8s-csgo-player-detection\\')\\\\nmodel.overrides[\\'conf\\'] = 0.25\\\\nmodel.overrides[\\'iou\\'] = 0.45\\\\nmodel.overrides[\\'agnostic_nms\\'] = False\\\\nmodel.overrides[\\'max_det\\'] = 1000\\\\nimage = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\\\\nresults = model.predict(image)\\\\nprint(results[0].boxes)\\\\nrender = render_result(model=model, image=image, result=results[0])\\\\nrender.show()\", \\'performance\\': {\\'dataset\\': \\'csgo-object-detection\\', \\'accuracy\\': 0.886}, \\'description\\': \"A YOLOv8 model for detecting Counter-Strike: Global Offensive (CS:GO) players. Supports the labels [\\'ct\\', \\'cthead\\', \\'t\\', \\'thead\\'].\"}', metadata={})]", "category": "generic"}
{"question_id": 467, "text": " We need a system to control the access into parking lot. Analyze an image and find out the license plate numbers to detect whether it's an authorized vehicle or not.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'License Plate Detection\\', \\'api_name\\': \\'keremberke/yolov5s-license-plate\\', \\'api_call\\': \"yolov5.load(\\'keremberke/yolov5s-license-plate\\')\", \\'api_arguments\\': {\\'img\\': \\'image url or path\\', \\'size\\': \\'image resize dimensions\\', \\'augment\\': \\'optional, test time augmentation\\'}, \\'python_environment_requirements\\': \\'pip install -U yolov5\\', \\'example_code\\': [\\'import yolov5\\', \"model = yolov5.load(\\'keremberke/yolov5s-license-plate\\')\", \\'model.conf = 0.25\\', \\'model.iou = 0.45\\', \\'model.agnostic = False\\', \\'model.multi_label = False\\', \\'model.max_det = 1000\\', \"img = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\", \\'results = model(img, size=640)\\', \\'results = model(img, augment=True)\\', \\'predictions = results.pred[0]\\', \\'boxes = predictions[:, :4]\\', \\'scores = predictions[:, 4]\\', \\'categories = predictions[:, 5]\\', \\'results.show()\\', \"results.save(save_dir=\\'results/\\')\"], \\'performance\\': {\\'dataset\\': \\'keremberke/license-plate-object-detection\\', \\'accuracy\\': 0.985}, \\'description\\': \\'A YOLOv5 based license plate detection model trained on a custom dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 468, "text": " Analyze an image of an urban scene to identify and separate regions with different semantics, such as streets, pedestrians, buildings, and vehicles.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Segmentation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Segmentation\\', \\'api_name\\': \\'keremberke/yolov8n-pothole-segmentation\\', \\'api_call\\': \"YOLO(\\'keremberke/yolov8n-pothole-segmentation\\')\", \\'api_arguments\\': {\\'image\\': \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\', \\'conf\\': 0.25, \\'iou\\': 0.45, \\'agnostic_nms\\': False, \\'max_det\\': 1000}, \\'python_environment_requirements\\': {\\'ultralyticsplus\\': \\'0.0.23\\', \\'ultralytics\\': \\'8.0.21\\'}, \\'example_code\\': \"from ultralyticsplus import YOLO, render_result\\\\nmodel = YOLO(\\'keremberke/yolov8n-pothole-segmentation\\')\\\\nmodel.overrides[\\'conf\\'] = 0.25\\\\nmodel.overrides[\\'iou\\'] = 0.45\\\\nmodel.overrides[\\'agnostic_nms\\'] = False\\\\nmodel.overrides[\\'max_det\\'] = 1000\\\\nimage = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\\\\nresults = model.predict(image)\\\\nprint(results[0].boxes)\\\\nprint(results[0].masks)\\\\nrender = render_result(model=model, image=image, result=results[0])\\\\nrender.show()\", \\'performance\\': {\\'dataset\\': \\'pothole-segmentation\\', \\'accuracy\\': {\\'mAP@0.5(box)\\': 0.995, \\'mAP@0.5(mask)\\': 0.995}}, \\'description\\': \\'A YOLOv8 model for pothole segmentation in images. The model is trained on the pothole-segmentation dataset and achieves high accuracy in detecting potholes.\\'}', metadata={})]", "category": "generic"}
{"question_id": 469, "text": " We are developing an application to help customers visualize themselves wearing clothes available on our e-commerce website. We need identification of clothing items in the image.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Zero-Shot Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Zero-Shot Image Classification\\', \\'api_name\\': \\'patrickjohncyh/fashion-clip\\', \\'api_call\\': \"CLIPModel.from_pretrained(\\'patrickjohncyh/fashion-clip\\')\", \\'api_arguments\\': {\\'image\\': \\'File\\', \\'class_names\\': \\'String (comma-separated)\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import CLIPProcessor, CLIPModel; model = CLIPModel.from_pretrained(\\'patrickjohncyh/fashion-clip\\'); processor = CLIPProcessor.from_pretrained(\\'patrickjohncyh/fashion-clip\\'); inputs = processor(text=\\'blue shoes\\', images=image, return_tensors=\\'pt\\', padding=True); logits_per_image = model(**inputs).logits_per_image; probs = logits_per_image.softmax(dim=-1).tolist()[0]\", \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'FMNIST\\', \\'accuracy\\': 0.83}, {\\'name\\': \\'KAGL\\', \\'accuracy\\': 0.73}, {\\'name\\': \\'DEEP\\', \\'accuracy\\': 0.62}]}, \\'description\\': \\'FashionCLIP is a CLIP-based model developed to produce general product representations for fashion concepts. Leveraging the pre-trained checkpoint (ViT-B/32) released by OpenAI, it is trained on a large, high-quality novel fashion dataset to study whether domain specific fine-tuning of CLIP-like models is sufficient to produce product representations that are zero-shot transferable to entirely new datasets and tasks.\\'}', metadata={})]", "category": "generic"}
{"question_id": 470, "text": " We need an image segmentation solution for our smart city planning project. The solution should be able to handle different types of segmentation tasks, from semantic to instance to panoptic.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Segmentation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Semantic Segmentation\\', \\'api_name\\': \\'nvidia/segformer-b2-finetuned-cityscapes-1024-1024\\', \\'api_call\\': \"SegformerForSemanticSegmentation.from_pretrained(\\'nvidia/segformer-b2-finetuned-cityscapes-1024-1024\\')\", \\'api_arguments\\': {\\'images\\': \\'image\\', \\'return_tensors\\': \\'pt\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'latest\\', \\'PIL\\': \\'latest\\', \\'requests\\': \\'latest\\'}, \\'example_code\\': \"from transformers import SegformerFeatureExtractor, SegformerForSemanticSegmentation\\\\nfrom PIL import Image\\\\nimport requests\\\\nfeature_extractor = SegformerFeatureExtractor.from_pretrained(\\'nvidia/segformer-b2-finetuned-cityscapes-1024-1024\\')\\\\nmodel = SegformerForSemanticSegmentation.from_pretrained(\\'nvidia/segformer-b2-finetuned-cityscapes-1024-1024\\')\\\\nurl = \\'http://images.cocodataset.org/val2017/000000039769.jpg\\'\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\ninputs = feature_extractor(images=image, return_tensors=\\'pt\\')\\\\noutputs = model(**inputs)\\\\nlogits = outputs.logits\", \\'performance\\': {\\'dataset\\': \\'Cityscapes\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'SegFormer model fine-tuned on CityScapes at resolution 1024x1024. It was introduced in the paper SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers by Xie et al. and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 471, "text": " We have a road safety app that helps identify potholes. Can we use an image segmentation model to detect if there are potholes in the pictures of roads?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Segmentation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Segmentation\\', \\'api_name\\': \\'keremberke/yolov8s-pothole-segmentation\\', \\'api_call\\': \"YOLO(\\'keremberke/yolov8s-pothole-segmentation\\')\", \\'api_arguments\\': {\\'image\\': \\'URL or local path to the image\\'}, \\'python_environment_requirements\\': {\\'ultralyticsplus\\': \\'0.0.23\\', \\'ultralytics\\': \\'8.0.21\\'}, \\'example_code\\': \"from ultralyticsplus import YOLO, render_result\\\\nmodel = YOLO(\\'keremberke/yolov8s-pothole-segmentation\\')\\\\nmodel.overrides[\\'conf\\'] = 0.25\\\\nmodel.overrides[\\'iou\\'] = 0.45\\\\nmodel.overrides[\\'agnostic_nms\\'] = False\\\\nmodel.overrides[\\'max_det\\'] = 1000\\\\nimage = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\\\\nresults = model.predict(image)\\\\nprint(results[0].boxes)\\\\nprint(results[0].masks)\\\\nrender = render_result(model=model, image=image, result=results[0])\\\\nrender.show()\", \\'performance\\': {\\'dataset\\': \\'pothole-segmentation\\', \\'accuracy\\': {\\'mAP@0.5(box)\\': 0.928, \\'mAP@0.5(mask)\\': 0.928}}, \\'description\\': \\'A YOLOv8 model for pothole segmentation. This model detects potholes in images and outputs bounding boxes and masks for the detected potholes.\\'}', metadata={})]", "category": "generic"}
{"question_id": 472, "text": " We are making a mobile app related to fitness. We need to estimate the human pose from an image of a user performing an exercise.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Human Pose Estimation\\', \\'api_name\\': \\'lllyasviel/sd-controlnet-openpose\\', \\'api_call\\': \"ControlNetModel.from_pretrained(\\'lllyasviel/sd-controlnet-openpose\\')\", \\'api_arguments\\': {\\'text\\': \\'chef in the kitchen\\', \\'image\\': \\'image\\', \\'num_inference_steps\\': 20}, \\'python_environment_requirements\\': {\\'diffusers\\': \\'pip install diffusers\\', \\'transformers\\': \\'pip install transformers\\', \\'accelerate\\': \\'pip install accelerate\\', \\'controlnet_aux\\': \\'pip install controlnet_aux\\'}, \\'example_code\\': \"from PIL import Image\\\\nfrom diffusers import StableDiffusionControlNetPipeline, ControlNetModel, UniPCMultistepScheduler\\\\nimport torch\\\\nfrom controlnet_aux import OpenposeDetector\\\\nfrom diffusers.utils import load_image\\\\nopenpose = OpenposeDetector.from_pretrained(\\'lllyasviel/ControlNet\\')\\\\nimage = load_image(https://huggingface.co/lllyasviel/sd-controlnet-openpose/resolve/main/images/pose.png)\\\\nimage = openpose(image)\\\\ncontrolnet = ControlNetModel.from_pretrained(\\\\n lllyasviel/sd-controlnet-openpose, torch_dtype=torch.float16\\\\n)\\\\npipe = StableDiffusionControlNetPipeline.from_pretrained(\\\\n runwayml/stable-diffusion-v1-5, controlnet=controlnet, safety_checker=None, torch_dtype=torch.float16\\\\n)\\\\npipe.scheduler = UniPCMultistepScheduler.from_config(pipe.scheduler.config)\\\\npipe.enable_xformers_memory_efficient_attention()\\\\npipe.enable_model_cpu_offload()\\\\nimage = pipe(chef in the kitchen, image, num_inference_steps=20).images[0]\\\\nimage.save(\\'images/chef_pose_out.png\\')\", \\'performance\\': {\\'dataset\\': \\'200k pose-image, caption pairs\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'ControlNet is a neural network structure to control diffusion models by adding extra conditions. This checkpoint corresponds to the ControlNet conditioned on Human Pose Estimation. It can be used in combination with Stable Diffusion.\\'}', metadata={})]", "category": "generic"}
{"question_id": 473, "text": " I wish to edit my images by detecting straight lines and controlling the diffusion models in the image's diffusion process.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image Diffusion Models\\', \\'api_name\\': \\'lllyasviel/control_v11p_sd15s2_lineart_anime\\', \\'api_call\\': \"ControlNetModel.from_pretrained(\\'lllyasviel/control_v11p_sd15s2_lineart_anime\\')\", \\'api_arguments\\': {\\'checkpoint\\': \\'lllyasviel/control_v11p_sd15s2_lineart_anime\\', \\'torch_dtype\\': \\'torch.float16\\'}, \\'python_environment_requirements\\': [\\'pip install diffusers transformers accelerate\\', \\'pip install controlnet_aux==0.3.0\\'], \\'example_code\\': [\\'import torch\\', \\'import os\\', \\'from huggingface_hub import HfApi\\', \\'from pathlib import Path\\', \\'from diffusers.utils import load_image\\', \\'from PIL import Image\\', \\'import numpy as np\\', \\'from controlnet_aux import LineartAnimeDetector\\', \\'from transformers import CLIPTextModel\\', \\'from diffusers import (\\', \\' ControlNetModel,\\', \\' StableDiffusionControlNetPipeline,\\', \\' UniPCMultistepScheduler,\\', \\')\\', \\'checkpoint = lllyasviel/control_v11p_sd15s2_lineart_anime\\', \\'image = load_image(\\', \\' https://huggingface.co/lllyasviel/control_v11p_sd15s2_lineart_anime/resolve/main/images/input.png\\', \\')\\', \\'image = image.resize((512, 512))\\', \\'prompt = A warrior girl in the jungle\\', \\'processor = LineartAnimeDetector.from_pretrained(lllyasviel/Annotators)\\', \\'control_image = processor(image)\\', \\'control_image.save(./images/control.png)\\', \\'text_encoder = CLIPTextModel.from_pretrained(runwayml/stable-diffusion-v1-5, subfolder=text_encoder, num_hidden_layers=11, torch_dtype=torch.float16)\\', \\'controlnet = ControlNetModel.from_pretrained(checkpoint, torch_dtype=torch.float16)\\', \\'pipe = StableDiffusionControlNetPipeline.from_pretrained(\\', \\' runwayml/stable-diffusion-v1-5, text_encoder=text_encoder, controlnet=controlnet, torch_dtype=torch.float16\\', \\')\\', \\'pipe.scheduler = UniPCMultistepScheduler.from_config(pipe.scheduler.config)\\', \\'pipe.enable_model_cpu_offload()\\', \\'generator = torch.manual_seed(0)\\', \\'image = pipe(prompt, num_inference_steps=30, generator=generator, image=control_image).images[0]\\', \"image.save(\\'images/image_out.png\\')\"], \\'performance\\': {\\'dataset\\': \\'Not specified\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'ControlNet is a neural network structure to control diffusion models by adding extra conditions. This checkpoint corresponds to the ControlNet conditioned on lineart_anime images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 474, "text": " A photographer wants to create artistic interpretations of some of her pictures. Let's help her to get variated styles of her image.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'vintedois-diffusion-v0-1\\', \\'api_call\\': \"pipeline(\\'text-to-image\\', model=\\'22h/vintedois-diffusion-v0-1\\')\", \\'api_arguments\\': [\\'prompt\\', \\'CFG Scale\\', \\'Scheduler\\', \\'Steps\\', \\'Seed\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"text2img(\\'photo of an old man in a jungle, looking at the camera\\', CFG Scale=7.5, Scheduler=\\'diffusers.EulerAncestralDiscreteScheduler\\', Steps=30, Seed=44)\", \\'performance\\': {\\'dataset\\': \\'large amount of high quality images\\', \\'accuracy\\': \\'not specified\\'}, \\'description\\': \\'Vintedois (22h) Diffusion model trained by Predogl and piEsposito with open weights, configs and prompts. This model generates beautiful images without a lot of prompt engineering. It can also generate high fidelity faces with a little amount of steps.\\'}', metadata={})]", "category": "generic"}
{"question_id": 475, "text": " We have a client that manages a car website for selling used cars and they need new images of cars for the website.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Image-to-Text\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Image Captioning\\', \\'api_name\\': \\'blip-image-captioning-large\\', \\'api_call\\': \\'BlipForConditionalGeneration.from_pretrained(Salesforce/blip-image-captioning-large)\\', \\'api_arguments\\': {\\'raw_image\\': \\'Image\\', \\'text\\': \\'Optional Text\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'BlipProcessor, BlipForConditionalGeneration\\', \\'PIL\\': \\'Image\\', \\'requests\\': \\'requests\\'}, \\'example_code\\': {\\'import_requests\\': \\'import requests\\', \\'import_PIL\\': \\'from PIL import Image\\', \\'import_transformers\\': \\'from transformers import BlipProcessor, BlipForConditionalGeneration\\', \\'load_processor\\': \\'processor = BlipProcessor.from_pretrained(Salesforce/blip-image-captioning-large)\\', \\'load_model\\': \\'model = BlipForConditionalGeneration.from_pretrained(Salesforce/blip-image-captioning-large)\\', \\'load_image\\': \"img_url = \\'https://storage.googleapis.com/sfr-vision-language-research/BLIP/demo.jpg\\'\\\\nraw_image = Image.open(requests.get(img_url, stream=True).raw).convert(\\'RGB\\')\", \\'conditional_captioning\\': \\'text = a photography of\\\\ninputs = processor(raw_image, text, return_tensors=pt)\\\\nout = model.generate(**inputs)\\\\nprint(processor.decode(out[0], skip_special_tokens=True))\\', \\'unconditional_captioning\\': \\'inputs = processor(raw_image, return_tensors=pt)\\\\nout = model.generate(**inputs)\\\\nprint(processor.decode(out[0], skip_special_tokens=True))\\'}, \\'performance\\': {\\'dataset\\': \\'COCO\\', \\'accuracy\\': {\\'image-text retrieval\\': \\'+2.7% recall@1\\', \\'image captioning\\': \\'+2.8% CIDEr\\', \\'VQA\\': \\'+1.6% VQA score\\'}}, \\'description\\': \\'BLIP is a Vision-Language Pre-training (VLP) framework that achieves state-of-the-art results on a wide range of vision-language tasks, such as image-text retrieval, image captioning, and VQA. It effectively utilizes noisy web data by bootstrapping the captions, where a captioner generates synthetic captions and a filter removes the noisy ones.\\'}', metadata={})]", "category": "generic"}
{"question_id": 476, "text": " I am interested in building a wall of generated pictures for my gallery. My specifications include a size of 256x256 pixels.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Unconditional Image Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Unconditional Image Generation\\', \\'api_name\\': \\'johnowhitaker/sd-class-wikiart-from-bedrooms\\', \\'api_call\\': \"DDPMPipeline.from_pretrained(\\'johnowhitaker/sd-class-wikiart-from-bedrooms\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'diffusers\\', \\'example_code\\': \"from diffusers import DDPMPipeline\\\\npipeline = DDPMPipeline.from_pretrained(\\'johnowhitaker/sd-class-wikiart-from-bedrooms\\')\\\\nimage = pipeline().images[0]\\\\nimage\", \\'performance\\': {\\'dataset\\': \\'https://huggingface.co/datasets/huggan/wikiart\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This model is a diffusion model initialized from https://huggingface.co/google/ddpm-bedroom-256 and trained for 5000 steps on https://huggingface.co/datasets/huggan/wikiart.\\'}', metadata={})]", "category": "generic"}
{"question_id": 477, "text": " We are a product company selling personalized gadgets, and we want to build a recommender system that shows generated, high-resolution images of human faces on our website.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'prompthero/openjourney-v4\\', \\'api_call\\': \"pipeline(\\'text-to-image\\', model=\\'prompthero/openjourney-v4\\')\", \\'api_arguments\\': {\\'text\\': \\'string\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"generate_image(\\'your text here\\')\", \\'performance\\': {\\'dataset\\': \\'Midjourney v4 images\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Openjourney v4 is trained on +124k Midjourney v4 images by PromptHero. It is used for generating images based on text inputs.\\'}', metadata={})]", "category": "generic"}
{"question_id": 478, "text": " I want to build an AI-based software that can identify the activities or actions in a video clip.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Voice Activity Detection\\', \\'api_name\\': \\'d4data/Indian-voice-cloning\\', \\'api_call\\': \"pipeline(\\'voice-activity-detection\\', model=\\'d4data/Indian-voice-cloning\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A model for detecting voice activity in Indian languages.\\'}', metadata={})]", "category": "generic"}
{"question_id": 479, "text": " The marketing team wants a tool to quickly classify new advertisement videos.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'chavinlo/TempoFunk\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'chavinlo/TempoFunk\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Text-to-Video model using Hugging Face Transformers library. Model is capable of generating video content based on the input text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 480, "text": " A sports league wants to analyze their videos and extract information on game highlights.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'text2text-generation\\', \\'api_name\\': \\'tuner007/pegasus_summarizer\\', \\'api_call\\': \"PegasusForConditionalGeneration.from_pretrained(\\'tuner007/pegasus_summarizer\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'pip install sentencepiece\\'], \\'example_code\\': \"context = \\\\nIndia wicket-keeper batsman Rishabh Pant has said someone from the crowd threw a ball on pacer Mohammed Siraj while he was fielding in the ongoing third Test against England on Wednesday. Pant revealed the incident made India skipper Virat Kohli upset. I think, somebody threw a ball inside, at Siraj, so he [Kohli] was upset, said Pant in a virtual press conference after the close of the first day\\'s play.You can say whatever you want to chant, but don\\'t throw things at the fielders and all those things. It is not good for cricket, I guess, he added.In the third session of the opening day of the third Test, a section of spectators seemed to have asked Siraj the score of the match to tease the pacer. The India pacer however came with a brilliant reply as he gestured 1-0 (India leading the Test series) towards the crowd.Earlier this month, during the second Test match, there was some bad crowd behaviour on a show as some unruly fans threw champagne corks at India batsman KL Rahul.Kohli also intervened and he was seen gesturing towards the opening batsman to know more about the incident. An over later, the TV visuals showed that many champagne corks were thrown inside the playing field, and the Indian players were visibly left frustrated.Coming back to the game, after bundling out India for 78, openers Rory Burns and Haseeb Hameed ensured that England took the honours on the opening day of the ongoing third Test.At stumps, England\\'s score reads 120/0 and the hosts have extended their lead to 42 runs. For the Three Lions, Burns (52) and Hameed (60) are currently unbeaten at the crease.Talking about the pitch on opening day, Pant said, They took the heavy roller, the wicket was much more settled down, and they batted nicely also, he said. But when we batted, the wicket was slightly soft, and they bowled in good areas, but we could have applied [ourselves] much better.Both England batsmen managed to see off the final session and the hosts concluded the opening day with all ten wickets intact, extending the lead to 42.(ANI)\\\\n\\\\nget_response(context)\", \\'performance\\': {\\'dataset\\': \\'cnn_dailymail\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.604, \\'ROUGE-2\\': 14.64, \\'ROUGE-L\\': 23.884, \\'ROUGE-LSUM\\': 32.902, \\'loss\\': 2.576, \\'gen_len\\': 76.398}}, \\'description\\': \\'PEGASUS fine-tuned for summarization\\'}', metadata={})]", "category": "generic"}
{"question_id": 481, "text": " We are a sports broadcasting company, and we need to automatically identify the sports events taking place in the videos we receive.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'chavinlo/TempoFunk\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'chavinlo/TempoFunk\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Text-to-Video model using Hugging Face Transformers library. Model is capable of generating video content based on the input text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 482, "text": " My organization wants to create a video categorization tool to classify and categorize various videos. Utilize the appropriate API to build this video classifier.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'camenduru/text2-video-zero\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'camenduru/text2-video-zero\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'This model is used for generating videos from text inputs. It is based on the Hugging Face framework and can be used with the transformers library. The model is trained on a variety of text and video datasets, and can be used for tasks such as video summarization, video generation from text prompts, and more.\\'}', metadata={})]", "category": "generic"}
{"question_id": 483, "text": " We have a surveillance camera in our backyard. We would like to analyze the captured videos to recognize the activities taking place in the backyard.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'ImRma/Brucelee\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'ImRma/Brucelee\\')\", \\'api_arguments\\': [\\'your_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Hugging Face model for converting Persian and English text into video.\\'}', metadata={})]", "category": "generic"}
{"question_id": 484, "text": " We need to classify an image's content and check if it contains a cat or a dog.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'julien-c/hotdog-not-hotdog\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'julien-c/hotdog-not-hotdog\\')\", \\'api_arguments\\': \\'image\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': 0.825}, \\'description\\': \\'A model that classifies images as hotdog or not hotdog.\\'}', metadata={})]", "category": "generic"}
{"question_id": 485, "text": " Analyze the type of plants in the image provided and provide the name of the probable plant.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Image-to-Image\\', \\'api_name\\': \\'GreeneryScenery/SheepsControlV3\\', \\'api_call\\': \"pipeline(\\'image-to-image\\', model=\\'GreeneryScenery/SheepsControlV3\\')\", \\'api_arguments\\': {\\'image\\': \\'Path to image file\\', \\'text_guidance\\': \\'Optional text guidance for the model\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'latest\\', \\'torch\\': \\'latest\\'}, \\'example_code\\': [\\'from transformers import pipeline\\', \"model = pipeline(\\'image-to-image\\', model=\\'GreeneryScenery/SheepsControlV3\\')\", \"result = model({\\'image\\': \\'path/to/image.jpg\\', \\'text_guidance\\': \\'Optional text guidance\\'})\"], \\'performance\\': {\\'dataset\\': \\'GreeneryScenery/SheepsControlV3\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'GreeneryScenery/SheepsControlV3 is a model for image-to-image tasks. It can be used to generate images based on the input image and optional text guidance. The model has some limitations, such as the conditioning image not affecting the output image much. Improvements can be made by training for more epochs, using better prompts, and preprocessing the data.\\'}', metadata={})]", "category": "generic"}
{"question_id": 486, "text": " I want to organize my images based on the scene content. The categories I want are landscape, cityscape, beach, forest, and animals.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'abhishek/autotrain-dog-vs-food\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'abhishek/autotrain-dog-vs-food\\')\", \\'api_arguments\\': \\'image_path\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'sasha/dog-food\\', \\'accuracy\\': 0.998}, \\'description\\': \"A pre-trained model for classifying images as either dog or food using Hugging Face\\'s AutoTrain framework.\"}', metadata={})]", "category": "generic"}
{"question_id": 487, "text": " We run an online store for selling electronic devices. We need to classify product images from our inventory and organize them.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'abhishek/autotrain-dog-vs-food\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'abhishek/autotrain-dog-vs-food\\')\", \\'api_arguments\\': \\'image_path\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'sasha/dog-food\\', \\'accuracy\\': 0.998}, \\'description\\': \"A pre-trained model for classifying images as either dog or food using Hugging Face\\'s AutoTrain framework.\"}', metadata={})]", "category": "generic"}
{"question_id": 488, "text": " Our customer wants to analyze the sentiment of their customers' feedback. The feedback is in Spanish.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'finiteautomata/beto-sentiment-analysis\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'finiteautomata/beto-sentiment-analysis\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'Hugging Face Transformers library\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'TASS 2020 corpus\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Model trained with TASS 2020 corpus (around ~5k tweets) of several dialects of Spanish. Base model is BETO, a BERT model trained in Spanish. Uses POS, NEG, NEU labels.\\'}', metadata={})]", "category": "generic"}
{"question_id": 489, "text": " A stock investor is looking to analyze the sentiment of a stock forum, such as StockTwits, to gain insights into the market sentiment for a specific stock.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Sentiment Inferencing for stock-related comments\\', \\'api_name\\': \\'zhayunduo/roberta-base-stocktwits-finetuned\\', \\'api_call\\': \"RobertaForSequenceClassification.from_pretrained(\\'zhayunduo/roberta-base-stocktwits-finetuned\\')\", \\'api_arguments\\': {\\'model\\': \\'RobertaForSequenceClassification\\', \\'tokenizer\\': \\'RobertaTokenizer\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import RobertaForSequenceClassification, RobertaTokenizer\\\\nfrom transformers import pipeline\\\\nimport pandas as pd\\\\nimport emoji\\\\ntokenizer_loaded = RobertaTokenizer.from_pretrained(\\'zhayunduo/roberta-base-stocktwits-finetuned\\')\\\\nmodel_loaded = RobertaForSequenceClassification.from_pretrained(\\'zhayunduo/roberta-base-stocktwits-finetuned\\')\\\\nnlp = pipeline(text-classification, model=model_loaded, tokenizer=tokenizer_loaded)\\\\nsentences = pd.Series([\\'just buy\\',\\'just sell it\\',\\'entity rocket to the sky!\\',\\'go down\\',\\'even though it is going up, I still think it will not keep this trend in the near future\\'])\\\\nsentences = list(sentences)\\\\nresults = nlp(sentences)\\\\nprint(results)\", \\'performance\\': {\\'dataset\\': \\'stocktwits\\', \\'accuracy\\': 0.9343}, \\'description\\': \"This model is fine-tuned with roberta-base model on 3,200,000 comments from stocktwits, with the user-labeled tags \\'Bullish\\' or \\'Bearish\\'.\"}', metadata={})]", "category": "generic"}
{"question_id": 490, "text": " A forum moderator wants a tool to assess user-generated comments for toxic content. How does this model help?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'martin-ha/toxic-comment-model\\', \\'api_call\\': \"pipeline(model=\\'martin-ha/toxic-comment-model\\')\", \\'api_arguments\\': {\\'model_path\\': \\'martin-ha/toxic-comment-model\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import AutoModelForSequenceClassification, AutoTokenizer, TextClassificationPipeline\\\\nmodel_path = martin-ha/toxic-comment-model\\\\ntokenizer = AutoTokenizer.from_pretrained(model_path)\\\\nmodel = AutoModelForSequenceClassification.from_pretrained(model_path)\\\\npipeline = TextClassificationPipeline(model=model, tokenizer=tokenizer)\\\\nprint(pipeline(\\'This is a test text.\\'))\", \\'performance\\': {\\'dataset\\': \\'held-out test set\\', \\'accuracy\\': 0.94, \\'f1-score\\': 0.59}, \\'description\\': \\'This model is a fine-tuned version of the DistilBERT model to classify toxic comments.\\'}', metadata={})]", "category": "generic"}
{"question_id": 491, "text": " I am developing a news analysis platform. I need to predict the named entities from the articles.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'finiteautomata/beto-sentiment-analysis\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'finiteautomata/beto-sentiment-analysis\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'Hugging Face Transformers library\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'TASS 2020 corpus\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Model trained with TASS 2020 corpus (around ~5k tweets) of several dialects of Spanish. Base model is BETO, a BERT model trained in Spanish. Uses POS, NEG, NEU labels.\\'}', metadata={})]", "category": "generic"}
{"question_id": 492, "text": " We need to analyze the user's text for extracting entities and improve our virtual assistant interaction.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'hyunwoongko/blenderbot-9B\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'hyunwoongko/blenderbot-9B\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'Input a message to start chatting with hyunwoongko/blenderbot-9B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high-performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi-turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 493, "text": " Extract the named entities from a given text snippet.\\n###Input: \\\"On June 7th, Jane Smith visited the Empire State Building in New York with an entry fee of 35 dollars.\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Token Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Named Entity Recognition\\', \\'api_name\\': \\'flair/ner-english-ontonotes\\', \\'api_call\\': \"SequenceTagger.load(\\'flair/ner-english-ontonotes\\')\", \\'api_arguments\\': [\\'sentence\\'], \\'python_environment_requirements\\': [\\'flair\\'], \\'example_code\\': \"from flair.data import Sentence\\\\nfrom flair.models import SequenceTagger\\\\ntagger = SequenceTagger.load(flair/ner-english-ontonotes)\\\\nsentence = Sentence(On September 1st George Washington won 1 dollar.)\\\\ntagger.predict(sentence)\\\\nprint(sentence)\\\\nfor entity in sentence.get_spans(\\'ner\\'):\\\\n    print(entity)\", \\'performance\\': {\\'dataset\\': \\'Ontonotes\\', \\'accuracy\\': \\'89.27\\'}, \\'description\\': \\'This is the 18-class NER model for English that ships with Flair. It predicts 18 tags such as cardinal value, date value, event name, building name, geo-political entity, language name, law name, location name, money name, affiliation, ordinal value, organization name, percent value, person name, product name, quantity value, time value, and name of work of art. Based on Flair embeddings and LSTM-CRF.\\'}', metadata={})]", "category": "generic"}
{"question_id": 494, "text": " As a surveillance analyst, my main task this week is to find out all the names and locations mentioned in the online chat rooms.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Generation\\', \\'api_name\\': \\'mywateriswet/ShuanBot\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'mywateriswet/ShuanBot\\')\", \\'api_arguments\\': \\'message\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"response = chatbot(\\'What is your name?\\')\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'ShuanBot is a conversational chatbot model based on the GPT-2 architecture. It can be used for generating human-like responses in a chat context.\\'}', metadata={})]", "category": "generic"}
{"question_id": 495, "text": " I need a program to identify the entities like persons, locations, organizations, and other names in a given German text.\\n \n Use this API documentation for reference: [Document(page_content=\"{'domain': 'Natural Language Processing Text Classification', 'framework': 'Hugging Face Transformers', 'functionality': 'German Sentiment Classification', 'api_name': 'oliverguhr/german-sentiment-bert', 'api_call': 'SentimentModel()', 'api_arguments': ['texts'], 'python_environment_requirements': 'pip install germansentiment', 'example_code': ['from germansentiment import SentimentModel', 'model = SentimentModel()', 'texts = [', ' Mit keinem guten Ergebniss,Das ist gar nicht mal so gut,', ' Total awesome!,nicht so schlecht wie erwartet,', ' Der Test verlief positiv.,Sie fhrt ein grnes Auto.]', 'result = model.predict_sentiment(texts)', 'print(result)'], 'performance': {'dataset': ['holidaycheck', 'scare', 'filmstarts', 'germeval', 'PotTS', 'emotions', 'sb10k', 'Leipzig Wikipedia Corpus 2016', 'all'], 'accuracy': [0.9568, 0.9418, 0.9021, 0.7536, 0.678, 0.9649, 0.7376, 0.9967, 0.9639]}, 'description': 'This model was trained for sentiment classification of German language texts. The model uses the Googles Bert architecture and was trained on 1.834 million German-language samples. The training data contains texts from various domains like Twitter, Facebook and movie, app and hotel reviews.'}\", metadata={})]", "category": "generic"}
{"question_id": 496, "text": " We are a multi-national company that would like to better understand global events and extract relevant named entities across 9 languages (de, en, es, fr, it, nl, pl, pt, ru). We would like to utilize a pre-trained NER model.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Token Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Named Entity Recognition\\', \\'api_name\\': \\'Davlan/bert-base-multilingual-cased-ner-hrl\\', \\'api_call\\': \"AutoModelForTokenClassification.from_pretrained(\\'Davlan/bert-base-multilingual-cased-ner-hrl\\')\", \\'api_arguments\\': {\\'model\\': \\'AutoModelForTokenClassification.from_pretrained(Davlan/bert-base-multilingual-cased-ner-hrl)\\', \\'tokenizer\\': \\'AutoTokenizer.from_pretrained(Davlan/bert-base-multilingual-cased-ner-hrl)\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'from transformers import AutoTokenizer, AutoModelForTokenClassification, pipeline\\'}, \\'example_code\\': \\'example = Nader Jokhadar had given Syria the lead with a well-struck header in the seventh minute.; ner_results = nlp(example); print(ner_results)\\', \\'performance\\': {\\'dataset\\': {\\'Arabic\\': \\'ANERcorp\\', \\'German\\': \\'conll 2003\\', \\'English\\': \\'conll 2003\\', \\'Spanish\\': \\'conll 2002\\', \\'French\\': \\'Europeana Newspapers\\', \\'Italian\\': \\'Italian I-CAB\\', \\'Latvian\\': \\'Latvian NER\\', \\'Dutch\\': \\'conll 2002\\', \\'Portuguese\\': \\'Paramopama + Second Harem\\', \\'Chinese\\': \\'MSRA\\'}, \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'bert-base-multilingual-cased-ner-hrl is a Named Entity Recognition model for 10 high resourced languages (Arabic, German, English, Spanish, French, Italian, Latvian, Dutch, Portuguese and Chinese) based on a fine-tuned mBERT base model. It has been trained to recognize three types of entities: location (LOC), organizations (ORG), and person (PER).\\'}', metadata={})]", "category": "generic"}
{"question_id": 497, "text": " We have a multimedia app in the Korean language. To deal with customer queries automatically, we want to incorporate question answering capability.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'monologg/koelectra-small-v2-distilled-korquad-384\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=\\'monologg/koelectra-small-v2-distilled-korquad-384\\')\", \\'api_arguments\\': {\\'model\\': \\'monologg/koelectra-small-v2-distilled-korquad-384\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"nlp(question=\\'your_question\\', context=\\'your_context\\')\", \\'performance\\': {\\'dataset\\': \\'KorQuAD\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'A Korean Question Answering model based on Electra and trained on the KorQuAD dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 498, "text": " Guide me to create a quiz project where I will summarize an article into a paragraph and from the summary I will develop a question with some multiple options. I need to check the correct answer for that question.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'text2text-generation\\', \\'api_name\\': \\'tuner007/pegasus_summarizer\\', \\'api_call\\': \"PegasusForConditionalGeneration.from_pretrained(\\'tuner007/pegasus_summarizer\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'pip install sentencepiece\\'], \\'example_code\\': \"context = \\\\nIndia wicket-keeper batsman Rishabh Pant has said someone from the crowd threw a ball on pacer Mohammed Siraj while he was fielding in the ongoing third Test against England on Wednesday. Pant revealed the incident made India skipper Virat Kohli upset. I think, somebody threw a ball inside, at Siraj, so he [Kohli] was upset, said Pant in a virtual press conference after the close of the first day\\'s play.You can say whatever you want to chant, but don\\'t throw things at the fielders and all those things. It is not good for cricket, I guess, he added.In the third session of the opening day of the third Test, a section of spectators seemed to have asked Siraj the score of the match to tease the pacer. The India pacer however came with a brilliant reply as he gestured 1-0 (India leading the Test series) towards the crowd.Earlier this month, during the second Test match, there was some bad crowd behaviour on a show as some unruly fans threw champagne corks at India batsman KL Rahul.Kohli also intervened and he was seen gesturing towards the opening batsman to know more about the incident. An over later, the TV visuals showed that many champagne corks were thrown inside the playing field, and the Indian players were visibly left frustrated.Coming back to the game, after bundling out India for 78, openers Rory Burns and Haseeb Hameed ensured that England took the honours on the opening day of the ongoing third Test.At stumps, England\\'s score reads 120/0 and the hosts have extended their lead to 42 runs. For the Three Lions, Burns (52) and Hameed (60) are currently unbeaten at the crease.Talking about the pitch on opening day, Pant said, They took the heavy roller, the wicket was much more settled down, and they batted nicely also, he said. But when we batted, the wicket was slightly soft, and they bowled in good areas, but we could have applied [ourselves] much better.Both England batsmen managed to see off the final session and the hosts concluded the opening day with all ten wickets intact, extending the lead to 42.(ANI)\\\\n\\\\nget_response(context)\", \\'performance\\': {\\'dataset\\': \\'cnn_dailymail\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.604, \\'ROUGE-2\\': 14.64, \\'ROUGE-L\\': 23.884, \\'ROUGE-LSUM\\': 32.902, \\'loss\\': 2.576, \\'gen_len\\': 76.398}}, \\'description\\': \\'PEGASUS fine-tuned for summarization\\'}', metadata={})]", "category": "generic"}
{"question_id": 499, "text": " I have a website text about technology and I want to know if it represents a positive sentiment or a negative one.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'it5-base-news-summarization\\', \\'api_call\\': \"pipeline(\\'summarization\\', model=\\'it5/it5-base-news-summarization\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'newsum(Dal 31 maggio \u00e8 infine partita la piattaforma ITsART, a pi\u00f9 di un anno da quando \u2013 durante il primo lockdown \u2013 il ministro della Cultura Dario Franceschini ne aveva parlato come di \u00abuna sorta di Netflix della cultura\u00bb, pensata per \u00aboffrire a tutto il mondo la cultura italiana a pagamento\u00bb. \u00c8 presto per dare giudizi definitivi sulla piattaforma, e di certo sar\u00e0 difficile farlo anche pi\u00f9 avanti senza numeri precisi. Al momento, l\u2019unica cosa che si pu\u00f2 fare \u00e8 guardare com\u2019\u00e8 fatto il sito, contare quanti contenuti ci sono (circa 700 \u201ctitoli\u201d, tra film, documentari, spettacoli teatrali e musicali e altri eventi) e provare a dare un giudizio sul loro valore e sulla loro variet\u00e0. Intanto, una cosa notata da pi\u00f9 parti \u00e8 che diversi contenuti di ITsART sono a pagamento sulla piattaforma sebbene altrove, per esempio su RaiPlay, siano invece disponibili gratuitamente.)\\', \\'performance\\': {\\'dataset\\': \\'NewsSum-IT\\', \\'accuracy\\': {\\'Rouge1\\': 0.339, \\'Rouge2\\': 0.16, \\'RougeL\\': 0.263}}, \\'description\\': \\'IT5 Base model fine-tuned on news summarization on the Fanpage and Il Post corpora for Italian Language Understanding and Generation.\\'}', metadata={})]", "category": "generic"}
{"question_id": 500, "text": " Translate a French website description containing information about the company, services, and copyright notice into English.\\n###Input: \\\"Bienvenue sur notre site ! Nous sommes une entreprise sp\\u00e9cialis\\u00e9e dans la gestion des projets informatiques et la cr\\u00e9ation de logiciels sur mesure. Nos services incluent la conception, le d\\u00e9veloppement, la maintenance et le support 24/7. Tous droits r\\u00e9serv\\u00e9s.\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Translation\\', \\'api_name\\': \\'Helsinki-NLP/opus-mt-en-fr\\', \\'api_call\\': \"translate(\\'input_text\\', model=\\'Helsinki-NLP/opus-mt-en-fr\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'opus\\', \\'accuracy\\': {\\'BLEU\\': {\\'newsdiscussdev2015-enfr.en.fr\\': 33.8, \\'newsdiscusstest2015-enfr.en.fr\\': 40.0, \\'newssyscomb2009.en.fr\\': 29.8, \\'news-test2008.en.fr\\': 27.5, \\'newstest2009.en.fr\\': 29.4, \\'newstest2010.en.fr\\': 32.7, \\'newstest2011.en.fr\\': 34.3, \\'newstest2012.en.fr\\': 31.8, \\'newstest2013.en.fr\\': 33.2, \\'Tatoeba.en.fr\\': 50.5}}}, \\'description\\': \\'Helsinki-NLP/opus-mt-en-fr is a translation model that translates English text to French using the Hugging Face Transformers library. It is based on the OPUS dataset and uses a transformer-align architecture with normalization and SentencePiece pre-processing.\\'}', metadata={})]", "category": "generic"}
{"question_id": 501, "text": " We have a list of documents written in multiple Romance languages, including texts in French, Spanish, and Italian. We want to make these texts accessible to our English-speaking audience by translating them.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Translation\\', \\'api_name\\': \\'opus-mt-en-ROMANCE\\', \\'api_call\\': \"pipeline(\\'translation_en_to_ROMANCE\\', model=\\'Helsinki-NLP/opus-mt-en-ROMANCE\\')\", \\'api_arguments\\': \\'source languages, target languages\\', \\'python_environment_requirements\\': \\'transformers, torch\\', \\'example_code\\': \"from transformers import pipeline\\\\ntranslation = pipeline(\\'translation_en_to_ROMANCE\\', model=\\'Helsinki-NLP/opus-mt-en-ROMANCE\\')\\\\ntranslated_text = translation(\\'Hello, how are you?\\', tgt_lang=\\'es\\')\", \\'performance\\': {\\'dataset\\': \\'opus\\', \\'accuracy\\': {\\'BLEU\\': 50.1, \\'chr-F\\': 0.693}}, \\'description\\': \\'A translation model trained on the OPUS dataset that supports translation between English and various Romance languages. It uses a transformer architecture and requires a sentence initial language token in the form of >>id<< (id = valid target language ID).\\'}', metadata={})]", "category": "generic"}
{"question_id": 502, "text": " We have an international audience for our website and require our website content to be translated into multiple languages for better understanding.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'PyTorch Transformers\\', \\'functionality\\': \\'text2text-generation\\', \\'api_name\\': \\'facebook/nllb-200-distilled-600M\\', \\'api_call\\': \"pipeline(\\'translation_xx_to_yy\\', model=\\'facebook/nllb-200-distilled-600M\\')\", \\'api_arguments\\': [\\'model\\', \\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\', \\'torch\\'], \\'example_code\\': \"from transformers import pipeline; translator = pipeline(\\'translation_xx_to_yy\\', model=\\'facebook/nllb-200-distilled-600M\\'); translator(\\'Hello World\\')\", \\'performance\\': {\\'dataset\\': \\'Flores-200\\', \\'accuracy\\': \\'BLEU, spBLEU, chrF++\\'}, \\'description\\': \\'NLLB-200 is a machine translation model primarily intended for research in machine translation, especially for low-resource languages. It allows for single sentence translation among 200 languages. The model was trained on general domain text data and is not intended to be used with domain specific texts, such as medical domain or legal domain. The model is not intended to be used for document translation.\\'}', metadata={})]", "category": "generic"}
{"question_id": 503, "text": " I am working for the review section of a book company. I want to convert a book summary into a positive book review.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'pszemraj/long-t5-tglobal-base-16384-book-summary\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'pszemraj/long-t5-tglobal-base-16384-book-summary\\')\", \\'api_arguments\\': [\\'long_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nimport torch\\\\nsummarizer = pipeline(\\\\n summarization,\\\\n pszemraj/long-t5-tglobal-base-16384-book-summary,\\\\n device=0 if torch.cuda.is_available() else -1,\\\\n)\\\\nlong_text = Here is a lot of text I don\\'t want to read. Replace me\\\\nresult = summarizer(long_text)\\\\nprint(result[0][summary_text])\", \\'performance\\': {\\'dataset\\': \\'kmfoda/booksum\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.408, \\'ROUGE-2\\': 6.065, \\'ROUGE-L\\': 16.721, \\'ROUGE-LSUM\\': 33.34}}, \\'description\\': \\'A fine-tuned version of google/long-t5-tglobal-base on the kmfoda/booksum dataset, which can be used to summarize long text and generate SparkNotes-esque summaries of arbitrary topics. The model generalizes reasonably well to academic and narrative text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 504, "text": " We are a company that wants to expand globally. We need to translate our website content from English to Italian.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Translation\\', \\'api_name\\': \\'Helsinki-NLP/opus-mt-it-en\\', \\'api_call\\': \"pipeline(\\'translation_it_to_en\\', model=\\'Helsinki-NLP/opus-mt-it-en\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\', \\'torch\\'], \\'example_code\\': \"from transformers import pipeline\\\\ntranslation = pipeline(\\'translation_it_to_en\\', model=\\'Helsinki-NLP/opus-mt-it-en\\')(\\'Ciao mondo!\\')\", \\'performance\\': {\\'dataset\\': \\'opus\\', \\'accuracy\\': {\\'BLEU\\': {\\'newssyscomb2009.it.en\\': 35.3, \\'newstest2009.it.en\\': 34.0, \\'Tatoeba.it.en\\': 70.9}, \\'chr-F\\': {\\'newssyscomb2009.it.en\\': 0.6, \\'newstest2009.it.en\\': 0.594, \\'Tatoeba.it.en\\': 0.808}}}, \\'description\\': \\'A transformer model for Italian to English translation trained on the OPUS dataset. It can be used for translating Italian text to English.\\'}', metadata={})]", "category": "generic"}
{"question_id": 505, "text": " Our company is working on a news website. We want to present summaries of news articles written in French to the users. \\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'it5-base-news-summarization\\', \\'api_call\\': \"pipeline(\\'summarization\\', model=\\'it5/it5-base-news-summarization\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'newsum(Dal 31 maggio \u00e8 infine partita la piattaforma ITsART, a pi\u00f9 di un anno da quando \u2013 durante il primo lockdown \u2013 il ministro della Cultura Dario Franceschini ne aveva parlato come di \u00abuna sorta di Netflix della cultura\u00bb, pensata per \u00aboffrire a tutto il mondo la cultura italiana a pagamento\u00bb. \u00c8 presto per dare giudizi definitivi sulla piattaforma, e di certo sar\u00e0 difficile farlo anche pi\u00f9 avanti senza numeri precisi. Al momento, l\u2019unica cosa che si pu\u00f2 fare \u00e8 guardare com\u2019\u00e8 fatto il sito, contare quanti contenuti ci sono (circa 700 \u201ctitoli\u201d, tra film, documentari, spettacoli teatrali e musicali e altri eventi) e provare a dare un giudizio sul loro valore e sulla loro variet\u00e0. Intanto, una cosa notata da pi\u00f9 parti \u00e8 che diversi contenuti di ITsART sono a pagamento sulla piattaforma sebbene altrove, per esempio su RaiPlay, siano invece disponibili gratuitamente.)\\', \\'performance\\': {\\'dataset\\': \\'NewsSum-IT\\', \\'accuracy\\': {\\'Rouge1\\': 0.339, \\'Rouge2\\': 0.16, \\'RougeL\\': 0.263}}, \\'description\\': \\'IT5 Base model fine-tuned on news summarization on the Fanpage and Il Post corpora for Italian Language Understanding and Generation.\\'}', metadata={})]", "category": "generic"}
{"question_id": 506, "text": " We are an international business and need to translate our documents from French to Spanish.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Translation\\', \\'api_name\\': \\'Helsinki-NLP/opus-mt-en-fr\\', \\'api_call\\': \"translate(\\'input_text\\', model=\\'Helsinki-NLP/opus-mt-en-fr\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'opus\\', \\'accuracy\\': {\\'BLEU\\': {\\'newsdiscussdev2015-enfr.en.fr\\': 33.8, \\'newsdiscusstest2015-enfr.en.fr\\': 40.0, \\'newssyscomb2009.en.fr\\': 29.8, \\'news-test2008.en.fr\\': 27.5, \\'newstest2009.en.fr\\': 29.4, \\'newstest2010.en.fr\\': 32.7, \\'newstest2011.en.fr\\': 34.3, \\'newstest2012.en.fr\\': 31.8, \\'newstest2013.en.fr\\': 33.2, \\'Tatoeba.en.fr\\': 50.5}}}, \\'description\\': \\'Helsinki-NLP/opus-mt-en-fr is a translation model that translates English text to French using the Hugging Face Transformers library. It is based on the OPUS dataset and uses a transformer-align architecture with normalization and SentencePiece pre-processing.\\'}', metadata={})]", "category": "generic"}
{"question_id": 507, "text": " The company needs to summarize articles for its news application. Provide guidelines to use PEGASUS for this purpose.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'google/pegasus-xsum\\', \\'api_call\\': \"pipeline(\\'summarization\\', model=\\'google/pegasus-xsum\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'xsum\\', \\'accuracy\\': {\\'ROUGE-1\\': 46.862, \\'ROUGE-2\\': 24.453, \\'ROUGE-L\\': 39.055, \\'ROUGE-LSUM\\': 39.099}}, {\\'name\\': \\'cnn_dailymail\\', \\'accuracy\\': {\\'ROUGE-1\\': 22.206, \\'ROUGE-2\\': 7.67, \\'ROUGE-L\\': 15.405, \\'ROUGE-LSUM\\': 19.218}}, {\\'name\\': \\'samsum\\', \\'accuracy\\': {\\'ROUGE-1\\': 21.81, \\'ROUGE-2\\': 4.253, \\'ROUGE-L\\': 17.447, \\'ROUGE-LSUM\\': 18.891}}]}, \\'description\\': \\'PEGASUS is a pre-trained model for abstractive summarization, developed by Google. It is based on the Transformer architecture and trained on both C4 and HugeNews datasets. The model is designed to extract gap sentences and generate summaries by stochastically sampling important sentences.\\'}', metadata={})]", "category": "generic"}
{"question_id": 508, "text": " You are working as a news curator for a media company and need to make a short summary of a long news article.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'it5-base-news-summarization\\', \\'api_call\\': \"pipeline(\\'summarization\\', model=\\'it5/it5-base-news-summarization\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'newsum(Dal 31 maggio \u00e8 infine partita la piattaforma ITsART, a pi\u00f9 di un anno da quando \u2013 durante il primo lockdown \u2013 il ministro della Cultura Dario Franceschini ne aveva parlato come di \u00abuna sorta di Netflix della cultura\u00bb, pensata per \u00aboffrire a tutto il mondo la cultura italiana a pagamento\u00bb. \u00c8 presto per dare giudizi definitivi sulla piattaforma, e di certo sar\u00e0 difficile farlo anche pi\u00f9 avanti senza numeri precisi. Al momento, l\u2019unica cosa che si pu\u00f2 fare \u00e8 guardare com\u2019\u00e8 fatto il sito, contare quanti contenuti ci sono (circa 700 \u201ctitoli\u201d, tra film, documentari, spettacoli teatrali e musicali e altri eventi) e provare a dare un giudizio sul loro valore e sulla loro variet\u00e0. Intanto, una cosa notata da pi\u00f9 parti \u00e8 che diversi contenuti di ITsART sono a pagamento sulla piattaforma sebbene altrove, per esempio su RaiPlay, siano invece disponibili gratuitamente.)\\', \\'performance\\': {\\'dataset\\': \\'NewsSum-IT\\', \\'accuracy\\': {\\'Rouge1\\': 0.339, \\'Rouge2\\': 0.16, \\'RougeL\\': 0.263}}, \\'description\\': \\'IT5 Base model fine-tuned on news summarization on the Fanpage and Il Post corpora for Italian Language Understanding and Generation.\\'}', metadata={})]", "category": "generic"}
{"question_id": 509, "text": " Assemble a conversational bot for my online business to answer questions regarding our products.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'hyunwoongko/blenderbot-9B\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'hyunwoongko/blenderbot-9B\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'Input a message to start chatting with hyunwoongko/blenderbot-9B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high-performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi-turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 510, "text": " We are developing a game and need a text-based AI to control the main character's behavior. Help us use this chatbot based on the Joshua character.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Generation\\', \\'api_name\\': \\'mywateriswet/ShuanBot\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'mywateriswet/ShuanBot\\')\", \\'api_arguments\\': \\'message\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"response = chatbot(\\'What is your name?\\')\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'ShuanBot is a conversational chatbot model based on the GPT-2 architecture. It can be used for generating human-like responses in a chat context.\\'}', metadata={})]", "category": "generic"}
{"question_id": 511, "text": " We are creating an AI assistant for banking clients. The customers should be able to talk to the bot to do various tasks. We need to have a meaningful dialogue with them.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'hyunwoongko/blenderbot-9B\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'hyunwoongko/blenderbot-9B\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'Input a message to start chatting with hyunwoongko/blenderbot-9B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high-performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi-turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 512, "text": " We want to facilitate a conversation with our Russian customers. Build a model that can respond to customer inquiries in Russian.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'hyunwoongko/blenderbot-9B\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'hyunwoongko/blenderbot-9B\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'Input a message to start chatting with hyunwoongko/blenderbot-9B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high-performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi-turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 513, "text": " We are launching a new blog and need a paragraph with tips on how to take care of houseplants.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Paraphrasing\\', \\'api_name\\': \\'prithivida/parrot_paraphraser_on_T5\\', \\'api_call\\': \"Parrot(model_tag=\\'prithivida/parrot_paraphraser_on_T5\\', use_gpu=False)\", \\'api_arguments\\': [\\'input_phrase\\', \\'diversity_ranker\\', \\'do_diverse\\', \\'max_return_phrases\\', \\'max_length\\', \\'adequacy_threshold\\', \\'fluency_threshold\\'], \\'python_environment_requirements\\': [\\'torch\\', \\'transformers\\'], \\'example_code\\': \\'from parrot import Parrot\\\\nimport torch\\\\nimport warnings\\\\nwarnings.filterwarnings(ignore)\\\\n\\\\nparrot = Parrot(model_tag=prithivida/parrot_paraphraser_on_T5, use_gpu=False)\\\\nphrases = [Can you recommed some upscale restaurants in Newyork?,\\\\n What are the famous places we should not miss in Russia?\\\\n]\\\\nfor phrase in phrases:\\\\n print(-*100)\\\\n print(Input_phrase: , phrase)\\\\n print(-*100)\\\\n para_phrases = parrot.augment(input_phrase=phrase)\\\\n for para_phrase in para_phrases:\\\\n  print(para_phrase)\\', \\'performance\\': {\\'dataset\\': \\'Not mentioned\\', \\'accuracy\\': \\'Not mentioned\\'}, \\'description\\': \\'Parrot is a paraphrase based utterance augmentation framework purpose built to accelerate training NLU models. It offers knobs to control Adequacy, Fluency, and Diversity as per your needs. It mainly focuses on augmenting texts typed-into or spoken-to conversational interfaces for building robust NLU models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 514, "text": " The development team is trying to create a function in Python to print \\\"Hello, World!\\\" but they're not sure how to proceed. Generate this function for them.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Paraphrasing\\', \\'api_name\\': \\'prithivida/parrot_paraphraser_on_T5\\', \\'api_call\\': \"Parrot(model_tag=\\'prithivida/parrot_paraphraser_on_T5\\', use_gpu=False)\", \\'api_arguments\\': [\\'input_phrase\\', \\'diversity_ranker\\', \\'do_diverse\\', \\'max_return_phrases\\', \\'max_length\\', \\'adequacy_threshold\\', \\'fluency_threshold\\'], \\'python_environment_requirements\\': [\\'torch\\', \\'transformers\\'], \\'example_code\\': \\'from parrot import Parrot\\\\nimport torch\\\\nimport warnings\\\\nwarnings.filterwarnings(ignore)\\\\n\\\\nparrot = Parrot(model_tag=prithivida/parrot_paraphraser_on_T5, use_gpu=False)\\\\nphrases = [Can you recommed some upscale restaurants in Newyork?,\\\\n What are the famous places we should not miss in Russia?\\\\n]\\\\nfor phrase in phrases:\\\\n print(-*100)\\\\n print(Input_phrase: , phrase)\\\\n print(-*100)\\\\n para_phrases = parrot.augment(input_phrase=phrase)\\\\n for para_phrase in para_phrases:\\\\n  print(para_phrase)\\', \\'performance\\': {\\'dataset\\': \\'Not mentioned\\', \\'accuracy\\': \\'Not mentioned\\'}, \\'description\\': \\'Parrot is a paraphrase based utterance augmentation framework purpose built to accelerate training NLU models. It offers knobs to control Adequacy, Fluency, and Diversity as per your needs. It mainly focuses on augmenting texts typed-into or spoken-to conversational interfaces for building robust NLU models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 515, "text": " We need to generate a motivational quote related to sports.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'text2text-generation\\', \\'api_name\\': \\'tuner007/pegasus_summarizer\\', \\'api_call\\': \"PegasusForConditionalGeneration.from_pretrained(\\'tuner007/pegasus_summarizer\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'pip install sentencepiece\\'], \\'example_code\\': \"context = \\\\nIndia wicket-keeper batsman Rishabh Pant has said someone from the crowd threw a ball on pacer Mohammed Siraj while he was fielding in the ongoing third Test against England on Wednesday. Pant revealed the incident made India skipper Virat Kohli upset. I think, somebody threw a ball inside, at Siraj, so he [Kohli] was upset, said Pant in a virtual press conference after the close of the first day\\'s play.You can say whatever you want to chant, but don\\'t throw things at the fielders and all those things. It is not good for cricket, I guess, he added.In the third session of the opening day of the third Test, a section of spectators seemed to have asked Siraj the score of the match to tease the pacer. The India pacer however came with a brilliant reply as he gestured 1-0 (India leading the Test series) towards the crowd.Earlier this month, during the second Test match, there was some bad crowd behaviour on a show as some unruly fans threw champagne corks at India batsman KL Rahul.Kohli also intervened and he was seen gesturing towards the opening batsman to know more about the incident. An over later, the TV visuals showed that many champagne corks were thrown inside the playing field, and the Indian players were visibly left frustrated.Coming back to the game, after bundling out India for 78, openers Rory Burns and Haseeb Hameed ensured that England took the honours on the opening day of the ongoing third Test.At stumps, England\\'s score reads 120/0 and the hosts have extended their lead to 42 runs. For the Three Lions, Burns (52) and Hameed (60) are currently unbeaten at the crease.Talking about the pitch on opening day, Pant said, They took the heavy roller, the wicket was much more settled down, and they batted nicely also, he said. But when we batted, the wicket was slightly soft, and they bowled in good areas, but we could have applied [ourselves] much better.Both England batsmen managed to see off the final session and the hosts concluded the opening day with all ten wickets intact, extending the lead to 42.(ANI)\\\\n\\\\nget_response(context)\", \\'performance\\': {\\'dataset\\': \\'cnn_dailymail\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.604, \\'ROUGE-2\\': 14.64, \\'ROUGE-L\\': 23.884, \\'ROUGE-LSUM\\': 32.902, \\'loss\\': 2.576, \\'gen_len\\': 76.398}}, \\'description\\': \\'PEGASUS fine-tuned for summarization\\'}', metadata={})]", "category": "generic"}
{"question_id": 516, "text": " Give me a tool that I can use to generate stories based on a starting phrase.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Paraphrasing\\', \\'api_name\\': \\'prithivida/parrot_paraphraser_on_T5\\', \\'api_call\\': \"Parrot(model_tag=\\'prithivida/parrot_paraphraser_on_T5\\', use_gpu=False)\", \\'api_arguments\\': [\\'input_phrase\\', \\'diversity_ranker\\', \\'do_diverse\\', \\'max_return_phrases\\', \\'max_length\\', \\'adequacy_threshold\\', \\'fluency_threshold\\'], \\'python_environment_requirements\\': [\\'torch\\', \\'transformers\\'], \\'example_code\\': \\'from parrot import Parrot\\\\nimport torch\\\\nimport warnings\\\\nwarnings.filterwarnings(ignore)\\\\n\\\\nparrot = Parrot(model_tag=prithivida/parrot_paraphraser_on_T5, use_gpu=False)\\\\nphrases = [Can you recommed some upscale restaurants in Newyork?,\\\\n What are the famous places we should not miss in Russia?\\\\n]\\\\nfor phrase in phrases:\\\\n print(-*100)\\\\n print(Input_phrase: , phrase)\\\\n print(-*100)\\\\n para_phrases = parrot.augment(input_phrase=phrase)\\\\n for para_phrase in para_phrases:\\\\n  print(para_phrase)\\', \\'performance\\': {\\'dataset\\': \\'Not mentioned\\', \\'accuracy\\': \\'Not mentioned\\'}, \\'description\\': \\'Parrot is a paraphrase based utterance augmentation framework purpose built to accelerate training NLU models. It offers knobs to control Adequacy, Fluency, and Diversity as per your needs. It mainly focuses on augmenting texts typed-into or spoken-to conversational interfaces for building robust NLU models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 517, "text": " Our PR team requires a machine learning approach to creative sentence generation for marketing content creation.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Paraphrasing\\', \\'api_name\\': \\'prithivida/parrot_paraphraser_on_T5\\', \\'api_call\\': \"Parrot(model_tag=\\'prithivida/parrot_paraphraser_on_T5\\', use_gpu=False)\", \\'api_arguments\\': [\\'input_phrase\\', \\'diversity_ranker\\', \\'do_diverse\\', \\'max_return_phrases\\', \\'max_length\\', \\'adequacy_threshold\\', \\'fluency_threshold\\'], \\'python_environment_requirements\\': [\\'torch\\', \\'transformers\\'], \\'example_code\\': \\'from parrot import Parrot\\\\nimport torch\\\\nimport warnings\\\\nwarnings.filterwarnings(ignore)\\\\n\\\\nparrot = Parrot(model_tag=prithivida/parrot_paraphraser_on_T5, use_gpu=False)\\\\nphrases = [Can you recommed some upscale restaurants in Newyork?,\\\\n What are the famous places we should not miss in Russia?\\\\n]\\\\nfor phrase in phrases:\\\\n print(-*100)\\\\n print(Input_phrase: , phrase)\\\\n print(-*100)\\\\n para_phrases = parrot.augment(input_phrase=phrase)\\\\n for para_phrase in para_phrases:\\\\n  print(para_phrase)\\', \\'performance\\': {\\'dataset\\': \\'Not mentioned\\', \\'accuracy\\': \\'Not mentioned\\'}, \\'description\\': \\'Parrot is a paraphrase based utterance augmentation framework purpose built to accelerate training NLU models. It offers knobs to control Adequacy, Fluency, and Diversity as per your needs. It mainly focuses on augmenting texts typed-into or spoken-to conversational interfaces for building robust NLU models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 518, "text": " We are developing an AI-powered code review system. Our model should provide a short summary of the provided code snippet.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'pszemraj/long-t5-tglobal-base-16384-book-summary\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'pszemraj/long-t5-tglobal-base-16384-book-summary\\')\", \\'api_arguments\\': [\\'long_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nimport torch\\\\nsummarizer = pipeline(\\\\n summarization,\\\\n pszemraj/long-t5-tglobal-base-16384-book-summary,\\\\n device=0 if torch.cuda.is_available() else -1,\\\\n)\\\\nlong_text = Here is a lot of text I don\\'t want to read. Replace me\\\\nresult = summarizer(long_text)\\\\nprint(result[0][summary_text])\", \\'performance\\': {\\'dataset\\': \\'kmfoda/booksum\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.408, \\'ROUGE-2\\': 6.065, \\'ROUGE-L\\': 16.721, \\'ROUGE-LSUM\\': 33.34}}, \\'description\\': \\'A fine-tuned version of google/long-t5-tglobal-base on the kmfoda/booksum dataset, which can be used to summarize long text and generate SparkNotes-esque summaries of arbitrary topics. The model generalizes reasonably well to academic and narrative text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 519, "text": " I want a language model that can fill in short blanks for example sentences, quizzes, or trivia questions.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Fill-Mask\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Masked Language Modeling\\', \\'api_name\\': \\'distilbert-base-multilingual-cased\\', \\'api_call\\': \"pipeline(\\'fill-mask\\', model=\\'distilbert-base-multilingual-cased\\')\", \\'api_arguments\\': [\\'pipeline\\', \\'fill-mask\\', \\'model\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nunmasker = pipeline(\\'fill-mask\\', model=\\'distilbert-base-multilingual-cased\\')\\\\nunmasker(Hello I\\'m a [MASK] model.)\", \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'XNLI\\', \\'accuracy\\': {\\'English\\': 78.2, \\'Spanish\\': 69.1, \\'Chinese\\': 64.0, \\'German\\': 66.3, \\'Arabic\\': 59.1, \\'Urdu\\': 54.7}}]}, \\'description\\': \\'This model is a distilled version of the BERT base multilingual model. It is trained on the concatenation of Wikipedia in 104 different languages. The model has 6 layers, 768 dimension and 12 heads, totalizing 134M parameters. On average, this model, referred to as DistilmBERT, is twice as fast as mBERT-base.\\'}', metadata={})]", "category": "generic"}
{"question_id": 520, "text": " As a language teacher, I sometimes need help completing Dutch sentences with the most suitable word. Can you fill in the gaps?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Fill-Mask\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Fill-Mask\\', \\'api_name\\': \\'GroNLP/bert-base-dutch-cased\\', \\'api_call\\': \"AutoModel.from_pretrained(\\'GroNLP/bert-base-dutch-cased\\')\", \\'api_arguments\\': [\\'pretrained_model_name_or_path\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'from transformers import AutoTokenizer, AutoModel, TFAutoModel\\\\ntokenizer = AutoTokenizer.from_pretrained(GroNLP/bert-base-dutch-cased)\\\\nmodel = AutoModel.from_pretrained(GroNLP/bert-base-dutch-cased)\\', \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'CoNLL-2002\\', \\'accuracy\\': \\'90.24\\'}, {\\'name\\': \\'SoNaR-1\\', \\'accuracy\\': \\'84.93\\'}, {\\'name\\': \\'spaCy UD LassySmall\\', \\'accuracy\\': \\'86.10\\'}]}, \\'description\\': \\'BERTje is a Dutch pre-trained BERT model developed at the University of Groningen.\\'}', metadata={})]", "category": "generic"}
{"question_id": 521, "text": " Translate some lyrics of a Portuguese song into English\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Speech Recognition\\', \\'api_name\\': \\'jonatasgrosman/wav2vec2-large-xlsr-53-portuguese\\', \\'api_call\\': \"SpeechRecognitionModel(\\'jonatasgrosman/wav2vec2-large-xlsr-53-portuguese\\')\", \\'api_arguments\\': [\\'audio_paths\\'], \\'python_environment_requirements\\': [\\'huggingsound\\', \\'torch\\', \\'librosa\\', \\'datasets\\', \\'transformers\\'], \\'example_code\\': \\'from huggingsound import SpeechRecognitionModel\\\\nmodel = SpeechRecognitionModel(jonatasgrosman/wav2vec2-large-xlsr-53-portuguese)\\\\naudio_paths = [/path/to/file.mp3, /path/to/another_file.wav]\\\\ntranscriptions = model.transcribe(audio_paths)\\', \\'performance\\': {\\'dataset\\': \\'mozilla-foundation/common_voice_6_0\\', \\'accuracy\\': {\\'Test WER\\': 11.31, \\'Test CER\\': 3.74, \\'Test WER (+LM)\\': 9.01, \\'Test CER (+LM)\\': 3.21}}, \\'description\\': \\'Fine-tuned facebook/wav2vec2-large-xlsr-53 on Portuguese using the train and validation splits of Common Voice 6.1. When using this model, make sure that your speech input is sampled at 16kHz.\\'}', metadata={})]", "category": "generic"}
{"question_id": 522, "text": " We are a news portal website and recently acquired a breaking news article. Determine the similarity between the main text of this article and other articles in our database.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'it5-base-news-summarization\\', \\'api_call\\': \"pipeline(\\'summarization\\', model=\\'it5/it5-base-news-summarization\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'newsum(Dal 31 maggio \u00e8 infine partita la piattaforma ITsART, a pi\u00f9 di un anno da quando \u2013 durante il primo lockdown \u2013 il ministro della Cultura Dario Franceschini ne aveva parlato come di \u00abuna sorta di Netflix della cultura\u00bb, pensata per \u00aboffrire a tutto il mondo la cultura italiana a pagamento\u00bb. \u00c8 presto per dare giudizi definitivi sulla piattaforma, e di certo sar\u00e0 difficile farlo anche pi\u00f9 avanti senza numeri precisi. Al momento, l\u2019unica cosa che si pu\u00f2 fare \u00e8 guardare com\u2019\u00e8 fatto il sito, contare quanti contenuti ci sono (circa 700 \u201ctitoli\u201d, tra film, documentari, spettacoli teatrali e musicali e altri eventi) e provare a dare un giudizio sul loro valore e sulla loro variet\u00e0. Intanto, una cosa notata da pi\u00f9 parti \u00e8 che diversi contenuti di ITsART sono a pagamento sulla piattaforma sebbene altrove, per esempio su RaiPlay, siano invece disponibili gratuitamente.)\\', \\'performance\\': {\\'dataset\\': \\'NewsSum-IT\\', \\'accuracy\\': {\\'Rouge1\\': 0.339, \\'Rouge2\\': 0.16, \\'RougeL\\': 0.263}}, \\'description\\': \\'IT5 Base model fine-tuned on news summarization on the Fanpage and Il Post corpora for Italian Language Understanding and Generation.\\'}', metadata={})]", "category": "generic"}
{"question_id": 523, "text": " As a business assistant of an international company, find the most relevant sentence among a list of sentences that answers a specific question.\\n###Input: {\\\"question\\\": \\\"What is the main purpose of photosynthesis?\\\", \\\"sentences\\\": [\\\"Photosynthesis is the process used by plants to convert light energy into chemical energy to fuel their growth.\\\", \\\"The Eiffel Tower is a famous landmark in Paris.\\\", \\\"Photosynthesis also produces oxygen as a byproduct, which is necessary for life on Earth.\\\"]}\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentence Correction\\', \\'api_name\\': \\'flexudy/t5-base-multi-sentence-doctor\\', \\'api_call\\': \"AutoModelWithLMHead.from_pretrained(\\'flexudy/t5-base-multi-sentence-doctor\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'from transformers import AutoTokenizer, AutoModelWithLMHead\\\\ntokenizer = AutoTokenizer.from_pretrained(flexudy/t5-base-multi-sentence-doctor)\\\\nmodel = AutoModelWithLMHead.from_pretrained(flexudy/t5-base-multi-sentence-doctor)\\\\ninput_text = repair_sentence: m a medical doct context: {That is my job I a}{or I save lives} </s>\\\\ninput_ids = tokenizer.encode(input_text, return_tensors=pt)\\\\noutputs = model.generate(input_ids, max_length=32, num_beams=1)\\\\nsentence = tokenizer.decode(outputs[0], skip_special_tokens=True, clean_up_tokenization_spaces=True)\\\\nassert sentence == I am a medical doctor.\\', \\'performance\\': {\\'dataset\\': \\'tatoeba\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'Sentence doctor is a T5 model that attempts to correct the errors or mistakes found in sentences. Model works on English, German and French text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 524, "text": " We need to cluster customer reviews based on their content similarity to understand the common issues customers are facing.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'michellejieli/NSFW_text_classifier\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'michellejieli/NSFW_text_classification\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'classifier(I see you\u2019ve set aside this special time to humiliate yourself in public.)\\', \\'performance\\': {\\'dataset\\': \\'Reddit posts\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'DistilBERT is a transformer model that performs sentiment analysis. I fine-tuned the model on Reddit posts with the purpose of classifying not safe for work (NSFW) content, specifically text that is considered inappropriate and unprofessional. The model predicts 2 classes, which are NSFW or safe for work (SFW). The model is a fine-tuned version of DistilBERT. It was fine-tuned on 14317 Reddit posts pulled from the Reddit API.\\'}', metadata={})]", "category": "generic"}
{"question_id": 525, "text": " Create a program to calculate sentence similarity scores between a list of sentences.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Sentence Similarity\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Sentence Transformers\\', \\'api_name\\': \\'sentence-transformers/distiluse-base-multilingual-cased-v2\\', \\'api_call\\': \"SentenceTransformer(\\'sentence-transformers/distiluse-base-multilingual-cased-v2\\')\", \\'api_arguments\\': [\\'sentences\\'], \\'python_environment_requirements\\': \\'pip install -U sentence-transformers\\', \\'example_code\\': \"from sentence_transformers import SentenceTransformer\\\\nsentences = [This is an example sentence, Each sentence is converted]\\\\nmodel = SentenceTransformer(\\'sentence-transformers/distiluse-base-multilingual-cased-v2\\')\\\\nembeddings = model.encode(sentences)\\\\nprint(embeddings)\", \\'performance\\': {\\'dataset\\': \\'https://seb.sbert.net\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This is a sentence-transformers model: It maps sentences & paragraphs to a 512 dimensional dense vector space and can be used for tasks like clustering or semantic search.\\'}', metadata={})]", "category": "generic"}
{"question_id": 526, "text": " We need to find a similar Chinese sentence to my source sentence for a suggestion.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Sentence Similarity\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'text2vec-large-chinese\\', \\'api_call\\': \"AutoModel.from_pretrained(\\'GanymedeNil/text2vec-large-chinese\\')\", \\'api_arguments\\': \\'source_sentence, sentences_to_compare\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"from transformers import AutoModel, AutoTokenizer\\\\n\\\\ntokenizer = AutoTokenizer.from_pretrained(\\'GanymedeNil/text2vec-large-chinese\\')\\\\nmodel = AutoModel.from_pretrained(\\'GanymedeNil/text2vec-large-chinese\\')\", \\'performance\\': {\\'dataset\\': \\'https://huggingface.co/shibing624/text2vec-base-chinese\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'A Chinese sentence similarity model based on the derivative model of https://huggingface.co/shibing624/text2vec-base-chinese, replacing MacBERT with LERT, and keeping other training conditions unchanged.\\'}', metadata={})]", "category": "generic"}
{"question_id": 527, "text": " I am building a voice assistant for my mobile app. Give me an example of how I can implement this Text-to-Speech model.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'mio/Artoria\\', \\'api_call\\': \"pipeline(\\'text-to-speech\\', model=\\'mio/Artoria\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline; tts = pipeline(\\'text-to-speech\\', model=\\'mio/Artoria\\'); tts(\\'s\\')\", \\'performance\\': {\\'dataset\\': \\'fate\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This model was trained by mio using fate recipe in espnet. It is a text-to-speech model that can convert text input into speech output.\\'}', metadata={})]", "category": "generic"}
{"question_id": 528, "text": " Our team wants to transcribe audio files into text including punctuation marks for usability studies.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Audio-to-Audio\\', \\'framework\\': \\'Fairseq\\', \\'functionality\\': \\'speech-to-speech-translation\\', \\'api_name\\': \\'facebook/textless_sm_ro_en\\', \\'api_call\\': \"pipeline(\\'audio-to-audio\\', model=\\'facebook/textless_sm_ro_en\\')\", \\'api_arguments\\': \\'audio file or recording\\', \\'python_environment_requirements\\': \\'fairseq, huggingface_hub\\', \\'example_code\\': \\'https://huggingface.co/facebook/textless_sm_cs_en\\', \\'performance\\': {\\'dataset\\': \\'unknown\\', \\'accuracy\\': \\'unknown\\'}, \\'description\\': \\'A speech-to-speech translation model for Romanian to English developed by Facebook AI\\'}', metadata={})]", "category": "generic"}
{"question_id": 529, "text": " Your client is a podcast aggregator website looking to transcribe episodes and display transcriptions on the site. Help them transcribe an audio file containing spoken English words.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'CTranslate2\\', \\'functionality\\': \\'Automatic Speech Recognition\\', \\'api_name\\': \\'guillaumekln/faster-whisper-large-v2\\', \\'api_call\\': \"WhisperModel(\\'large-v2\\')\", \\'api_arguments\\': [\\'audio.mp3\\'], \\'python_environment_requirements\\': [\\'faster_whisper\\'], \\'example_code\\': \\'from faster_whisper import WhisperModel\\\\nmodel = WhisperModel(large-v2)\\\\nsegments, info = model.transcribe(audio.mp3)\\\\nfor segment in segments:\\\\n print([%.2fs -&gt; %.2fs] %s % (segment.start, segment.end, segment.text))\\', \\'performance\\': {\\'dataset\\': \\'99 languages\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Whisper large-v2 model for CTranslate2. This model can be used in CTranslate2 or projets based on CTranslate2 such as faster-whisper.\\'}', metadata={})]", "category": "generic"}
{"question_id": 530, "text": " Our app offers assistance to people with hearing problems by enhancing the clarity of speech. We need a feature to clean and enhance the audio.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'CTranslate2\\', \\'functionality\\': \\'Automatic Speech Recognition\\', \\'api_name\\': \\'guillaumekln/faster-whisper-large-v2\\', \\'api_call\\': \"WhisperModel(\\'large-v2\\')\", \\'api_arguments\\': [\\'audio.mp3\\'], \\'python_environment_requirements\\': [\\'faster_whisper\\'], \\'example_code\\': \\'from faster_whisper import WhisperModel\\\\nmodel = WhisperModel(large-v2)\\\\nsegments, info = model.transcribe(audio.mp3)\\\\nfor segment in segments:\\\\n print([%.2fs -&gt; %.2fs] %s % (segment.start, segment.end, segment.text))\\', \\'performance\\': {\\'dataset\\': \\'99 languages\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Whisper large-v2 model for CTranslate2. This model can be used in CTranslate2 or projets based on CTranslate2 such as faster-whisper.\\'}', metadata={})]", "category": "generic"}
{"question_id": 531, "text": " Find a way to separate the background music and vocal from an audio file.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Audio-to-Audio\\', \\'framework\\': \\'SpeechBrain\\', \\'functionality\\': \\'Audio Source Separation\\', \\'api_name\\': \\'sepformer-wsj02mix\\', \\'api_call\\': \"separator.from_hparams(source=\\'speechbrain/sepformer-wsj02mix\\')\", \\'api_arguments\\': [\\'path\\'], \\'python_environment_requirements\\': [\\'pip install speechbrain\\'], \\'example_code\\': [\\'from speechbrain.pretrained import SepformerSeparation as separator\\', \\'import torchaudio\\', \"model = separator.from_hparams(source=speechbrain/sepformer-wsj02mix, savedir=\\'pretrained_models/sepformer-wsj02mix\\')\", \"est_sources = model.separate_file(path=\\'speechbrain/sepformer-wsj02mix/test_mixture.wav\\')\", \\'torchaudio.save(source1hat.wav, est_sources[:, :, 0].detach().cpu(), 8000)\\', \\'torchaudio.save(source2hat.wav, est_sources[:, :, 1].detach().cpu(), 8000)\\'], \\'performance\\': {\\'dataset\\': \\'WSJ0-2Mix\\', \\'accuracy\\': \\'22.4 dB\\'}, \\'description\\': \\'This repository provides all the necessary tools to perform audio source separation with a SepFormer model, implemented with SpeechBrain, and pretrained on WSJ0-2Mix dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 532, "text": " An educational software company needs an advanced method to separate voices from background noise. Develop a solution.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Voice Activity Detection\\', \\'api_name\\': \\'d4data/Indian-voice-cloning\\', \\'api_call\\': \"pipeline(\\'voice-activity-detection\\', model=\\'d4data/Indian-voice-cloning\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A model for detecting voice activity in Indian languages.\\'}', metadata={})]", "category": "generic"}
{"question_id": 533, "text": " Develop a voice command security system that distinguishes between specific command phrases like \\\"disarm security\\\" or \\\"activate alarm\\\".\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Voice Activity Detection\\', \\'api_name\\': \\'d4data/Indian-voice-cloning\\', \\'api_call\\': \"pipeline(\\'voice-activity-detection\\', model=\\'d4data/Indian-voice-cloning\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A model for detecting voice activity in Indian languages.\\'}', metadata={})]", "category": "generic"}
{"question_id": 534, "text": " I have an audio recording that I want to analyze to understand the emotion of the speaker.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Audio Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'padmalcom/wav2vec2-large-emotion-detection-german\\', \\'api_call\\': \"pipeline(\\'audio-classification\\', model=\\'padmalcom/wav2vec2-large-emotion-detection-german\\')\", \\'api_arguments\\': \\'audio_file\\', \\'python_environment_requirements\\': \\'transformers, torch\\', \\'example_code\\': \"from transformers import pipeline\\\\naudio_classifier = pipeline(\\'audio-classification\\', model=\\'padmalcom/wav2vec2-large-emotion-detection-german\\')\\\\nresult = audio_classifier(audio_file)\", \\'performance\\': {\\'dataset\\': \\'emo-DB\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This wav2vec2 based emotion detection model is trained on the emo-DB dataset. It can classify emotions in German audio files into seven classes: anger, boredom, disgust, fear, happiness, sadness, and neutral.\\'}', metadata={})]", "category": "generic"}
{"question_id": 535, "text": " We are working on an assistant that can help tourists translate spoken language from one language to another. Let's build a solution for Hokkien to English translation.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'Fairseq\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'unit_hifigan_HK_layer12.km2500_frame_TAT-TTS\\', \\'api_call\\': \"load_model_ensemble_and_task_from_hf_hub(\\'facebook/unit_hifigan_HK_layer12.km2500_frame_TAT-TT\\')\", \\'api_arguments\\': {\\'unit\\': \\'Text input for the TTS model\\'}, \\'python_environment_requirements\\': [\\'fairseq\\', \\'huggingface_hub\\', \\'torchaudio\\'], \\'example_code\\': \"import json\\\\nimport os\\\\nfrom pathlib import Path\\\\nimport IPython.display as ipd\\\\nfrom fairseq import hub_utils\\\\nfrom fairseq.checkpoint_utils import load_model_ensemble_and_task_from_hf_hub\\\\nfrom fairseq.models.speech_to_text.hub_interface import S2THubInterface\\\\nfrom fairseq.models.text_to_speech import CodeHiFiGANVocoder\\\\nfrom fairseq.models.text_to_speech.hub_interface import VocoderHubInterface\\\\nfrom huggingface_hub import snapshot_download\\\\nimport torchaudio\\\\ncache_dir = os.getenv(HUGGINGFACE_HUB_CACHE)\\\\nlibrary_name = fairseq\\\\ncache_dir = (\\\\n cache_dir or (Path.home() / .cache / library_name).as_posix()\\\\n)\\\\ncache_dir = snapshot_download(\\\\n ffacebook/unit_hifigan_HK_layer12.km2500_frame_TAT-TTS, cache_dir=cache_dir, library_name=library_name\\\\n)\\\\nx = hub_utils.from_pretrained(\\\\n cache_dir,\\\\n model.pt,\\\\n .,\\\\n archive_map=CodeHiFiGANVocoder.hub_models(),\\\\n config_yaml=config.json,\\\\n fp16=False,\\\\n is_vocoder=True,\\\\n)\\\\nwith open(f{x[\\'args\\'][\\'data\\']}/config.json) as f:\\\\n vocoder_cfg = json.load(f)\\\\nassert (\\\\n len(x[args][model_path]) == 1\\\\n), Too many vocoder models in the input\\\\nvocoder = CodeHiFiGANVocoder(x[args][model_path][0], vocoder_cfg)\\\\ntts_model = VocoderHubInterface(vocoder_cfg, vocoder)\\\\ntts_sample = tts_model.get_model_input(unit)\\\\nwav, sr = tts_model.get_prediction(tts_sample)\\\\nipd.Audio(wav, rate=sr)\", \\'performance\\': {\\'dataset\\': \\'TAT-TTS\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Hokkien unit HiFiGAN based vocoder from fairseq. Trained with TAT-TTS data with 4 speakers in Taiwanese Hokkien accent.\\'}', metadata={})]", "category": "generic"}
{"question_id": 536, "text": " Trying to implement an assistant that detects user's emotions on a given audio file to help them manage their emotions better.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'Rajaram1996/Hubert_emotion\\', \\'api_call\\': \"HubertForSpeechClassification.from_pretrained(\\'Rajaram1996/Hubert_emotion\\')\", \\'api_arguments\\': {\\'audio_file\\': \\'string\\'}, \\'python_environment_requirements\\': [\\'audio_models\\', \\'transformers\\', \\'torch\\', \\'numpy\\', \\'pydub\\'], \\'example_code\\': \"def predict_emotion_hubert(audio_file):\\\\n from audio_models import HubertForSpeechClassification\\\\n from transformers import Wav2Vec2FeatureExtractor, AutoConfig\\\\n import torch.nn.functional as F\\\\n import torch\\\\n import numpy as np\\\\n from pydub import AudioSegment\\\\nmodel = HubertForSpeechClassification.from_pretrained(Rajaram1996/Hubert_emotion)\\\\nfeature_extractor = Wav2Vec2FeatureExtractor.from_pretrained(facebook/hubert-base-ls960)\\\\nsampling_rate=16000\\\\nconfig = AutoConfig.from_pretrained(Rajaram1996/Hubert_emotion)\\\\ndef speech_file_to_array(path, sampling_rate):\\\\n sound = AudioSegment.from_file(path)\\\\n sound = sound.set_frame_rate(sampling_rate)\\\\n sound_array = np.array(sound.get_array_of_samples())\\\\n return sound_array\\\\nsound_array = speech_file_to_array(audio_file, sampling_rate)\\\\ninputs = feature_extractor(sound_array, sampling_rate=sampling_rate, return_tensors=pt, padding=True)\\\\ninputs = {key: inputs[key].to(cpu).float() for key in inputs}\\\\nwith torch.no_grad():\\\\n logits = model(**inputs).logits\\\\nscores = F.softmax(logits, dim=1).detach().cpu().numpy()[0]\\\\noutputs = [{\\\\n emo: config.id2label[i],\\\\n score: round(score * 100, 1)}\\\\n for i, score in enumerate(scores)\\\\n]\\\\nreturn [row for row in sorted(outputs, key=lambda x:x[score], reverse=True) if row[\\'score\\'] != \\'0.0%\\'][:2]\\\\nresult = predict_emotion_hubert(male-crying.mp3)\\\\nresult\", \\'performance\\': {\\'dataset\\': \\'unknown\\', \\'accuracy\\': \\'unknown\\'}, \\'description\\': \\'A pretrained model for predicting emotion in local audio files using Hubert.\\'}', metadata={})]", "category": "generic"}
{"question_id": 537, "text": " We are building a customer service analytics software. The software should recognize the voice of the customers.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Voice Activity Detection\\', \\'api_name\\': \\'d4data/Indian-voice-cloning\\', \\'api_call\\': \"pipeline(\\'voice-activity-detection\\', model=\\'d4data/Indian-voice-cloning\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A model for detecting voice activity in Indian languages.\\'}', metadata={})]", "category": "generic"}
{"question_id": 538, "text": " We are developing a voice-controlled device. Help categorize spoken commands into specific keywords to trigger various actions.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Voice Activity Detection\\', \\'api_name\\': \\'d4data/Indian-voice-cloning\\', \\'api_call\\': \"pipeline(\\'voice-activity-detection\\', model=\\'d4data/Indian-voice-cloning\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A model for detecting voice activity in Indian languages.\\'}', metadata={})]", "category": "generic"}
{"question_id": 539, "text": " Develop an emotion analysis system to understand customer satisfaction over the phone for a telecommunication company in Russia.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Emotion Classification\\', \\'api_name\\': \\'j-hartmann/emotion-english-distilroberta-base\\', \\'api_call\\': \"pipeline(\\'text-classification\\', model=\\'j-hartmann/emotion-english-distilroberta-base\\', return_all_scores=True)\", \\'api_arguments\\': {\\'text\\': \\'string\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'latest\\'}, \\'example_code\\': \\'from transformers import pipeline\\\\nclassifier = pipeline(text-classification, model=j-hartmann/emotion-english-distilroberta-base, return_all_scores=True)\\\\nclassifier(I love this!)\\', \\'performance\\': {\\'dataset\\': \\'Balanced subset from 6 diverse datasets\\', \\'accuracy\\': \\'66%\\'}, \\'description\\': \"This model classifies emotions in English text data. It predicts Ekman\\'s 6 basic emotions, plus a neutral class: anger, disgust, fear, joy, neutral, sadness, and surprise. The model is a fine-tuned checkpoint of DistilRoBERTa-base.\"}', metadata={})]", "category": "generic"}
{"question_id": 540, "text": " We are a company that provides transcription services. We require voice activity detection in our audio recordings.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Voice Activity Detection\\', \\'api_name\\': \\'d4data/Indian-voice-cloning\\', \\'api_call\\': \"pipeline(\\'voice-activity-detection\\', model=\\'d4data/Indian-voice-cloning\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A model for detecting voice activity in Indian languages.\\'}', metadata={})]", "category": "generic"}
{"question_id": 541, "text": " We need to predict the survival of passengers on the Titanic based on certain demographics like age, gender, etc.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Binary Classification\\', \\'api_name\\': \\'harithapliyal/autotrain-tatanic-survival-51030121311\\', \\'api_call\\': \"AutoModel.from_pretrained(\\'harithapliyal/autotrain-tatanic-survival-51030121311\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'harithapliyal/autotrain-data-tatanic-survival\\', \\'accuracy\\': 0.872}, \\'description\\': \\'A tabular classification model trained on the Titanic survival dataset using Hugging Face AutoTrain. The model predicts whether a passenger survived or not based on features such as age, gender, and passenger class.\\'}', metadata={})]", "category": "generic"}
{"question_id": 542, "text": " Implement a system that predicts the carbon emissions in a given dataset.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'kochetkovIT/autotrain-ironhack-49741119788\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'json\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'kochetkovIT/autotrain-data-ironhack\\', \\'accuracy\\': {\\'Loss\\': 2.603, \\'R2\\': 0.013, \\'MSE\\': 6.776, \\'MAE\\': 1.666, \\'RMSLE\\': 0.502}}, \\'description\\': \\'A tabular regression model trained using AutoTrain to predict carbon emissions (in grams) with an R2 score of 0.013.\\'}', metadata={})]", "category": "generic"}
{"question_id": 543, "text": " Predict the carbon emissions of several power plants based on their characteristics.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Joblib\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'pachi107/autotrain-in-class-test-1780161764\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data.csv\\'], \\'python_environment_requirements\\': [\\'json\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(data.csv)\\\\ndata = data[features]\\\\ndata.columns = [feat_ + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'Validation Metrics\\', \\'accuracy\\': 0.974}, \\'description\\': \\'A binary classification model for predicting CO2 emissions based on tabular data. Trained using AutoTrain with a model ID of 1780161764.\\'}', metadata={})]", "category": "generic"}
{"question_id": 544, "text": " Our robotics team needs a way to stabilize a two-wheeled self-balancing robot. Would you suggest and apply any RL algorithm for this purpose?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Reinforcement Learning\\', \\'framework\\': \\'Stable-Baselines3\\', \\'functionality\\': \\'LunarLander-v2\\', \\'api_name\\': \\'araffin/dqn-LunarLander-v2\\', \\'api_call\\': \"DQN.load(load_from_hub(\\'araffin/dqn-LunarLander-v2\\', \\'dqn-LunarLander-v2.zip\\'), **kwargs)\", \\'api_arguments\\': {\\'checkpoint\\': \\'araffin/dqn-LunarLander-v2\\', \\'kwargs\\': {\\'target_update_interval\\': 30}}, \\'python_environment_requirements\\': [\\'huggingface_sb3\\', \\'stable_baselines3\\'], \\'example_code\\': {\\'load_model\\': \\'from huggingface_sb3 import load_from_hub\\\\nfrom stable_baselines3 import DQN\\\\nfrom stable_baselines3.common.env_util import make_vec_env\\\\nfrom stable_baselines3.common.evaluation import evaluate_policy\\\\n\\\\ncheckpoint = load_from_hub(araffin/dqn-LunarLander-v2, dqn-LunarLander-v2.zip)\\\\n\\\\nkwargs = dict(target_update_interval=30)\\\\n\\\\nmodel = DQN.load(checkpoint, **kwargs)\\\\nenv = make_vec_env(LunarLander-v2, n_envs=1)\\', \\'evaluate\\': \\'mean_reward, std_reward = evaluate_policy(\\\\n model,\\\\n env,\\\\n n_eval_episodes=20,\\\\n deterministic=True,\\\\n)\\\\nprint(fMean reward = {mean_reward:.2f} +/- {std_reward:.2f})\\'}, \\'performance\\': {\\'dataset\\': \\'LunarLander-v2\\', \\'accuracy\\': \\'280.22 +/- 13.03\\'}, \\'description\\': \\'This is a trained model of a DQN agent playing LunarLander-v2 using the stable-baselines3 library.\\'}', metadata={})]", "category": "generic"}
{"question_id": 545, "text": " We have a large text dataset and want to extract some important features from it for our data analysis.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'sentiment_analysis_generic_dataset\\', \\'api_call\\': \"pipeline(\\'text-classification\\', model=\\'Seethal/sentiment_analysis_generic_dataset\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"sentiment_analysis(\\'I love this product!\\')\", \\'performance\\': {\\'dataset\\': \\'generic_dataset\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'This is a fine-tuned downstream version of the bert-base-uncased model for sentiment analysis, this model is not intended for further downstream fine-tuning for any other tasks. This model is trained on a classified dataset for text classification.\\'}', metadata={})]", "category": "generic"}
{"question_id": 546, "text": " I'm working on a biomedical research project and need to extract features from a set of entity names to help me understand their relationships.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Token Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Named Entity Recognition\\', \\'api_name\\': \\'d4data/biomedical-ner-all\\', \\'api_call\\': \"AutoModelForTokenClassification.from_pretrained(\\'d4data/biomedical-ner-all\\')\", \\'api_arguments\\': {\\'model\\': \\'AutoModelForTokenClassification.from_pretrained(d4data/biomedical-ner-all)\\', \\'tokenizer\\': \\'AutoTokenizer.from_pretrained(d4data/biomedical-ner-all)\\', \\'aggregation_strategy\\': \\'simple\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'from transformers import pipeline, AutoTokenizer, AutoModelForTokenClassification\\'}, \\'example_code\\': \\'pipe(The patient reported no recurrence of palpitations at follow-up 6 months after the ablation.)\\', \\'performance\\': {\\'dataset\\': \\'Maccrobat\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'An English Named Entity Recognition model, trained on Maccrobat to recognize the bio-medical entities (107 entities) from a given text corpus (case reports etc.). This model was built on top of distilbert-base-uncased.\\'}', metadata={})]", "category": "generic"}
{"question_id": 547, "text": " Our company is working on an app that allows music producers to detect beats in a sample. We want to use the Hubert-large-ll60k model for this.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Feature Extraction\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Feature Extraction\\', \\'api_name\\': \\'hubert-large-ll60k\\', \\'api_call\\': \"HubertModel.from_pretrained(\\'facebook/hubert-large-ll60k\\')\", \\'api_arguments\\': \\'pretrained model name\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"hubert = HubertModel.from_pretrained(\\'facebook/hubert-large-ll60k\\')\", \\'performance\\': {\\'dataset\\': \\'Libri-Light\\', \\'accuracy\\': \\'matches or improves upon the state-of-the-art wav2vec 2.0 performance\\'}, \\'description\\': \\'Hubert-Large is a self-supervised speech representation learning model pretrained on 16kHz sampled speech audio. It is designed to deal with the unique problems in speech representation learning, such as multiple sound units in each input utterance, no lexicon of input sound units during the pre-training phase, and variable lengths of sound units with no explicit segmentation. The model relies on an offline clustering step to provide aligned target labels for a BERT-like prediction loss.\\'}', metadata={})]", "category": "generic"}
{"question_id": 548, "text": " As an art director, generating ideas from descriptions can be difficult. Produce an image of a serene lake at sunset.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'dreamlike-art/dreamlike-photoreal-2.0\\', \\'api_call\\': \"StableDiffusionPipeline.from_pretrained(\\'dreamlike-art/dreamlike-photoreal-2.0\\', torch_dtype=torch.float16)(prompt).images[0]\", \\'api_arguments\\': {\\'prompt\\': \\'photo, a church in the middle of a field of crops, bright cinematic lighting, gopro, fisheye lens\\'}, \\'python_environment_requirements\\': {\\'torch\\': \\'torch.float16\\', \\'diffusers\\': \\'StableDiffusionPipeline\\'}, \\'example_code\\': \\'from diffusers import StableDiffusionPipeline\\\\nimport torch\\\\nmodel_id = dreamlike-art/dreamlike-photoreal-2.0\\\\npipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16)\\\\npipe = pipe.to(cuda)\\\\nprompt = photo, a church in the middle of a field of crops, bright cinematic lighting, gopro, fisheye lens\\\\nimage = pipe(prompt).images[0]\\\\nimage.save(./result.jpg)\\', \\'performance\\': {\\'dataset\\': \\'Stable Diffusion 1.5\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'Dreamlike Photoreal 2.0 is a photorealistic model based on Stable Diffusion 1.5, made by dreamlike.art. It can be used to generate photorealistic images from text prompts.\\'}', metadata={})]", "category": "generic"}
{"question_id": 549, "text": " There is a new small restaurant opening and they need a store sign featuring a kangaroo eating pizza.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'text2text-generation\\', \\'api_name\\': \\'tuner007/pegasus_summarizer\\', \\'api_call\\': \"PegasusForConditionalGeneration.from_pretrained(\\'tuner007/pegasus_summarizer\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'pip install sentencepiece\\'], \\'example_code\\': \"context = \\\\nIndia wicket-keeper batsman Rishabh Pant has said someone from the crowd threw a ball on pacer Mohammed Siraj while he was fielding in the ongoing third Test against England on Wednesday. Pant revealed the incident made India skipper Virat Kohli upset. I think, somebody threw a ball inside, at Siraj, so he [Kohli] was upset, said Pant in a virtual press conference after the close of the first day\\'s play.You can say whatever you want to chant, but don\\'t throw things at the fielders and all those things. It is not good for cricket, I guess, he added.In the third session of the opening day of the third Test, a section of spectators seemed to have asked Siraj the score of the match to tease the pacer. The India pacer however came with a brilliant reply as he gestured 1-0 (India leading the Test series) towards the crowd.Earlier this month, during the second Test match, there was some bad crowd behaviour on a show as some unruly fans threw champagne corks at India batsman KL Rahul.Kohli also intervened and he was seen gesturing towards the opening batsman to know more about the incident. An over later, the TV visuals showed that many champagne corks were thrown inside the playing field, and the Indian players were visibly left frustrated.Coming back to the game, after bundling out India for 78, openers Rory Burns and Haseeb Hameed ensured that England took the honours on the opening day of the ongoing third Test.At stumps, England\\'s score reads 120/0 and the hosts have extended their lead to 42 runs. For the Three Lions, Burns (52) and Hameed (60) are currently unbeaten at the crease.Talking about the pitch on opening day, Pant said, They took the heavy roller, the wicket was much more settled down, and they batted nicely also, he said. But when we batted, the wicket was slightly soft, and they bowled in good areas, but we could have applied [ourselves] much better.Both England batsmen managed to see off the final session and the hosts concluded the opening day with all ten wickets intact, extending the lead to 42.(ANI)\\\\n\\\\nget_response(context)\", \\'performance\\': {\\'dataset\\': \\'cnn_dailymail\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.604, \\'ROUGE-2\\': 14.64, \\'ROUGE-L\\': 23.884, \\'ROUGE-LSUM\\': 32.902, \\'loss\\': 2.576, \\'gen_len\\': 76.398}}, \\'description\\': \\'PEGASUS fine-tuned for summarization\\'}', metadata={})]", "category": "generic"}
{"question_id": 550, "text": " Design a promotional poster for a new line of summer clothing featuring people wearing the clothes with a beach background.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'prompthero/openjourney-v4\\', \\'api_call\\': \"pipeline(\\'text-to-image\\', model=\\'prompthero/openjourney-v4\\')\", \\'api_arguments\\': {\\'text\\': \\'string\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"generate_image(\\'your text here\\')\", \\'performance\\': {\\'dataset\\': \\'Midjourney v4 images\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Openjourney v4 is trained on +124k Midjourney v4 images by PromptHero. It is used for generating images based on text inputs.\\'}', metadata={})]", "category": "generic"}
{"question_id": 551, "text": " We are developing an application for Japanese language learners. We need to extract text from Japanese manga pages for our users.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Image-to-Text\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'kha-white/manga-ocr-base\\', \\'api_call\\': \"pipeline(\\'ocr\\', model=\\'kha-white/manga-ocr-base\\')\", \\'api_arguments\\': \\'image\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'manga109s\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Optical character recognition for Japanese text, with the main focus being Japanese manga. It uses Vision Encoder Decoder framework. Manga OCR can be used as a general purpose printed Japanese OCR, but its main goal was to provide a high quality text recognition, robust against various scenarios specific to manga: both vertical and horizontal text, text with furigana, text overlaid on images, wide variety of fonts and font styles, and low quality images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 552, "text": " I am a museum curator, I need to learn more about an artwork in my archive so I can create better exhibition materials.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 553, "text": " We want an application that can answer questions about an image. For example, how many people are in this photo?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 554, "text": " I am developing a home security software which can detect intruders entering the house. In case any door or object is tampered, the application will ask, \\\"Who entered the room?\\\", to assist quick analysis of the CCTV recordings.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Detect Bordered and Borderless tables in documents\\', \\'api_name\\': \\'TahaDouaji/detr-doc-table-detection\\', \\'api_call\\': \"DetrForObjectDetection.from_pretrained(\\'TahaDouaji/detr-doc-table-detection\\')\", \\'api_arguments\\': [\\'images\\', \\'return_tensors\\', \\'threshold\\'], \\'python_environment_requirements\\': [\\'transformers\\', \\'torch\\', \\'PIL\\', \\'requests\\'], \\'example_code\\': \\'from transformers import DetrImageProcessor, DetrForObjectDetection\\\\nimport torch\\\\nfrom PIL import Image\\\\nimport requests\\\\nimage = Image.open(IMAGE_PATH)\\\\nprocessor = DetrImageProcessor.from_pretrained(TahaDouaji/detr-doc-table-detection)\\\\nmodel = DetrForObjectDetection.from_pretrained(TahaDouaji/detr-doc-table-detection)\\\\ninputs = processor(images=image, return_tensors=pt)\\\\noutputs = model(**inputs)\\\\ntarget_sizes = torch.tensor([image.size[::-1]])\\\\nresults = processor.post_process_object_detection(outputs, target_sizes=target_sizes, threshold=0.9)[0]\\\\nfor score, label, box in zip(results[scores], results[labels], results[boxes]):\\\\n box = [round(i, 2) for i in box.tolist()]\\\\n print(\\\\n fDetected {model.config.id2label[label.item()]} with confidence \\\\n f{round(score.item(), 3)} at location {box}\\\\n )\\', \\'performance\\': {\\'dataset\\': \\'ICDAR2019 Table Dataset\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'detr-doc-table-detection is a model trained to detect both Bordered and Borderless tables in documents, based on facebook/detr-resnet-50.\\'}', metadata={})]", "category": "generic"}
{"question_id": 555, "text": " Our business is growing rapidly, and we've received an increasing number of questions related to product images. We need the model to provide answers based on images.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 556, "text": " Our company receives invoices in different formats. We need to extract specific information from these documents to process payments and keep records.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'layoutlm-invoices\\', \\'api_call\\': \"AutoModelForDocumentQuestionAnswering.from_pretrained(\\'impira/layoutlm-invoices\\')\", \\'api_arguments\\': \\'question, context\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"nlp(question=\\'What is the total amount?\\', context=\\'your_invoice_text\\')\", \\'performance\\': {\\'dataset\\': \\'proprietary dataset of invoices, SQuAD2.0, and DocVQA\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'A fine-tuned version of the multi-modal LayoutLM model for the task of question answering on invoices and other documents. It has been fine-tuned on a proprietary dataset of invoices as well as both SQuAD2.0 and DocVQA for general comprehension. Unlike other QA models, which can only extract consecutive tokens, this model can predict longer-range, non-consecutive sequences with an additional classifier head.\\'}', metadata={})]", "category": "generic"}
{"question_id": 557, "text": " Design an AI algorithm to answer questions from scanned documents.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 558, "text": " I want to analyze the text and images in a document and extract answers to questions based on the content.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 559, "text": " Locate specific information from an invoice image, such as total amount due, invoice number, and due date.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'layoutlm-invoices\\', \\'api_call\\': \"AutoModelForDocumentQuestionAnswering.from_pretrained(\\'impira/layoutlm-invoices\\')\", \\'api_arguments\\': \\'question, context\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"nlp(question=\\'What is the total amount?\\', context=\\'your_invoice_text\\')\", \\'performance\\': {\\'dataset\\': \\'proprietary dataset of invoices, SQuAD2.0, and DocVQA\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'A fine-tuned version of the multi-modal LayoutLM model for the task of question answering on invoices and other documents. It has been fine-tuned on a proprietary dataset of invoices as well as both SQuAD2.0 and DocVQA for general comprehension. Unlike other QA models, which can only extract consecutive tokens, this model can predict longer-range, non-consecutive sequences with an additional classifier head.\\'}', metadata={})]", "category": "generic"}
{"question_id": 560, "text": " The company director needs a summary of a recent financial report. You should provide the answer of following question: What were the total revenues for the last quarter?\\n###Input: In the last quarter, the company's total revenues were reported at $3.2 million with a gross profit of $1.5 million. The operating expenses during the same quarter were $1 million.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'text2text-generation\\', \\'api_name\\': \\'financial-summarization-pegasus\\', \\'api_call\\': \"PegasusForConditionalGeneration.from_pretrained(\\'human-centered-summarization/financial-summarization-pegasus\\')\", \\'api_arguments\\': [\\'model_name\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'from transformers import PegasusTokenizer, PegasusForConditionalGeneration, TFPegasusForConditionalGeneration\\\\nmodel_name = human-centered-summarization/financial-summarization-pegasus\\\\ntokenizer = PegasusTokenizer.from_pretrained(model_name)\\\\nmodel = PegasusForConditionalGeneration.from_pretrained(model_name)\\\\ntext_to_summarize = National Commercial Bank (NCB), Saudi Arabia\u2019s largest lender by assets, agreed to buy rival Samba Financial Group for $15 billion in the biggest banking takeover this year.NCB will pay 28.45 riyals ($7.58) for each Samba share, according to a statement on Sunday, valuing it at about 55.7 billion riyals. NCB will offer 0.739 new shares for each Samba share, at the lower end of the 0.736-0.787 ratio the banks set when they signed an initial framework agreement in June.The offer is a 3.5% premium to Samba\u2019s Oct. 8 closing price of 27.50 riyals and about 24% higher than the level the shares traded at before the talks were made public. Bloomberg News first reported the merger discussions.The new bank will have total assets of more than $220 billion, creating the Gulf region\u2019s third-largest lender. The entity\u2019s $46 billion market capitalization nearly matches that of Qatar National Bank QPSC, which is still the Middle East\u2019s biggest lender with about $268 billion of assets.\\\\ninput_ids = tokenizer(text_to_summarize, return_tensors=pt).input_ids\\\\noutput = model.generate(input_ids, max_length=32, num_beams=5, early_stopping=True)\\\\nprint(tokenizer.decode(output[0], skip_special_tokens=True))\\', \\'performance\\': {\\'dataset\\': \\'xsum\\', \\'accuracy\\': {\\'ROUGE-1\\': 35.206, \\'ROUGE-2\\': 16.569, \\'ROUGE-L\\': 30.128, \\'ROUGE-LSUM\\': 30.171}}, \\'description\\': \\'This model was fine-tuned on a novel financial news dataset, which consists of 2K articles from Bloomberg, on topics such as stock, markets, currencies, rate and cryptocurrencies. It is based on the PEGASUS model and in particular PEGASUS fine-tuned on the Extreme Summarization (XSum) dataset: google/pegasus-xsum model. PEGASUS was originally proposed by Jingqing Zhang, Yao Zhao, Mohammad Saleh and Peter J. Liu in PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization.\\'}', metadata={})]", "category": "generic"}
{"question_id": 561, "text": " You must create a tool to assess safety risks automatically in construction sites by estimating the distance of tools and workers.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'kochetkovIT/autotrain-ironhack-49741119788\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'json\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'kochetkovIT/autotrain-data-ironhack\\', \\'accuracy\\': {\\'Loss\\': 2.603, \\'R2\\': 0.013, \\'MSE\\': 6.776, \\'MAE\\': 1.666, \\'RMSLE\\': 0.502}}, \\'description\\': \\'A tabular regression model trained using AutoTrain to predict carbon emissions (in grams) with an R2 score of 0.013.\\'}', metadata={})]", "category": "generic"}
{"question_id": 562, "text": " Develop a solution to estimate the depth of objects in an image.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Depth Estimation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Depth Estimation\\', \\'api_name\\': \\'glpn-kitti\\', \\'api_call\\': \"GLPNForDepthEstimation.from_pretrained(\\'vinvino02/glpn-kitti\\')\", \\'api_arguments\\': \\'images, return_tensors\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'from transformers import GLPNFeatureExtractor, GLPNForDepthEstimation\\\\nimport torch\\\\nimport numpy as np\\\\nfrom PIL import Image\\\\nimport requests\\\\nurl = http://images.cocodataset.org/val2017/000000039769.jpg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nfeature_extractor = GLPNFeatureExtractor.from_pretrained(vinvino02/glpn-kitti)\\\\nmodel = GLPNForDepthEstimation.from_pretrained(vinvino02/glpn-kitti)\\\\ninputs = feature_extractor(images=image, return_tensors=pt)\\\\nwith torch.no_grad():\\\\n outputs = model(**inputs)\\\\n predicted_depth = outputs.predicted_depth\\\\nprediction = torch.nn.functional.interpolate(\\\\n predicted_depth.unsqueeze(1),\\\\n size=image.size[::-1],\\\\n mode=bicubic,\\\\n align_corners=False,\\\\n)\\\\noutput = prediction.squeeze().cpu().numpy()\\\\nformatted = (output * 255 / np.max(output)).astype(uint8)\\\\ndepth = Image.fromarray(formatted)\\', \\'performance\\': {\\'dataset\\': \\'KITTI\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Global-Local Path Networks (GLPN) model trained on KITTI for monocular depth estimation. It was introduced in the paper Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth by Kim et al. and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 563, "text": " We need to analyze some pictures from nature and classify them to protect some species of animals.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'michellejieli/NSFW_text_classifier\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'michellejieli/NSFW_text_classification\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'classifier(I see you\u2019ve set aside this special time to humiliate yourself in public.)\\', \\'performance\\': {\\'dataset\\': \\'Reddit posts\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'DistilBERT is a transformer model that performs sentiment analysis. I fine-tuned the model on Reddit posts with the purpose of classifying not safe for work (NSFW) content, specifically text that is considered inappropriate and unprofessional. The model predicts 2 classes, which are NSFW or safe for work (SFW). The model is a fine-tuned version of DistilBERT. It was fine-tuned on 14317 Reddit posts pulled from the Reddit API.\\'}', metadata={})]", "category": "generic"}
{"question_id": 564, "text": " We need to detect objects in an image provided by a user and describe their locations.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Object Detection\\', \\'api_name\\': \\'keremberke/yolov8n-csgo-player-detection\\', \\'api_call\\': \"YOLO(\\'keremberke/yolov8n-csgo-player-detection\\')\", \\'api_arguments\\': {\\'image\\': \\'URL or local path to image\\'}, \\'python_environment_requirements\\': \\'pip install ultralyticsplus==0.0.23 ultralytics==8.0.21\\', \\'example_code\\': \"from ultralyticsplus import YOLO, render_result\\\\n\\\\nmodel = YOLO(\\'keremberke/yolov8n-csgo-player-detection\\')\\\\n\\\\nmodel.overrides[\\'conf\\'] = 0.25\\\\nmodel.overrides[\\'iou\\'] = 0.45\\\\nmodel.overrides[\\'agnostic_nms\\'] = False\\\\nmodel.overrides[\\'max_det\\'] = 1000\\\\n\\\\nimage = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\\\\n\\\\nresults = model.predict(image)\\\\n\\\\nprint(results[0].boxes)\\\\nrender = render_result(model=model, image=image, result=results[0])\\\\nrender.show()\", \\'performance\\': {\\'dataset\\': \\'csgo-object-detection\\', \\'accuracy\\': 0.844}, \\'description\\': \"A YOLOv8 model for detecting Counter-Strike: Global Offensive (CS:GO) players with supported labels: [\\'ct\\', \\'cthead\\', \\'t\\', \\'thead\\'].\"}', metadata={})]", "category": "generic"}
{"question_id": 565, "text": " We are creating an analysis platform for Counter-Strike: Global Offensive. Detect and locate players in the given image.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Object Detection\\', \\'api_name\\': \\'keremberke/yolov8n-csgo-player-detection\\', \\'api_call\\': \"YOLO(\\'keremberke/yolov8n-csgo-player-detection\\')\", \\'api_arguments\\': {\\'image\\': \\'URL or local path to image\\'}, \\'python_environment_requirements\\': \\'pip install ultralyticsplus==0.0.23 ultralytics==8.0.21\\', \\'example_code\\': \"from ultralyticsplus import YOLO, render_result\\\\n\\\\nmodel = YOLO(\\'keremberke/yolov8n-csgo-player-detection\\')\\\\n\\\\nmodel.overrides[\\'conf\\'] = 0.25\\\\nmodel.overrides[\\'iou\\'] = 0.45\\\\nmodel.overrides[\\'agnostic_nms\\'] = False\\\\nmodel.overrides[\\'max_det\\'] = 1000\\\\n\\\\nimage = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\\\\n\\\\nresults = model.predict(image)\\\\n\\\\nprint(results[0].boxes)\\\\nrender = render_result(model=model, image=image, result=results[0])\\\\nrender.show()\", \\'performance\\': {\\'dataset\\': \\'csgo-object-detection\\', \\'accuracy\\': 0.844}, \\'description\\': \"A YOLOv8 model for detecting Counter-Strike: Global Offensive (CS:GO) players with supported labels: [\\'ct\\', \\'cthead\\', \\'t\\', \\'thead\\'].\"}', metadata={})]", "category": "generic"}
{"question_id": 566, "text": " Build an AI model that can recognize the objects in a given image and draw a boundary around them.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'abhishek/autotrain-dog-vs-food\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'abhishek/autotrain-dog-vs-food\\')\", \\'api_arguments\\': \\'image_path\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'sasha/dog-food\\', \\'accuracy\\': 0.998}, \\'description\\': \"A pre-trained model for classifying images as either dog or food using Hugging Face\\'s AutoTrain framework.\"}', metadata={})]", "category": "generic"}
{"question_id": 567, "text": " Design a pipeline to create artistic variations of an input image.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Image Variations\\', \\'api_name\\': \\'lambdalabs/sd-image-variations-diffusers\\', \\'api_call\\': \"StableDiffusionImageVariationPipeline.from_pretrained(\\'lambdalabs/sd-image-variations-diffusers\\', revision=\\'v2.0\\')\", \\'api_arguments\\': {\\'revision\\': \\'v2.0\\'}, \\'python_environment_requirements\\': \\'Diffusers >=0.8.0\\', \\'example_code\\': \\'from diffusers import StableDiffusionImageVariationPipeline\\\\nfrom PIL import Image\\\\ndevice = cuda:0\\\\nsd_pipe = StableDiffusionImageVariationPipeline.from_pretrained(\\\\n lambdalabs/sd-image-variations-diffusers,\\\\n revision=v2.0,\\\\n)\\\\nsd_pipe = sd_pipe.to(device)\\\\nim = Image.open(path/to/image.jpg)\\\\ntform = transforms.Compose([\\\\n transforms.ToTensor(),\\\\n transforms.Resize(\\\\n  (224, 224),\\\\n  interpolation=transforms.InterpolationMode.BICUBIC,\\\\n  antialias=False,\\\\n ),\\\\n transforms.Normalize(\\\\n  [0.48145466, 0.4578275, 0.40821073],\\\\n  [0.26862954, 0.26130258, 0.27577711]),\\\\n])\\\\ninp = tform(im).to(device).unsqueeze(0)\\\\nout = sd_pipe(inp, guidance_scale=3)\\\\nout[images][0].save(result.jpg)\\', \\'performance\\': {\\'dataset\\': \\'ChristophSchuhmann/improved_aesthetics_6plus\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'This version of Stable Diffusion has been fine tuned from CompVis/stable-diffusion-v1-4-original to accept CLIP image embedding rather than text embeddings. This allows the creation of image variations similar to DALLE-2 using Stable Diffusion.\\'}', metadata={})]", "category": "generic"}
{"question_id": 568, "text": " A user wants to improve the quality of a small sized picture from her vacation. We need to upscale the picture to 2x its size without losing quality.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Image Upscaling\\', \\'api_name\\': \\'stabilityai/sd-x2-latent-upscaler\\', \\'api_call\\': \\'StableDiffusionLatentUpscalePipeline.from_pretrained(stabilityai/sd-x2-latent-upscaler, torch_dtype=torch.float16)\\', \\'api_arguments\\': {\\'prompt\\': \\'text prompt\\', \\'image\\': \\'low resolution latents\\', \\'num_inference_steps\\': 20, \\'guidance_scale\\': 0, \\'generator\\': \\'torch generator\\'}, \\'python_environment_requirements\\': [\\'git+https://github.com/huggingface/diffusers.git\\', \\'transformers\\', \\'accelerate\\', \\'scipy\\', \\'safetensors\\'], \\'example_code\\': \\'from diffusers import StableDiffusionLatentUpscalePipeline, StableDiffusionPipeline\\\\nimport torch\\\\npipeline = StableDiffusionPipeline.from_pretrained(CompVis/stable-diffusion-v1-4, torch_dtype=torch.float16)\\\\npipeline.to(cuda)\\\\nupscaler = StableDiffusionLatentUpscalePipeline.from_pretrained(stabilityai/sd-x2-latent-upscaler, torch_dtype=torch.float16)\\\\nupscaler.to(cuda)\\\\nprompt = a photo of an astronaut high resolution, unreal engine, ultra realistic\\\\ngenerator = torch.manual_seed(33)\\\\nlow_res_latents = pipeline(prompt, generator=generator, output_type=latent).images\\\\nupscaled_image = upscaler(prompt=prompt, image=low_res_latents, num_inference_steps=20, guidance_scale=0, generator=generator).images[0]\\\\nupscaled_image.save(astronaut_1024.png)\\', \\'performance\\': {\\'dataset\\': \\'LAION-2B\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \"Stable Diffusion x2 latent upscaler is a diffusion-based upscaler model developed by Katherine Crowson in collaboration with Stability AI. It is designed to upscale Stable Diffusion\\'s latent denoised image embeddings, allowing for fast text-to-image and upscaling pipelines. The model was trained on a high-resolution subset of the LAION-2B dataset and works with all Stable Diffusion checkpoints.\"}', metadata={})]", "category": "generic"}
{"question_id": 569, "text": " I want to create a visual representation based on a short description: \\\"A magical forest with unicorns and a rainbow.\\\".\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'text2text-generation\\', \\'api_name\\': \\'tuner007/pegasus_summarizer\\', \\'api_call\\': \"PegasusForConditionalGeneration.from_pretrained(\\'tuner007/pegasus_summarizer\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'pip install sentencepiece\\'], \\'example_code\\': \"context = \\\\nIndia wicket-keeper batsman Rishabh Pant has said someone from the crowd threw a ball on pacer Mohammed Siraj while he was fielding in the ongoing third Test against England on Wednesday. Pant revealed the incident made India skipper Virat Kohli upset. I think, somebody threw a ball inside, at Siraj, so he [Kohli] was upset, said Pant in a virtual press conference after the close of the first day\\'s play.You can say whatever you want to chant, but don\\'t throw things at the fielders and all those things. It is not good for cricket, I guess, he added.In the third session of the opening day of the third Test, a section of spectators seemed to have asked Siraj the score of the match to tease the pacer. The India pacer however came with a brilliant reply as he gestured 1-0 (India leading the Test series) towards the crowd.Earlier this month, during the second Test match, there was some bad crowd behaviour on a show as some unruly fans threw champagne corks at India batsman KL Rahul.Kohli also intervened and he was seen gesturing towards the opening batsman to know more about the incident. An over later, the TV visuals showed that many champagne corks were thrown inside the playing field, and the Indian players were visibly left frustrated.Coming back to the game, after bundling out India for 78, openers Rory Burns and Haseeb Hameed ensured that England took the honours on the opening day of the ongoing third Test.At stumps, England\\'s score reads 120/0 and the hosts have extended their lead to 42 runs. For the Three Lions, Burns (52) and Hameed (60) are currently unbeaten at the crease.Talking about the pitch on opening day, Pant said, They took the heavy roller, the wicket was much more settled down, and they batted nicely also, he said. But when we batted, the wicket was slightly soft, and they bowled in good areas, but we could have applied [ourselves] much better.Both England batsmen managed to see off the final session and the hosts concluded the opening day with all ten wickets intact, extending the lead to 42.(ANI)\\\\n\\\\nget_response(context)\", \\'performance\\': {\\'dataset\\': \\'cnn_dailymail\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.604, \\'ROUGE-2\\': 14.64, \\'ROUGE-L\\': 23.884, \\'ROUGE-LSUM\\': 32.902, \\'loss\\': 2.576, \\'gen_len\\': 76.398}}, \\'description\\': \\'PEGASUS fine-tuned for summarization\\'}', metadata={})]", "category": "generic"}
{"question_id": 570, "text": " Explain how to use GPT-3 to create a slogan for an e-commerce website that sells eco-friendly products.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Conversational\\', \\'api_name\\': \\'ingen51/DialoGPT-medium-GPT4\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'ingen51/DialoGPT-medium-GPT4\\')\", \\'api_arguments\\': [\\'message\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"conversational_pipeline(\\'Hello, how are you?\\')\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'A GPT-4 model for generating conversational responses in a dialogue setting.\\'}', metadata={})]", "category": "generic"}
{"question_id": 571, "text": " I'm working on a project that requires generating high-quality images of faces for a set of characters in a video game.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'gsdf/Counterfeit-V2.5\\', \\'api_call\\': \"pipeline(\\'text-to-image\\', model=\\'gsdf/Counterfeit-V2.5\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'((masterpiece,best quality)),1girl, solo, animal ears, rabbit, barefoot, knees up, dress, sitting, rabbit ears, short sleeves, looking at viewer, grass, short hair, smile, white hair, puffy sleeves, outdoors, puffy short sleeves, bangs, on ground, full body, animal, white dress, sunlight, brown eyes, dappled sunlight, day, depth of field\\', \\'performance\\': {\\'dataset\\': \\'EasyNegative\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Counterfeit-V2.5 is a text-to-image model that generates anime-style images based on text prompts. It has been updated for ease of use and can be used with negative prompts to create high-quality images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 572, "text": " Our client wants to analyze videos for their marketing website. They need a quick solution to categorize video content without worrying about accuracy.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'chavinlo/TempoFunk\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'chavinlo/TempoFunk\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Text-to-Video model using Hugging Face Transformers library. Model is capable of generating video content based on the input text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 573, "text": " I want to classify images of vehicles including cars, motorcycles, trucks, and bicycles, based on their appearance.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'abhishek/autotrain-dog-vs-food\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'abhishek/autotrain-dog-vs-food\\')\", \\'api_arguments\\': \\'image_path\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'sasha/dog-food\\', \\'accuracy\\': 0.998}, \\'description\\': \"A pre-trained model for classifying images as either dog or food using Hugging Face\\'s AutoTrain framework.\"}', metadata={})]", "category": "generic"}
{"question_id": 574, "text": " I need to classify images of animals into their specific categories.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'abhishek/autotrain-dog-vs-food\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'abhishek/autotrain-dog-vs-food\\')\", \\'api_arguments\\': \\'image_path\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'sasha/dog-food\\', \\'accuracy\\': 0.998}, \\'description\\': \"A pre-trained model for classifying images as either dog or food using Hugging Face\\'s AutoTrain framework.\"}', metadata={})]", "category": "generic"}
{"question_id": 575, "text": " We are building a location recommendation system that identifies possible locations for new stores based on images from potential locations. Use the StreetCLIP model to generate probabilities for various cities.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Zero-Shot Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Geolocalization\\', \\'api_name\\': \\'geolocal/StreetCLIP\\', \\'api_call\\': \"CLIPModel.from_pretrained(\\'geolocal/StreetCLIP\\')\", \\'api_arguments\\': {\\'pretrained_model_name_or_path\\': \\'geolocal/StreetCLIP\\'}, \\'python_environment_requirements\\': [\\'transformers\\', \\'PIL\\', \\'requests\\'], \\'example_code\\': \\'from PIL import Image\\\\nimport requests\\\\nfrom transformers import CLIPProcessor, CLIPModel\\\\nmodel = CLIPModel.from_pretrained(geolocal/StreetCLIP)\\\\nprocessor = CLIPProcessor.from_pretrained(geolocal/StreetCLIP)\\\\nurl = https://huggingface.co/geolocal/StreetCLIP/resolve/main/sanfrancisco.jpeg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nchoices = [San Jose, San Diego, Los Angeles, Las Vegas, San Francisco]\\\\ninputs = processor(text=choices, images=image, return_tensors=pt, padding=True)\\\\noutputs = model(**inputs)\\\\nlogits_per_image = outputs.logits_per_image\\\\nprobs = logits_per_image.softmax(dim=1)\\', \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'IM2GPS\\', \\'accuracy\\': {\\'25km\\': 28.3, \\'200km\\': 45.1, \\'750km\\': 74.7, \\'2500km\\': 88.2}}, {\\'name\\': \\'IM2GPS3K\\', \\'accuracy\\': {\\'25km\\': 22.4, \\'200km\\': 37.4, \\'750km\\': 61.3, \\'2500km\\': 80.4}}]}, \\'description\\': \\'StreetCLIP is a robust foundation model for open-domain image geolocalization and other geographic and climate-related tasks. Trained on an original dataset of 1.1 million street-level urban and rural geo-tagged images, it achieves state-of-the-art performance on multiple open-domain image geolocalization benchmarks in zero-shot, outperforming supervised models trained on millions of images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 576, "text": " We are receiving mixed reviews about our new product and we need to understand customers' reactions.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'michellejieli/NSFW_text_classifier\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'michellejieli/NSFW_text_classification\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'classifier(I see you\u2019ve set aside this special time to humiliate yourself in public.)\\', \\'performance\\': {\\'dataset\\': \\'Reddit posts\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'DistilBERT is a transformer model that performs sentiment analysis. I fine-tuned the model on Reddit posts with the purpose of classifying not safe for work (NSFW) content, specifically text that is considered inappropriate and unprofessional. The model predicts 2 classes, which are NSFW or safe for work (SFW). The model is a fine-tuned version of DistilBERT. It was fine-tuned on 14317 Reddit posts pulled from the Reddit API.\\'}', metadata={})]", "category": "generic"}
{"question_id": 577, "text": " Detect if there are any harmful messages in a chat room.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'michellejieli/NSFW_text_classifier\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'michellejieli/NSFW_text_classification\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'classifier(I see you\u2019ve set aside this special time to humiliate yourself in public.)\\', \\'performance\\': {\\'dataset\\': \\'Reddit posts\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'DistilBERT is a transformer model that performs sentiment analysis. I fine-tuned the model on Reddit posts with the purpose of classifying not safe for work (NSFW) content, specifically text that is considered inappropriate and unprofessional. The model predicts 2 classes, which are NSFW or safe for work (SFW). The model is a fine-tuned version of DistilBERT. It was fine-tuned on 14317 Reddit posts pulled from the Reddit API.\\'}', metadata={})]", "category": "generic"}
{"question_id": 578, "text": " I am developing software that needs to retrieve relevant information from a collection of documents based on a user's query.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Text2Text Generation\\', \\'api_name\\': \\'castorini/doc2query-t5-base-msmarco\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'castorini/doc2query-t5-base-msmarco\\')\", \\'api_arguments\\': \\'text, max_length\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'MS MARCO\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'A T5 model trained on the MS MARCO dataset for generating queries from documents.\\'}', metadata={})]", "category": "generic"}
{"question_id": 579, "text": " Our client is a media organization that wants to extract named entities from a large collection of news articles in order to identify key people and places\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Token Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Named Entity Recognition\\', \\'api_name\\': \\'dslim/bert-base-NER-uncased\\', \\'api_call\\': \"pipeline(\\'ner\\', model=\\'dslim/bert-base-NER-uncased\\')\", \\'api_arguments\\': {}, \\'python_environment_requirements\\': {\\'transformers\\': \\'>=4.0.0\\'}, \\'example_code\\': \"nlp(\\'My name is John and I live in New York.\\')\", \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A pretrained BERT model for Named Entity Recognition (NER) on uncased text. It can be used to extract entities such as person names, locations, and organizations from text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 580, "text": " Extract entities from a provided sentence mentioning various companies and their CEOs.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Sentence Similarity\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Sentence Transformers\\', \\'api_name\\': \\'sentence-transformers/all-roberta-large-v1\\', \\'api_call\\': \"SentenceTransformer(\\'sentence-transformers/all-roberta-large-v1\\')\", \\'api_arguments\\': [\\'sentences\\'], \\'python_environment_requirements\\': \\'pip install -U sentence-transformers\\', \\'example_code\\': \"from sentence_transformers import SentenceTransformer\\\\nsentences = [This is an example sentence, Each sentence is converted]\\\\nmodel = SentenceTransformer(\\'sentence-transformers/all-roberta-large-v1\\')\\\\nembeddings = model.encode(sentences)\\\\nprint(embeddings)\", \\'performance\\': {\\'dataset\\': \\'https://seb.sbert.net\\', \\'accuracy\\': \\'Automated evaluation\\'}, \\'description\\': \\'This is a sentence-transformers model: It maps sentences & paragraphs to a 1024 dimensional dense vector space and can be used for tasks like clustering or semantic search.\\'}', metadata={})]", "category": "generic"}
{"question_id": 581, "text": " I am developing a food application where food keywords need to be extracted from user's input text. The model should be able to recognize food-related named entities.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Token Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Named Entity Recognition\\', \\'api_name\\': \\'Dizex/InstaFoodRoBERTa-NER\\', \\'api_call\\': \"AutoModelForTokenClassification.from_pretrained(\\'Dizex/InstaFoodRoBERTa-NER\\')\", \\'api_arguments\\': {\\'model\\': \"AutoModelForTokenClassification.from_pretrained(\\'Dizex/InstaFoodRoBERTa-NER\\')\", \\'tokenizer\\': \"AutoTokenizer.from_pretrained(\\'Dizex/InstaFoodRoBERTa-NER\\')\"}, \\'python_environment_requirements\\': {\\'transformers\\': \\'from transformers import AutoTokenizer, AutoModelForTokenClassification, pipeline\\'}, \\'example_code\\': \"example = Today\\'s meal: Fresh olive pok bowl topped with chia seeds. Very delicious!\\\\nner_entity_results = pipe(example, aggregation_strategy=\\'simple\\')\\\\nprint(ner_entity_results)\", \\'performance\\': {\\'dataset\\': \\'Dizex/InstaFoodSet\\', \\'accuracy\\': {\\'f1\\': 0.91, \\'precision\\': 0.89, \\'recall\\': 0.93}}, \\'description\\': \\'InstaFoodRoBERTa-NER is a fine-tuned BERT model that is ready to use for Named Entity Recognition of Food entities on informal text (social media like). It has been trained to recognize a single entity: food (FOOD). Specifically, this model is a roberta-base model that was fine-tuned on a dataset consisting of 400 English Instagram posts related to food.\\'}', metadata={})]", "category": "generic"}
{"question_id": 582, "text": " Create a script for an AI bot that automatically adds punctuation to users' messages in a chat app.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'Zixtrauce/BaekBot\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'Zixtrauce/BaekBot\\')\", \\'api_arguments\\': \\'input_message\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"conversational_pipeline(\\'input_message\\')\", \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'BaekBot is a conversational model based on the GPT-2 architecture for text generation. It can be used for generating human-like responses in a chat-like environment.\\'}', metadata={})]", "category": "generic"}
{"question_id": 583, "text": " I am writing an article on the history of technology companies, and I want to extract the names of companies and people mentioned in the text.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'text2text-generation\\', \\'api_name\\': \\'google/pegasus-newsroom\\', \\'api_call\\': \"pipeline(\\'summarization\\', model=\\'google/pegasus-newsroom\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'This model can be loaded on the Inference API on-demand.\\', \\'performance\\': {\\'dataset\\': \\'newsroom\\', \\'accuracy\\': \\'45.98/34.20/42.18\\'}, \\'description\\': \\'PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization by Jingqing Zhang, Yao Zhao, Mohammad Saleh, and Peter J. Liu. The model is trained on both C4 and HugeNews datasets and is designed for summarization tasks.\\'}', metadata={})]", "category": "generic"}
{"question_id": 584, "text": " I have a diary entry and want to identify the names of people and locations mentioned in it.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'tiennvcs/layoutlmv2-large-uncased-finetuned-vi-infovqa\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=\\'tiennvcs/layoutlmv2-large-uncased-finetuned-vi-infovqa\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers==4.15.0, torch==1.8.0+cu101, datasets==1.17.0, tokenizers==0.10.3\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'unknown\\', \\'accuracy\\': {\\'Loss\\': 8.5806}}, \\'description\\': \\'This model is a fine-tuned version of microsoft/layoutlmv2-large-uncased on an unknown dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 585, "text": " Our company has a deal with international clients, and it's important to detect the proper locations of meetings from multilingual texts provided to us.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'PyTorch Transformers\\', \\'functionality\\': \\'text2text-generation\\', \\'api_name\\': \\'facebook/nllb-200-distilled-600M\\', \\'api_call\\': \"pipeline(\\'translation_xx_to_yy\\', model=\\'facebook/nllb-200-distilled-600M\\')\", \\'api_arguments\\': [\\'model\\', \\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\', \\'torch\\'], \\'example_code\\': \"from transformers import pipeline; translator = pipeline(\\'translation_xx_to_yy\\', model=\\'facebook/nllb-200-distilled-600M\\'); translator(\\'Hello World\\')\", \\'performance\\': {\\'dataset\\': \\'Flores-200\\', \\'accuracy\\': \\'BLEU, spBLEU, chrF++\\'}, \\'description\\': \\'NLLB-200 is a machine translation model primarily intended for research in machine translation, especially for low-resource languages. It allows for single sentence translation among 200 languages. The model was trained on general domain text data and is not intended to be used with domain specific texts, such as medical domain or legal domain. The model is not intended to be used for document translation.\\'}', metadata={})]", "category": "generic"}
{"question_id": 586, "text": " We need to extract information from a table to answer a user's question regarding a company's revenue.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Table Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Table Question Answering\\', \\'api_name\\': \\'google/tapas-small-finetuned-sqa\\', \\'api_call\\': \"pipeline(\\'table-question-answering\\', model=\\'google/tapas-small-finetuned-sqa\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'msr_sqa\\', \\'accuracy\\': 0.6155}, \\'description\\': \\'TAPAS small model fine-tuned on Sequential Question Answering (SQA). It uses relative position embeddings (i.e. resetting the position index at every cell of the table).\\'}', metadata={})]", "category": "generic"}
{"question_id": 587, "text": " Develop a table-based question answering system that can answer queries based on the input data.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Table Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Table Question Answering\\', \\'api_name\\': \\'google/tapas-small-finetuned-sqa\\', \\'api_call\\': \"pipeline(\\'table-question-answering\\', model=\\'google/tapas-small-finetuned-sqa\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'msr_sqa\\', \\'accuracy\\': 0.6155}, \\'description\\': \\'TAPAS small model fine-tuned on Sequential Question Answering (SQA). It uses relative position embeddings (i.e. resetting the position index at every cell of the table).\\'}', metadata={})]", "category": "generic"}
{"question_id": 588, "text": " Create a tool that receives a table and a question in natural language, and returns an answer to the question based on the inputted table.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Table Question Answering\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Table Question Answering\\', \\'api_name\\': \\'lysandre/tapas-temporary-repo\\', \\'api_call\\': \"TapasForQuestionAnswering.from_pretrained(\\'lysandre/tapas-temporary-repo\\')\", \\'api_arguments\\': [\\'model_name\\', \\'question\\', \\'table\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"tokenizer = TapasTokenizer.from_pretrained(\\'lysandre/tapas-temporary-repo\\')\\\\nmodel = TapasForQuestionAnswering.from_pretrained(\\'lysandre/tapas-temporary-repo\\')\\\\ninputs = tokenizer(table=table, queries=question, return_tensors=\\'pt\\')\\\\noutputs = model(**inputs)\\\\npredicted_answer_coordinates, predicted_aggregation_indices = tokenizer.convert_logits_to_predictions(inputs, outputs.logits.detach(), outputs.logits_aggregation.detach())\", \\'performance\\': {\\'dataset\\': \\'SQA\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'TAPAS base model fine-tuned on Sequential Question Answering (SQA). This model is pretrained on a large corpus of English data from Wikipedia in a self-supervised fashion and can be used for answering questions related to a table in a conversational set-up.\\'}', metadata={})]", "category": "generic"}
{"question_id": 589, "text": " We are a medical company providing health FAQs. We need to answer customers' questions accurately.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'deepset/roberta-base-squad2-covid\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=RobertaForQuestionAnswering.from_pretrained(\\'deepset/roberta-base-squad2-covid\\'), tokenizer=RobertaTokenizer.from_pretrained(\\'deepset/roberta-base-squad2-covid\\'))\", \\'api_arguments\\': {\\'model_name\\': \\'deepset/roberta-base-squad2-covid\\', \\'tokenizer\\': \\'deepset/roberta-base-squad2-covid\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': {\\'QA_input\\': {\\'question\\': \\'Why is model conversion important?\\', \\'context\\': \\'The option to convert models between FARM and transformers gives freedom to the user and let people easily switch between frameworks.\\'}, \\'res\\': \\'nlp(QA_input)\\'}, \\'performance\\': {\\'dataset\\': \\'squad_v2\\', \\'accuracy\\': {\\'XVAL_EM\\': 0.17890995260663506, \\'XVAL_f1\\': 0.49925444207319924, \\'XVAL_top_3_recall\\': 0.8021327014218009}}, \\'description\\': \\'This model is a Roberta-based model fine-tuned on SQuAD-style CORD-19 annotations for the task of extractive question answering in the context of COVID-19. It can be used with the Hugging Face Transformers library for question answering tasks.\\'}', metadata={})]", "category": "generic"}
{"question_id": 590, "text": " A group of students are doing a project on European capitals. They need to know the capital city of Germany.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Zero-Shot Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Classification\\', \\'api_name\\': \\'svalabs/gbert-large-zeroshot-nli\\', \\'api_call\\': \"pipeline(\\'zero-shot-classification\\', model=\\'svalabs/gbert-large-zeroshot-nli\\')\", \\'api_arguments\\': [\\'sequence\\', \\'labels\\', \\'hypothesis_template\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'from transformers import pipeline\\\\nzershot_pipeline = pipeline(zero-shot-classification, model=svalabs/gbert-large-zeroshot-nli)\\\\nsequence = Ich habe ein Problem mit meinem Iphone das so schnell wie m\u00f6glich gel\u00f6st werden muss\\\\nlabels = [Computer, Handy, Tablet, dringend, nicht dringend]\\\\nhypothesis_template = In diesem Satz geht es um das Thema {}. \\', \\'performance\\': {\\'dataset\\': \\'XNLI TEST-Set\\', \\'accuracy\\': \\'85.6%\\'}, \\'description\\': \\'A German zeroshot classification model based on the German BERT large model from deepset.ai and finetuned for natural language inference using machine-translated nli sentence pairs from mnli, anli, and snli datasets.\\'}', metadata={})]", "category": "generic"}
{"question_id": 591, "text": " Assist us in creating a question answering system to provide quick answers for customer inquiries.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 592, "text": " We are worried about price inflation in our country. Can you answer our questions on price inflation using the BERT large cased whole word masking finetuned model on SQuAD?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'bert-large-uncased-whole-word-masking-squad2\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=AutoModel.from_pretrained(\\'deepset/bert-large-uncased-whole-word-masking-squad2\\'), tokenizer=AutoTokenizer.from_pretrained(\\'deepset/bert-large-uncased-whole-word-masking-squad2\\'))\", \\'api_arguments\\': {\\'model_name\\': \\'deepset/bert-large-uncased-whole-word-masking-squad2\\', \\'tokenizer\\': \\'deepset/bert-large-uncased-whole-word-masking-squad2\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"QA_input = {\\\\n \\'question\\': \\'Why is model conversion important?\\',\\\\n \\'context\\': \\'The option to convert models between FARM and transformers gives freedom to the user and let people easily switch between frameworks.\\'\\\\n}\\\\nres = nlp(QA_input)\", \\'performance\\': {\\'dataset\\': \\'squad_v2\\', \\'accuracy\\': {\\'Exact Match\\': 80.885, \\'F1\\': 83.876}}, \\'description\\': \\'This is a bert-large model, fine-tuned using the SQuAD2.0 dataset for the task of question answering. It is designed for extractive question answering and supports English language.\\'}', metadata={})]", "category": "generic"}
{"question_id": 593, "text": " I am interested in developing a system that can categorize my text messages into different subjects like finances, health, and entertainment.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'michellejieli/NSFW_text_classifier\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'michellejieli/NSFW_text_classification\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'classifier(I see you\u2019ve set aside this special time to humiliate yourself in public.)\\', \\'performance\\': {\\'dataset\\': \\'Reddit posts\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'DistilBERT is a transformer model that performs sentiment analysis. I fine-tuned the model on Reddit posts with the purpose of classifying not safe for work (NSFW) content, specifically text that is considered inappropriate and unprofessional. The model predicts 2 classes, which are NSFW or safe for work (SFW). The model is a fine-tuned version of DistilBERT. It was fine-tuned on 14317 Reddit posts pulled from the Reddit API.\\'}', metadata={})]", "category": "generic"}
{"question_id": 594, "text": " An online magazine editor is trying to clujster articles into topics. We want to classify articles into categories like sport, politics, health, and technology written in French.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'it5-base-news-summarization\\', \\'api_call\\': \"pipeline(\\'summarization\\', model=\\'it5/it5-base-news-summarization\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'newsum(Dal 31 maggio \u00e8 infine partita la piattaforma ITsART, a pi\u00f9 di un anno da quando \u2013 durante il primo lockdown \u2013 il ministro della Cultura Dario Franceschini ne aveva parlato come di \u00abuna sorta di Netflix della cultura\u00bb, pensata per \u00aboffrire a tutto il mondo la cultura italiana a pagamento\u00bb. \u00c8 presto per dare giudizi definitivi sulla piattaforma, e di certo sar\u00e0 difficile farlo anche pi\u00f9 avanti senza numeri precisi. Al momento, l\u2019unica cosa che si pu\u00f2 fare \u00e8 guardare com\u2019\u00e8 fatto il sito, contare quanti contenuti ci sono (circa 700 \u201ctitoli\u201d, tra film, documentari, spettacoli teatrali e musicali e altri eventi) e provare a dare un giudizio sul loro valore e sulla loro variet\u00e0. Intanto, una cosa notata da pi\u00f9 parti \u00e8 che diversi contenuti di ITsART sono a pagamento sulla piattaforma sebbene altrove, per esempio su RaiPlay, siano invece disponibili gratuitamente.)\\', \\'performance\\': {\\'dataset\\': \\'NewsSum-IT\\', \\'accuracy\\': {\\'Rouge1\\': 0.339, \\'Rouge2\\': 0.16, \\'RougeL\\': 0.263}}, \\'description\\': \\'IT5 Base model fine-tuned on news summarization on the Fanpage and Il Post corpora for Italian Language Understanding and Generation.\\'}', metadata={})]", "category": "generic"}
{"question_id": 595, "text": " We have a user manual in English that needs to be translated into Spanish to be displayed on our Spanish website.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Translation\\', \\'api_name\\': \\'Helsinki-NLP/opus-mt-es-en\\', \\'api_call\\': \"pipeline(\\'translation_es_to_en\\', model=\\'Helsinki-NLP/opus-mt-es-en\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\ntranslation = pipeline(\\'translation_es_to_en\\', model=\\'Helsinki-NLP/opus-mt-es-en\\')(\\'Hola, \u00bfc\u00f3mo est\u00e1s?\\')\", \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'newssyscomb2009-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 30.6, \\'chr-F\\': 0.57}}, {\\'name\\': \\'news-test2008-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 27.9, \\'chr-F\\': 0.553}}, {\\'name\\': \\'newstest2009-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 30.4, \\'chr-F\\': 0.572}}, {\\'name\\': \\'newstest2010-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 36.1, \\'chr-F\\': 0.614}}, {\\'name\\': \\'newstest2011-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 34.2, \\'chr-F\\': 0.599}}, {\\'name\\': \\'newstest2012-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 37.9, \\'chr-F\\': 0.624}}, {\\'name\\': \\'newstest2013-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 35.3, \\'chr-F\\': 0.609}}, {\\'name\\': \\'Tatoeba-test.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 59.6, \\'chr-F\\': 0.739}}]}, \\'description\\': \\'Helsinki-NLP/opus-mt-es-en is a machine translation model trained to translate from Spanish to English using the Hugging Face Transformers library. The model is based on the Marian framework and was trained on the OPUS dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 596, "text": " We need a classifier that can sort movie synopses in German into three categories: crime, tragedy, and theft.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Zero-Shot Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Zero-Shot Classification\\', \\'api_name\\': \"AutoModelForSequenceClassification.from_pretrained(\\'Sahajtomar/German_Zeroshot\\')\", \\'api_call\\': \\'classifier(sequence, candidate_labels, hypothesis_template=hypothesis_template)\\', \\'api_arguments\\': {\\'sequence\\': \\'string\\', \\'candidate_labels\\': \\'list of strings\\', \\'hypothesis_template\\': \\'string\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'from transformers import pipeline\\\\nclassifier = pipeline(zero-shot-classification, model=Sahajtomar/German_Zeroshot)\\\\nsequence = Letzte Woche gab es einen Selbstmord in einer nahe gelegenen kolonie\\\\ncandidate_labels = [Verbrechen,Trag\u00f6die,Stehlen]\\\\nhypothesis_template = In deisem geht es um {}. ## Since monolingual model,its sensitive to hypothesis template. This can be experimented\\\\nclassifier(sequence, candidate_labels, hypothesis_template=hypothesis_template)\\', \\'performance\\': {\\'dataset\\': {\\'XNLI DEV (german)\\': {\\'accuracy\\': 85.5}, \\'XNLI TEST (german)\\': {\\'accuracy\\': 83.6}}}, \\'description\\': \\'This model has GBERT Large as base model and fine-tuned it on xnli de dataset. The default hypothesis template is in English: This text is {}. While using this model, change it to In deisem geht es um {}. or something different. While inferencing through huggingface api may give poor results as it uses by default english template. Since model is monolingual and not multilingual, hypothesis template needs to be changed accordingly.\\'}', metadata={})]", "category": "generic"}
{"question_id": 597, "text": " Write an abstract about the impacts of social media on mental health, summarizing key findings from previous studies.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'moussaKam/barthez-orangesum-abstract\\', \\'api_call\\': \"BarthezModel.from_pretrained(\\'moussaKam/barthez-orangesum-abstract\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'orangeSum\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Barthez model finetuned on orangeSum for abstract generation in French language\\'}', metadata={})]", "category": "generic"}
{"question_id": 598, "text": " Help me to translate a Spanish text to English. The Spanish text is: \\\"Lo siento, pero no puedo ir a la reuni\\u00f3n debido a una emergencia personal. Avisar\\u00e9 al equipo y nos pondremos en contacto para reprogramar la reuni\\u00f3n.\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Translation\\', \\'api_name\\': \\'Helsinki-NLP/opus-mt-en-es\\', \\'api_call\\': \"pipeline(\\'translation_en_to_es\\', model=\\'Helsinki-NLP/opus-mt-en-es\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"from transformers import pipeline\\\\ntranslation = pipeline(\\'translation_en_to_es\\', model=\\'Helsinki-NLP/opus-mt-en-es\\')\\\\ntranslated_text = translation(\\'Hello, how are you?\\')\", \\'performance\\': {\\'dataset\\': \\'Tatoeba-test.eng.spa\\', \\'accuracy\\': 54.9}, \\'description\\': \\'This model is a translation model from English to Spanish using the Hugging Face Transformers library. It is based on the Marian framework and trained on the OPUS dataset. The model achieves a BLEU score of 54.9 on the Tatoeba test set.\\'}', metadata={})]", "category": "generic"}
{"question_id": 599, "text": " I'm building a chatbot and I need it to respond to user questions based on the persona of an elderly person.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'hyunwoongko/blenderbot-9B\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'hyunwoongko/blenderbot-9B\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'Input a message to start chatting with hyunwoongko/blenderbot-9B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high-performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi-turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 600, "text": " Design a chatbot that can have conversations about a variety of topics with the users.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'hyunwoongko/blenderbot-9B\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'hyunwoongko/blenderbot-9B\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'Input a message to start chatting with hyunwoongko/blenderbot-9B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high-performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi-turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 601, "text": " We have a new platform that offers various services related to digital art, and we want it to be able to produce creative story ideas based on a short description. Can you help me providing suggestions?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'it5-base-news-summarization\\', \\'api_call\\': \"pipeline(\\'summarization\\', model=\\'it5/it5-base-news-summarization\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'newsum(Dal 31 maggio \u00e8 infine partita la piattaforma ITsART, a pi\u00f9 di un anno da quando \u2013 durante il primo lockdown \u2013 il ministro della Cultura Dario Franceschini ne aveva parlato come di \u00abuna sorta di Netflix della cultura\u00bb, pensata per \u00aboffrire a tutto il mondo la cultura italiana a pagamento\u00bb. \u00c8 presto per dare giudizi definitivi sulla piattaforma, e di certo sar\u00e0 difficile farlo anche pi\u00f9 avanti senza numeri precisi. Al momento, l\u2019unica cosa che si pu\u00f2 fare \u00e8 guardare com\u2019\u00e8 fatto il sito, contare quanti contenuti ci sono (circa 700 \u201ctitoli\u201d, tra film, documentari, spettacoli teatrali e musicali e altri eventi) e provare a dare un giudizio sul loro valore e sulla loro variet\u00e0. Intanto, una cosa notata da pi\u00f9 parti \u00e8 che diversi contenuti di ITsART sono a pagamento sulla piattaforma sebbene altrove, per esempio su RaiPlay, siano invece disponibili gratuitamente.)\\', \\'performance\\': {\\'dataset\\': \\'NewsSum-IT\\', \\'accuracy\\': {\\'Rouge1\\': 0.339, \\'Rouge2\\': 0.16, \\'RougeL\\': 0.263}}, \\'description\\': \\'IT5 Base model fine-tuned on news summarization on the Fanpage and Il Post corpora for Italian Language Understanding and Generation.\\'}', metadata={})]", "category": "generic"}
{"question_id": 602, "text": " We are creating a chatbot to answer customer queries on our website. Please provide a way to generate human-like text.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'hyunwoongko/blenderbot-9B\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'hyunwoongko/blenderbot-9B\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'Input a message to start chatting with hyunwoongko/blenderbot-9B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high-performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi-turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 603, "text": " To improve customer satisfaction, we want to automate the creation of written explanations that sound more natural and less robotic. Can you suggest a way to generate text that sounds conscious and alive?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Paraphrasing\\', \\'api_name\\': \\'prithivida/parrot_paraphraser_on_T5\\', \\'api_call\\': \"Parrot(model_tag=\\'prithivida/parrot_paraphraser_on_T5\\', use_gpu=False)\", \\'api_arguments\\': [\\'input_phrase\\', \\'diversity_ranker\\', \\'do_diverse\\', \\'max_return_phrases\\', \\'max_length\\', \\'adequacy_threshold\\', \\'fluency_threshold\\'], \\'python_environment_requirements\\': [\\'torch\\', \\'transformers\\'], \\'example_code\\': \\'from parrot import Parrot\\\\nimport torch\\\\nimport warnings\\\\nwarnings.filterwarnings(ignore)\\\\n\\\\nparrot = Parrot(model_tag=prithivida/parrot_paraphraser_on_T5, use_gpu=False)\\\\nphrases = [Can you recommed some upscale restaurants in Newyork?,\\\\n What are the famous places we should not miss in Russia?\\\\n]\\\\nfor phrase in phrases:\\\\n print(-*100)\\\\n print(Input_phrase: , phrase)\\\\n print(-*100)\\\\n para_phrases = parrot.augment(input_phrase=phrase)\\\\n for para_phrase in para_phrases:\\\\n  print(para_phrase)\\', \\'performance\\': {\\'dataset\\': \\'Not mentioned\\', \\'accuracy\\': \\'Not mentioned\\'}, \\'description\\': \\'Parrot is a paraphrase based utterance augmentation framework purpose built to accelerate training NLU models. It offers knobs to control Adequacy, Fluency, and Diversity as per your needs. It mainly focuses on augmenting texts typed-into or spoken-to conversational interfaces for building robust NLU models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 604, "text": " Our company is growing and expanding into multiple countries. We need a solution for translating English sentences to German to ensure effective communication.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'sshleifer/tiny-marian-en-de\\', \\'api_call\\': \"pipeline(\\'translation_en_to_de\\', model=\\'sshleifer/tiny-marian-en-de\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A tiny English to German translation model using the Marian framework in Hugging Face Transformers.\\'}', metadata={})]", "category": "generic"}
{"question_id": 605, "text": " We need a solution for creating Polish subtitles for YouTube videos in Spanish. The AI should provide the translation.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Translation\\', \\'api_name\\': \\'Helsinki-NLP/opus-mt-es-en\\', \\'api_call\\': \"pipeline(\\'translation_es_to_en\\', model=\\'Helsinki-NLP/opus-mt-es-en\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\ntranslation = pipeline(\\'translation_es_to_en\\', model=\\'Helsinki-NLP/opus-mt-es-en\\')(\\'Hola, \u00bfc\u00f3mo est\u00e1s?\\')\", \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'newssyscomb2009-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 30.6, \\'chr-F\\': 0.57}}, {\\'name\\': \\'news-test2008-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 27.9, \\'chr-F\\': 0.553}}, {\\'name\\': \\'newstest2009-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 30.4, \\'chr-F\\': 0.572}}, {\\'name\\': \\'newstest2010-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 36.1, \\'chr-F\\': 0.614}}, {\\'name\\': \\'newstest2011-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 34.2, \\'chr-F\\': 0.599}}, {\\'name\\': \\'newstest2012-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 37.9, \\'chr-F\\': 0.624}}, {\\'name\\': \\'newstest2013-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 35.3, \\'chr-F\\': 0.609}}, {\\'name\\': \\'Tatoeba-test.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 59.6, \\'chr-F\\': 0.739}}]}, \\'description\\': \\'Helsinki-NLP/opus-mt-es-en is a machine translation model trained to translate from Spanish to English using the Hugging Face Transformers library. The model is based on the Marian framework and was trained on the OPUS dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 606, "text": " I am an English teacher looking to brainstorm synonyms for the word \\\"happy.\\\" Can a model help me generate similar words?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Grammar Correction\\', \\'api_name\\': \\'vennify/t5-base-grammar-correction\\', \\'api_call\\': \"HappyTextToText(\\'T5\\', \\'vennify/t5-base-grammar-correction\\')\", \\'api_arguments\\': {\\'num_beams\\': 5, \\'min_length\\': 1}, \\'python_environment_requirements\\': {\\'package\\': \\'happytransformer\\', \\'installation\\': \\'pip install happytransformer\\'}, \\'example_code\\': \\'from happytransformer import HappyTextToText, TTSettings\\\\nhappy_tt = HappyTextToText(T5, vennify/t5-base-grammar-correction)\\\\nargs = TTSettings(num_beams=5, min_length=1)\\\\nresult = happy_tt.generate_text(grammar: This sentences has has bads grammar., args=args)\\\\nprint(result.text)\\', \\'performance\\': {\\'dataset\\': \\'jfleg\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This model generates a revised version of inputted text with the goal of containing fewer grammatical errors. It was trained with Happy Transformer using a dataset called JFLEG.\\'}', metadata={})]", "category": "generic"}
{"question_id": 607, "text": " Generate a sentence which will complete the following: \\\"Hello, I'm a ...\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Conversational\\', \\'api_name\\': \\'Pi3141/DialoGPT-medium-elon-3\\', \\'api_call\\': \"pipeline(\\'text-generation\\', model=\\'Pi3141/DialoGPT-medium-elon-3\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'Input a message to start chatting with Pi3141/DialoGPT-medium-elon-3.\\', \\'performance\\': {\\'dataset\\': \\'Twitter tweets by Elon Musk\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'DialoGPT model that talks like Elon Musk, trained on Twitter tweets by Elon Musk. This model will spew meaningless shit about 40% of the time. Trained on 8 epochs. But with a larger dataset this time. The AI can now use more emojis, I think.\\'}', metadata={})]", "category": "generic"}
{"question_id": 608, "text": " Can you help me find the best model to fill in the gap in my legal document? I want a smaller model with higher efficiency but maintains a high level of accuracy.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Fill-Mask\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Fill-Mask\\', \\'api_name\\': \\'nlpaueb/legal-bert-small-uncased\\', \\'api_call\\': \"AutoModel.from_pretrained(\\'nlpaueb/legal-bert-small-uncased\\')\", \\'api_arguments\\': {\\'pretrained_model_name_or_path\\': \\'nlpaueb/legal-bert-small-uncased\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'AutoTokenizer, AutoModel\\'}, \\'example_code\\': \"from transformers import AutoTokenizer, AutoModel\\\\ntokenizer = AutoTokenizer.from_pretrained(\\'nlpaueb/legal-bert-small-uncased\\')\\\\nmodel = AutoModel.from_pretrained(\\'nlpaueb/legal-bert-small-uncased\\')\", \\'performance\\': {\\'dataset\\': \\'Legal Corpora\\', \\'accuracy\\': \\'Comparable to larger models\\'}, \\'description\\': \\'LEGAL-BERT is a family of BERT models for the legal domain, intended to assist legal NLP research, computational law, and legal technology applications. This is the light-weight version of BERT-BASE (33% the size of BERT-BASE) pre-trained from scratch on legal data, which achieves comparable performance to larger models, while being much more efficient (approximately 4 times faster) with a smaller environmental footprint.\\'}', metadata={})]", "category": "generic"}
{"question_id": 609, "text": " I am learning English literature. I plan to build a system that has a huge database of English sentences and keeps the important ones.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Token Classification\\', \\'framework\\': \\'Flair\\', \\'functionality\\': \\'Named Entity Recognition\\', \\'api_name\\': \\'flair/ner-english-ontonotes-large\\', \\'api_call\\': \"SequenceTagger.load(\\'flair/ner-english-ontonotes-large\\')\", \\'api_arguments\\': [\\'sentence\\'], \\'python_environment_requirements\\': [\\'flair\\'], \\'example_code\\': \"from flair.data import Sentence\\\\nfrom flair.models import SequenceTagger\\\\n\\\\n# load tagger\\\\ntagger = SequenceTagger.load(\\'flair/ner-english-ontonotes-large\\')\\\\n\\\\n# make example sentence\\\\nsentence = Sentence(\\'On September 1st George won 1 dollar while watching Game of Thrones.\\')\\\\n\\\\n# predict NER tags\\\\ntagger.predict(sentence)\\\\n\\\\n# print sentence\\\\nprint(sentence)\\\\n\\\\n# print predicted NER spans\\\\nprint(\\'The following NER tags are found:\\')\\\\n\\\\n# iterate over entities and print\\\\nfor entity in sentence.get_spans(\\'ner\\'):\\\\n    print(entity)\", \\'performance\\': {\\'dataset\\': \\'Ontonotes\\', \\'accuracy\\': 90.93}, \\'description\\': \\'English NER in Flair (Ontonotes large model). This is the large 18-class NER model for English that ships with Flair. It predicts 18 tags such as cardinal value, date value, event name, building name, geo-political entity, language name, law name, location name, money name, affiliation, ordinal value, organization name, percent value, person name, product name, quantity value, time value, and name of work of art. The model is based on document-level XLM-R embeddings and FLERT.\\'}', metadata={})]", "category": "generic"}
{"question_id": 610, "text": " Our company specializes in providing information on similar topics. We want to find similar sentences in a text document.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Sentence Similarity\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Sentence Embeddings\\', \\'api_name\\': \\'sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2\\', \\'api_call\\': \"SentenceTransformer(\\'sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2\\')\", \\'api_arguments\\': [\\'sentences\\'], \\'python_environment_requirements\\': \\'pip install -U sentence-transformers\\', \\'example_code\\': \"from sentence_transformers import SentenceTransformer\\\\nsentences = [This is an example sentence, Each sentence is converted]\\\\nmodel = SentenceTransformer(\\'sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2\\')\\\\nembeddings = model.encode(sentences)\\\\nprint(embeddings)\", \\'performance\\': {\\'dataset\\': \\'https://seb.sbert.net\\', \\'accuracy\\': \\'Automated evaluation\\'}, \\'description\\': \\'This is a sentence-transformers model: It maps sentences & paragraphs to a 384 dimensional dense vector space and can be used for tasks like clustering or semantic search.\\'}', metadata={})]", "category": "generic"}
{"question_id": 611, "text": " The company is building an AI chatbot, and we need a way to process users' questions and tell if two questions are similar.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'hyunwoongko/blenderbot-9B\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'hyunwoongko/blenderbot-9B\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'Input a message to start chatting with hyunwoongko/blenderbot-9B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high-performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi-turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 612, "text": " I am a CEO, and I want to create an audio announcement that will be played to my customers. The text announcement is: \\\"Dear valued customers, we are glad to announce a 30% discount on our special offer. Hurry up!\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'CTranslate2\\', \\'functionality\\': \\'Automatic Speech Recognition\\', \\'api_name\\': \\'guillaumekln/faster-whisper-large-v2\\', \\'api_call\\': \"WhisperModel(\\'large-v2\\')\", \\'api_arguments\\': [\\'audio.mp3\\'], \\'python_environment_requirements\\': [\\'faster_whisper\\'], \\'example_code\\': \\'from faster_whisper import WhisperModel\\\\nmodel = WhisperModel(large-v2)\\\\nsegments, info = model.transcribe(audio.mp3)\\\\nfor segment in segments:\\\\n print([%.2fs -&gt; %.2fs] %s % (segment.start, segment.end, segment.text))\\', \\'performance\\': {\\'dataset\\': \\'99 languages\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Whisper large-v2 model for CTranslate2. This model can be used in CTranslate2 or projets based on CTranslate2 such as faster-whisper.\\'}', metadata={})]", "category": "generic"}
{"question_id": 613, "text": " Our user would like to develop an audiobook using a Text-to-Speech API. Convert the text of a book into an audio file.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'lakahaga/novel_reading_tts\\', \\'api_call\\': \"AutoModelForTTS.from_pretrained(\\'lakahaga/novel_reading_tts\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"inputs = processor(text, return_tensors=\\'pt\\'); generated_audio = model.generate(**inputs);\", \\'performance\\': {\\'dataset\\': \\'novelspeech\\', \\'accuracy\\': None}, \\'description\\': \\'This model was trained by lakahaga using novelspeech recipe in espnet. It is designed for Korean text-to-speech tasks.\\'}', metadata={})]", "category": "generic"}
{"question_id": 614, "text": " In our language app, we want to include text to speech functionality for Hokkien, a dialect of Chinese, using the TAT-TTS dataset.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'Fairseq\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'unit_hifigan_HK_layer12.km2500_frame_TAT-TTS\\', \\'api_call\\': \"load_model_ensemble_and_task_from_hf_hub(\\'facebook/unit_hifigan_HK_layer12.km2500_frame_TAT-TT\\')\", \\'api_arguments\\': {\\'unit\\': \\'Text input for the TTS model\\'}, \\'python_environment_requirements\\': [\\'fairseq\\', \\'huggingface_hub\\', \\'torchaudio\\'], \\'example_code\\': \"import json\\\\nimport os\\\\nfrom pathlib import Path\\\\nimport IPython.display as ipd\\\\nfrom fairseq import hub_utils\\\\nfrom fairseq.checkpoint_utils import load_model_ensemble_and_task_from_hf_hub\\\\nfrom fairseq.models.speech_to_text.hub_interface import S2THubInterface\\\\nfrom fairseq.models.text_to_speech import CodeHiFiGANVocoder\\\\nfrom fairseq.models.text_to_speech.hub_interface import VocoderHubInterface\\\\nfrom huggingface_hub import snapshot_download\\\\nimport torchaudio\\\\ncache_dir = os.getenv(HUGGINGFACE_HUB_CACHE)\\\\nlibrary_name = fairseq\\\\ncache_dir = (\\\\n cache_dir or (Path.home() / .cache / library_name).as_posix()\\\\n)\\\\ncache_dir = snapshot_download(\\\\n ffacebook/unit_hifigan_HK_layer12.km2500_frame_TAT-TTS, cache_dir=cache_dir, library_name=library_name\\\\n)\\\\nx = hub_utils.from_pretrained(\\\\n cache_dir,\\\\n model.pt,\\\\n .,\\\\n archive_map=CodeHiFiGANVocoder.hub_models(),\\\\n config_yaml=config.json,\\\\n fp16=False,\\\\n is_vocoder=True,\\\\n)\\\\nwith open(f{x[\\'args\\'][\\'data\\']}/config.json) as f:\\\\n vocoder_cfg = json.load(f)\\\\nassert (\\\\n len(x[args][model_path]) == 1\\\\n), Too many vocoder models in the input\\\\nvocoder = CodeHiFiGANVocoder(x[args][model_path][0], vocoder_cfg)\\\\ntts_model = VocoderHubInterface(vocoder_cfg, vocoder)\\\\ntts_sample = tts_model.get_model_input(unit)\\\\nwav, sr = tts_model.get_prediction(tts_sample)\\\\nipd.Audio(wav, rate=sr)\", \\'performance\\': {\\'dataset\\': \\'TAT-TTS\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Hokkien unit HiFiGAN based vocoder from fairseq. Trained with TAT-TTS data with 4 speakers in Taiwanese Hokkien accent.\\'}', metadata={})]", "category": "generic"}
{"question_id": 615, "text": " Develop an application to transcribe audio files with punctuation marks for a podcast platform.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Audio-to-Audio\\', \\'framework\\': \\'Fairseq\\', \\'functionality\\': \\'speech-to-speech-translation\\', \\'api_name\\': \\'facebook/textless_sm_ro_en\\', \\'api_call\\': \"pipeline(\\'audio-to-audio\\', model=\\'facebook/textless_sm_ro_en\\')\", \\'api_arguments\\': \\'audio file or recording\\', \\'python_environment_requirements\\': \\'fairseq, huggingface_hub\\', \\'example_code\\': \\'https://huggingface.co/facebook/textless_sm_cs_en\\', \\'performance\\': {\\'dataset\\': \\'unknown\\', \\'accuracy\\': \\'unknown\\'}, \\'description\\': \\'A speech-to-speech translation model for Romanian to English developed by Facebook AI\\'}', metadata={})]", "category": "generic"}
{"question_id": 616, "text": " Our company has an audio archive of Chinese podcasts, and we want to create transcripts for them.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'kan-bayashi_csmsc_tts_train_tacotron2_raw_phn_pypinyin_g2p_phone_train.loss.best\\', \\'api_call\\': \"Text2Speech.from_pretrained(\\'espnet/kan-bayashi_csmsc_tts_train_tacotron2_raw_phn_pypinyin_g2p_phone_train.loss.best\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'torch\\', \\'espnet_model_zoo\\'], \\'example_code\\': \\'import soundfile\\\\nfrom espnet2.bin.tts_inference import Text2Speech\\\\ntext2speech = Text2Speech.from_pretrained(espnet/kan-bayashi_csmsc_tts_train_tacotron2_raw_phn_pypinyin_g2p_phone_train.loss.best)\\\\ntext = \u6625\u6c5f\u6f6e\u6c34\u8fde\u6d77\u5e73\uff0c\u6d77\u4e0a\u660e\u6708\u5171\u6f6e\u751f\\\\nspeech = text2speech(text)[wav]\\\\nsoundfile.write(out.wav, speech.numpy(), text2speech.fs, PCM_16)\\', \\'performance\\': {\\'dataset\\': \\'csmsc\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'A pre-trained Text-to-Speech model for Chinese language using ESPnet framework. It can be used to convert text input into speech output in Chinese.\\'}', metadata={})]", "category": "generic"}
{"question_id": 617, "text": " We are a podcast platform. We have a library of multiple podcasts that are mixed with background noise. We need to enhance the podcast audio quality to improve the user experience on our platform.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'SYSPIN/Marathi_Male_TTS\\', \\'api_call\\': \"api.load(\\'ESPnet/espnet_model_zoo:SYSPIN/Marathi_Male_TTS\\').\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'huggingface_hub\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Marathi Male Text-to-Speech model using ESPnet framework.\\'}', metadata={})]", "category": "generic"}
{"question_id": 618, "text": " A podcast producer is looking to improve the quality of their audio files by removing background noise. What can they do?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Voice Activity Detection, Speech-to-Noise Ratio, and C50 Room Acoustics Estimation\\', \\'api_name\\': \\'pyannote/brouhaha\\', \\'api_call\\': \"Model.from_pretrained(\\'pyannote/brouhaha\\', use_auth_token=\\'ACCESS_TOKEN_GOES_HERE\\')\", \\'api_arguments\\': [\\'audio.wav\\'], \\'python_environment_requirements\\': [\\'pyannote-audio\\', \\'brouhaha-vad\\'], \\'example_code\\': [\\'from pyannote.audio import Model\\', \\'model = Model.from_pretrained(pyannote/brouhaha, use_auth_token=ACCESS_TOKEN_GOES_HERE)\\', \\'from pyannote.audio import Inference\\', \\'inference = Inference(model)\\', \\'output = inference(audio.wav)\\', \\'for frame, (vad, snr, c50) in output:\\', \\'  t = frame.middle\\', \\'  print(f{t:8.3f} vad={100*vad:.0f}% snr={snr:.0f} c50={c50:.0f})\\'], \\'performance\\': {\\'dataset\\': \\'LibriSpeech, AudioSet, EchoThief, MIT-Acoustical-Reverberation-Scene\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Brouhaha is a joint voice activity detection, speech-to-noise ratio, and C50 room acoustics estimation model. It is based on the PyTorch framework and uses the pyannote.audio library.\\'}', metadata={})]", "category": "generic"}
{"question_id": 619, "text": " I work at a call center, and I need to convert a recorded customer call into an audio file with a different voice without changing the content.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'CTranslate2\\', \\'functionality\\': \\'Automatic Speech Recognition\\', \\'api_name\\': \\'guillaumekln/faster-whisper-large-v2\\', \\'api_call\\': \"WhisperModel(\\'large-v2\\')\", \\'api_arguments\\': [\\'audio.mp3\\'], \\'python_environment_requirements\\': [\\'faster_whisper\\'], \\'example_code\\': \\'from faster_whisper import WhisperModel\\\\nmodel = WhisperModel(large-v2)\\\\nsegments, info = model.transcribe(audio.mp3)\\\\nfor segment in segments:\\\\n print([%.2fs -&gt; %.2fs] %s % (segment.start, segment.end, segment.text))\\', \\'performance\\': {\\'dataset\\': \\'99 languages\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Whisper large-v2 model for CTranslate2. This model can be used in CTranslate2 or projets based on CTranslate2 such as faster-whisper.\\'}', metadata={})]", "category": "generic"}
{"question_id": 620, "text": " In a customer support system, we have received a recorded audio file from a customer complaining about an issue in Czech language. Translate it to English language preserving the audio format.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'CTranslate2\\', \\'functionality\\': \\'Automatic Speech Recognition\\', \\'api_name\\': \\'guillaumekln/faster-whisper-large-v2\\', \\'api_call\\': \"WhisperModel(\\'large-v2\\')\", \\'api_arguments\\': [\\'audio.mp3\\'], \\'python_environment_requirements\\': [\\'faster_whisper\\'], \\'example_code\\': \\'from faster_whisper import WhisperModel\\\\nmodel = WhisperModel(large-v2)\\\\nsegments, info = model.transcribe(audio.mp3)\\\\nfor segment in segments:\\\\n print([%.2fs -&gt; %.2fs] %s % (segment.start, segment.end, segment.text))\\', \\'performance\\': {\\'dataset\\': \\'99 languages\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Whisper large-v2 model for CTranslate2. This model can be used in CTranslate2 or projets based on CTranslate2 such as faster-whisper.\\'}', metadata={})]", "category": "generic"}
{"question_id": 621, "text": " We are a podcast editing company aiming at separating speakers from a recorded audio.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'pyannote.audio\\', \\'functionality\\': \\'Speaker diarization\\', \\'api_name\\': \\'johnislarry/cloned-pyannote-speaker-diarization-endpoint\\', \\'api_call\\': \"Pipeline.from_pretrained(\\'pyannote/speaker-diarization@2.1\\',use_auth_token=\\'ACCESS_TOKEN_GOES_HERE\\')\", \\'api_arguments\\': [\\'num_speakers\\', \\'min_speakers\\', \\'max_speakers\\', \\'segmentation_onset\\'], \\'python_environment_requirements\\': \\'pyannote.audio 2.0\\', \\'example_code\\': {\\'load_pipeline\\': \\'from pyannote.audio import Pipeline\\\\npipeline = Pipeline.from_pretrained(pyannote/speaker-diarization@2022.07)\\', \\'apply_pipeline\\': \\'diarization = pipeline(audio.wav)\\', \\'save_output\\': \\'with open(audio.rttm, w) as rttm:\\\\n  diarization.write_rttm(rttm)\\'}, \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'AISHELL-4\\', \\'accuracy\\': {\\'DER%\\': 14.61, \\'FA%\\': 3.31, \\'Miss%\\': 4.35, \\'Conf%\\': 6.95}}, {\\'name\\': \\'AMI Mix-Headset only_words\\', \\'accuracy\\': {\\'DER%\\': 18.21, \\'FA%\\': 3.28, \\'Miss%\\': 11.07, \\'Conf%\\': 3.87}}, {\\'name\\': \\'AMI Array1-01 only_words\\', \\'accuracy\\': {\\'DER%\\': 29.0, \\'FA%\\': 2.71, \\'Miss%\\': 21.61, \\'Conf%\\': 4.68}}, {\\'name\\': \\'CALLHOME Part2\\', \\'accuracy\\': {\\'DER%\\': 30.24, \\'FA%\\': 3.71, \\'Miss%\\': 16.86, \\'Conf%\\': 9.66}}, {\\'name\\': \\'DIHARD 3 Full\\', \\'accuracy\\': {\\'DER%\\': 20.99, \\'FA%\\': 4.25, \\'Miss%\\': 10.74, \\'Conf%\\': 6.0}}, {\\'name\\': \\'REPERE Phase 2\\', \\'accuracy\\': {\\'DER%\\': 12.62, \\'FA%\\': 1.55, \\'Miss%\\': 3.3, \\'Conf%\\': 7.76}}, {\\'name\\': \\'VoxConverse v0.0.2\\', \\'accuracy\\': {\\'DER%\\': 12.76, \\'FA%\\': 3.45, \\'Miss%\\': 3.85, \\'Conf%\\': 5.46}}]}, \\'description\\': \\'This API provides speaker diarization functionality using the pyannote.audio framework. It is capable of processing audio files and outputting speaker diarization results in RTTM format. The API supports providing the number of speakers, minimum and maximum number of speakers, and adjusting the segmentation onset threshold.\\'}', metadata={})]", "category": "generic"}
{"question_id": 622, "text": " We are building a virtual assistant which can translate English speech input to Hokkien in real-time.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'Fairseq\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'unit_hifigan_HK_layer12.km2500_frame_TAT-TTS\\', \\'api_call\\': \"load_model_ensemble_and_task_from_hf_hub(\\'facebook/unit_hifigan_HK_layer12.km2500_frame_TAT-TT\\')\", \\'api_arguments\\': {\\'unit\\': \\'Text input for the TTS model\\'}, \\'python_environment_requirements\\': [\\'fairseq\\', \\'huggingface_hub\\', \\'torchaudio\\'], \\'example_code\\': \"import json\\\\nimport os\\\\nfrom pathlib import Path\\\\nimport IPython.display as ipd\\\\nfrom fairseq import hub_utils\\\\nfrom fairseq.checkpoint_utils import load_model_ensemble_and_task_from_hf_hub\\\\nfrom fairseq.models.speech_to_text.hub_interface import S2THubInterface\\\\nfrom fairseq.models.text_to_speech import CodeHiFiGANVocoder\\\\nfrom fairseq.models.text_to_speech.hub_interface import VocoderHubInterface\\\\nfrom huggingface_hub import snapshot_download\\\\nimport torchaudio\\\\ncache_dir = os.getenv(HUGGINGFACE_HUB_CACHE)\\\\nlibrary_name = fairseq\\\\ncache_dir = (\\\\n cache_dir or (Path.home() / .cache / library_name).as_posix()\\\\n)\\\\ncache_dir = snapshot_download(\\\\n ffacebook/unit_hifigan_HK_layer12.km2500_frame_TAT-TTS, cache_dir=cache_dir, library_name=library_name\\\\n)\\\\nx = hub_utils.from_pretrained(\\\\n cache_dir,\\\\n model.pt,\\\\n .,\\\\n archive_map=CodeHiFiGANVocoder.hub_models(),\\\\n config_yaml=config.json,\\\\n fp16=False,\\\\n is_vocoder=True,\\\\n)\\\\nwith open(f{x[\\'args\\'][\\'data\\']}/config.json) as f:\\\\n vocoder_cfg = json.load(f)\\\\nassert (\\\\n len(x[args][model_path]) == 1\\\\n), Too many vocoder models in the input\\\\nvocoder = CodeHiFiGANVocoder(x[args][model_path][0], vocoder_cfg)\\\\ntts_model = VocoderHubInterface(vocoder_cfg, vocoder)\\\\ntts_sample = tts_model.get_model_input(unit)\\\\nwav, sr = tts_model.get_prediction(tts_sample)\\\\nipd.Audio(wav, rate=sr)\", \\'performance\\': {\\'dataset\\': \\'TAT-TTS\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Hokkien unit HiFiGAN based vocoder from fairseq. Trained with TAT-TTS data with 4 speakers in Taiwanese Hokkien accent.\\'}', metadata={})]", "category": "generic"}
{"question_id": 623, "text": " A language service wants to incorporate a speech-to-speech translation feature that assists users in translating Hokkien to English on an audio file.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'Fairseq\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'unit_hifigan_HK_layer12.km2500_frame_TAT-TTS\\', \\'api_call\\': \"load_model_ensemble_and_task_from_hf_hub(\\'facebook/unit_hifigan_HK_layer12.km2500_frame_TAT-TT\\')\", \\'api_arguments\\': {\\'unit\\': \\'Text input for the TTS model\\'}, \\'python_environment_requirements\\': [\\'fairseq\\', \\'huggingface_hub\\', \\'torchaudio\\'], \\'example_code\\': \"import json\\\\nimport os\\\\nfrom pathlib import Path\\\\nimport IPython.display as ipd\\\\nfrom fairseq import hub_utils\\\\nfrom fairseq.checkpoint_utils import load_model_ensemble_and_task_from_hf_hub\\\\nfrom fairseq.models.speech_to_text.hub_interface import S2THubInterface\\\\nfrom fairseq.models.text_to_speech import CodeHiFiGANVocoder\\\\nfrom fairseq.models.text_to_speech.hub_interface import VocoderHubInterface\\\\nfrom huggingface_hub import snapshot_download\\\\nimport torchaudio\\\\ncache_dir = os.getenv(HUGGINGFACE_HUB_CACHE)\\\\nlibrary_name = fairseq\\\\ncache_dir = (\\\\n cache_dir or (Path.home() / .cache / library_name).as_posix()\\\\n)\\\\ncache_dir = snapshot_download(\\\\n ffacebook/unit_hifigan_HK_layer12.km2500_frame_TAT-TTS, cache_dir=cache_dir, library_name=library_name\\\\n)\\\\nx = hub_utils.from_pretrained(\\\\n cache_dir,\\\\n model.pt,\\\\n .,\\\\n archive_map=CodeHiFiGANVocoder.hub_models(),\\\\n config_yaml=config.json,\\\\n fp16=False,\\\\n is_vocoder=True,\\\\n)\\\\nwith open(f{x[\\'args\\'][\\'data\\']}/config.json) as f:\\\\n vocoder_cfg = json.load(f)\\\\nassert (\\\\n len(x[args][model_path]) == 1\\\\n), Too many vocoder models in the input\\\\nvocoder = CodeHiFiGANVocoder(x[args][model_path][0], vocoder_cfg)\\\\ntts_model = VocoderHubInterface(vocoder_cfg, vocoder)\\\\ntts_sample = tts_model.get_model_input(unit)\\\\nwav, sr = tts_model.get_prediction(tts_sample)\\\\nipd.Audio(wav, rate=sr)\", \\'performance\\': {\\'dataset\\': \\'TAT-TTS\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Hokkien unit HiFiGAN based vocoder from fairseq. Trained with TAT-TTS data with 4 speakers in Taiwanese Hokkien accent.\\'}', metadata={})]", "category": "generic"}
{"question_id": 624, "text": " Our company is building an interactive platform for children to learn random trivia quickly. We need a model to understand what is being spoken and what category it belongs to.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Paraphrase-based utterance augmentation\\', \\'api_name\\': \\'prithivida/parrot_fluency_model\\', \\'api_call\\': \"pipeline(\\'text-classification\\', model=\\'prithivida/parrot_fluency_model\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"parrot(\\'your input text\\')\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'Parrot is a paraphrase-based utterance augmentation framework purpose-built to accelerate training NLU models. A paraphrase framework is more than just a paraphrasing model.\\'}', metadata={})]", "category": "generic"}
{"question_id": 625, "text": " We are a company that develops voice assistants, and we need to verify a user's voice to authenticate them.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Voice Activity Detection\\', \\'api_name\\': \\'d4data/Indian-voice-cloning\\', \\'api_call\\': \"pipeline(\\'voice-activity-detection\\', model=\\'d4data/Indian-voice-cloning\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A model for detecting voice activity in Indian languages.\\'}', metadata={})]", "category": "generic"}
{"question_id": 626, "text": " We recently launched a customer service hotline, and we want to estimate the demographics of callers.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 627, "text": " Create a smart speaker that can recognize voice commands such as \\\"Turn on the lights,\\\" \\\"Play music,\\\" or \\\"Set a timer.\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'lakahaga/novel_reading_tts\\', \\'api_call\\': \"AutoModelForTTS.from_pretrained(\\'lakahaga/novel_reading_tts\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"inputs = processor(text, return_tensors=\\'pt\\'); generated_audio = model.generate(**inputs);\", \\'performance\\': {\\'dataset\\': \\'novelspeech\\', \\'accuracy\\': None}, \\'description\\': \\'This model was trained by lakahaga using novelspeech recipe in espnet. It is designed for Korean text-to-speech tasks.\\'}', metadata={})]", "category": "generic"}
{"question_id": 628, "text": " Our company's voice assistant needs to be able to detect voice activity in a conversation.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Voice Activity Detection\\', \\'api_name\\': \\'d4data/Indian-voice-cloning\\', \\'api_call\\': \"pipeline(\\'voice-activity-detection\\', model=\\'d4data/Indian-voice-cloning\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A model for detecting voice activity in Indian languages.\\'}', metadata={})]", "category": "generic"}
{"question_id": 629, "text": " Implement a conference call control program which can recognize if there are any interruptions among the speakers during the meeting.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'pyannote.audio\\', \\'functionality\\': \\'Speaker diarization\\', \\'api_name\\': \\'johnislarry/cloned-pyannote-speaker-diarization-endpoint\\', \\'api_call\\': \"Pipeline.from_pretrained(\\'pyannote/speaker-diarization@2.1\\',use_auth_token=\\'ACCESS_TOKEN_GOES_HERE\\')\", \\'api_arguments\\': [\\'num_speakers\\', \\'min_speakers\\', \\'max_speakers\\', \\'segmentation_onset\\'], \\'python_environment_requirements\\': \\'pyannote.audio 2.0\\', \\'example_code\\': {\\'load_pipeline\\': \\'from pyannote.audio import Pipeline\\\\npipeline = Pipeline.from_pretrained(pyannote/speaker-diarization@2022.07)\\', \\'apply_pipeline\\': \\'diarization = pipeline(audio.wav)\\', \\'save_output\\': \\'with open(audio.rttm, w) as rttm:\\\\n  diarization.write_rttm(rttm)\\'}, \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'AISHELL-4\\', \\'accuracy\\': {\\'DER%\\': 14.61, \\'FA%\\': 3.31, \\'Miss%\\': 4.35, \\'Conf%\\': 6.95}}, {\\'name\\': \\'AMI Mix-Headset only_words\\', \\'accuracy\\': {\\'DER%\\': 18.21, \\'FA%\\': 3.28, \\'Miss%\\': 11.07, \\'Conf%\\': 3.87}}, {\\'name\\': \\'AMI Array1-01 only_words\\', \\'accuracy\\': {\\'DER%\\': 29.0, \\'FA%\\': 2.71, \\'Miss%\\': 21.61, \\'Conf%\\': 4.68}}, {\\'name\\': \\'CALLHOME Part2\\', \\'accuracy\\': {\\'DER%\\': 30.24, \\'FA%\\': 3.71, \\'Miss%\\': 16.86, \\'Conf%\\': 9.66}}, {\\'name\\': \\'DIHARD 3 Full\\', \\'accuracy\\': {\\'DER%\\': 20.99, \\'FA%\\': 4.25, \\'Miss%\\': 10.74, \\'Conf%\\': 6.0}}, {\\'name\\': \\'REPERE Phase 2\\', \\'accuracy\\': {\\'DER%\\': 12.62, \\'FA%\\': 1.55, \\'Miss%\\': 3.3, \\'Conf%\\': 7.76}}, {\\'name\\': \\'VoxConverse v0.0.2\\', \\'accuracy\\': {\\'DER%\\': 12.76, \\'FA%\\': 3.45, \\'Miss%\\': 3.85, \\'Conf%\\': 5.46}}]}, \\'description\\': \\'This API provides speaker diarization functionality using the pyannote.audio framework. It is capable of processing audio files and outputting speaker diarization results in RTTM format. The API supports providing the number of speakers, minimum and maximum number of speakers, and adjusting the segmentation onset threshold.\\'}', metadata={})]", "category": "generic"}
{"question_id": 630, "text": " We are trying to create a solution for an HR department to predict whether a candidate would be a potential employee based on a list of background information.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Predicting Pokemon HP\\', \\'api_name\\': \\'julien-c/pokemon-predict-hp\\', \\'api_call\\': \"pipeline(\\'regression\\', model=\\'julien-c/pokemon-predict-hp\\')\", \\'api_arguments\\': \\'input_data\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'julien-c/kaggle-rounakbanik-pokemon\\', \\'accuracy\\': {\\'mean_absolute_error\\': 15.909, \\'model_loss\\': 647.605}}, \\'description\\': \\'A tabular regression model trained on the julien-c/kaggle-rounakbanik-pokemon dataset to predict the HP of Pokemon.\\'}', metadata={})]", "category": "generic"}
{"question_id": 631, "text": " I have just downloaded a dataset that measures CO2 Emissions. I want to classify this dataset using a pre-trained model.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Joblib\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'pachi107/autotrain-in-class-test-1780161764\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data.csv\\'], \\'python_environment_requirements\\': [\\'json\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(data.csv)\\\\ndata = data[features]\\\\ndata.columns = [feat_ + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'Validation Metrics\\', \\'accuracy\\': 0.974}, \\'description\\': \\'A binary classification model for predicting CO2 emissions based on tabular data. Trained using AutoTrain with a model ID of 1780161764.\\'}', metadata={})]", "category": "generic"}
{"question_id": 632, "text": " We are building an app that simulates Pokemon battles. Can you help us predict the HP of a Pokemon given its input attributes?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Predicting Pokemon HP\\', \\'api_name\\': \\'julien-c/pokemon-predict-hp\\', \\'api_call\\': \"pipeline(\\'regression\\', model=\\'julien-c/pokemon-predict-hp\\')\", \\'api_arguments\\': \\'input_data\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'julien-c/kaggle-rounakbanik-pokemon\\', \\'accuracy\\': {\\'mean_absolute_error\\': 15.909, \\'model_loss\\': 647.605}}, \\'description\\': \\'A tabular regression model trained on the julien-c/kaggle-rounakbanik-pokemon dataset to predict the HP of Pokemon.\\'}', metadata={})]", "category": "generic"}
{"question_id": 633, "text": " We are a content marketing agency and we are focusing on promoting our clients' products. We want to use a tool to generate interesting marketing messages.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'it5-base-news-summarization\\', \\'api_call\\': \"pipeline(\\'summarization\\', model=\\'it5/it5-base-news-summarization\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'newsum(Dal 31 maggio \u00e8 infine partita la piattaforma ITsART, a pi\u00f9 di un anno da quando \u2013 durante il primo lockdown \u2013 il ministro della Cultura Dario Franceschini ne aveva parlato come di \u00abuna sorta di Netflix della cultura\u00bb, pensata per \u00aboffrire a tutto il mondo la cultura italiana a pagamento\u00bb. \u00c8 presto per dare giudizi definitivi sulla piattaforma, e di certo sar\u00e0 difficile farlo anche pi\u00f9 avanti senza numeri precisi. Al momento, l\u2019unica cosa che si pu\u00f2 fare \u00e8 guardare com\u2019\u00e8 fatto il sito, contare quanti contenuti ci sono (circa 700 \u201ctitoli\u201d, tra film, documentari, spettacoli teatrali e musicali e altri eventi) e provare a dare un giudizio sul loro valore e sulla loro variet\u00e0. Intanto, una cosa notata da pi\u00f9 parti \u00e8 che diversi contenuti di ITsART sono a pagamento sulla piattaforma sebbene altrove, per esempio su RaiPlay, siano invece disponibili gratuitamente.)\\', \\'performance\\': {\\'dataset\\': \\'NewsSum-IT\\', \\'accuracy\\': {\\'Rouge1\\': 0.339, \\'Rouge2\\': 0.16, \\'RougeL\\': 0.263}}, \\'description\\': \\'IT5 Base model fine-tuned on news summarization on the Fanpage and Il Post corpora for Italian Language Understanding and Generation.\\'}', metadata={})]", "category": "generic"}
{"question_id": 634, "text": " Analyze a set of sentences to find the most similar pairs.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Sentence Similarity\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Sentence Transformers\\', \\'api_name\\': \\'sentence-transformers/distiluse-base-multilingual-cased-v1\\', \\'api_call\\': \"SentenceTransformer(\\'sentence-transformers/distiluse-base-multilingual-cased-v1\\')\", \\'api_arguments\\': [\\'sentences\\'], \\'python_environment_requirements\\': \\'pip install -U sentence-transformers\\', \\'example_code\\': \"from sentence_transformers import SentenceTransformer\\\\nsentences = [This is an example sentence, Each sentence is converted]\\\\nmodel = SentenceTransformer(\\'sentence-transformers/distiluse-base-multilingual-cased-v1\\')\\\\nembeddings = model.encode(sentences)\\\\nprint(embeddings)\", \\'performance\\': {\\'dataset\\': \\'https://seb.sbert.net\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'This is a sentence-transformers model: It maps sentences & paragraphs to a 512 dimensional dense vector space and can be used for tasks like clustering or semantic search.\\'}', metadata={})]", "category": "generic"}
{"question_id": 635, "text": " I need a tool to extract code syntax and named entities from a text taken from StackOverflow.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Text2Text Generation\\', \\'api_name\\': \\'castorini/doc2query-t5-base-msmarco\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'castorini/doc2query-t5-base-msmarco\\')\", \\'api_arguments\\': \\'text, max_length\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'MS MARCO\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'A T5 model trained on the MS MARCO dataset for generating queries from documents.\\'}', metadata={})]", "category": "generic"}
{"question_id": 636, "text": " Our video streaming platform wants to categorize thousands of movies into genres. Please help us classify them without any genre labels.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Joblib\\', \\'functionality\\': \\'Multi-class Classification\\', \\'api_name\\': \\'Alexei1/imdb\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'json\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'IMDB\\', \\'accuracy\\': 0.487}, \\'description\\': \\'A tabular classification model trained using AutoTrain for sentiment analysis on the IMDB dataset. The model has a CO2 emission of 0.0186 grams and an accuracy of 0.487.\\'}', metadata={})]", "category": "generic"}
{"question_id": 637, "text": " Can you generate an image of a lighthouse on a foggy island, based on my description?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Image-to-Image\\', \\'api_name\\': \\'lllyasviel/control_v11p_sd15_canny\\', \\'api_call\\': \"ControlNetModel.from_pretrained(\\'lllyasviel/control_v11p_sd15_canny\\')\", \\'api_arguments\\': {\\'text\\': \\'a blue paradise bird in the jungle\\', \\'num_inference_steps\\': 20, \\'generator\\': \\'torch.manual_seed(33)\\', \\'image\\': \\'control_image\\'}, \\'python_environment_requirements\\': [\\'pip install opencv-contrib-python\\', \\'pip install diffusers transformers accelerate\\'], \\'example_code\\': [\\'import torch\\', \\'import os\\', \\'from huggingface_hub import HfApi\\', \\'from pathlib import Path\\', \\'from diffusers.utils import load_image\\', \\'import numpy as np\\', \\'import cv2\\', \\'from PIL import Image\\', \\'from diffusers import (\\', \\' ControlNetModel,\\', \\' StableDiffusionControlNetPipeline,\\', \\' UniPCMultistepScheduler,\\', \\')\\', \\'checkpoint = lllyasviel/control_v11p_sd15_canny\\', \\'image = load_image(\\', \\' https://huggingface.co/lllyasviel/control_v11p_sd15_canny/resolve/main/images/input.png\\', \\')\\', \\'image = np.array(image)\\', \\'low_threshold = 100\\', \\'high_threshold = 200\\', \\'image = cv2.Canny(image, low_threshold, high_threshold)\\', \\'image = image[:, :, None]\\', \\'image = np.concatenate([image, image, image], axis=2)\\', \\'control_image = Image.fromarray(image)\\', \\'control_image.save(./images/control.png)\\', \\'controlnet = ControlNetModel.from_pretrained(checkpoint, torch_dtype=torch.float16)\\', \\'pipe = StableDiffusionControlNetPipeline.from_pretrained(\\', \\' runwayml/stable-diffusion-v1-5, controlnet=controlnet, torch_dtype=torch.float16\\', \\')\\', \\'pipe.scheduler = UniPCMultistepScheduler.from_config(pipe.scheduler.config)\\', \\'pipe.enable_model_cpu_offload()\\', \\'generator = torch.manual_seed(33)\\', \\'image = pipe(a blue paradise bird in the jungle, num_inference_steps=20, generator=generator, image=control_image).images[0]\\', \"image.save(\\'images/image_out.png\\')\"], \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'Controlnet v1.1 is a neural network structure to control diffusion models by adding extra conditions. This checkpoint corresponds to the ControlNet conditioned on Canny edges. It can be used in combination with Stable Diffusion, such as runwayml/stable-diffusion-v1-5.\\'}', metadata={})]", "category": "generic"}
{"question_id": 638, "text": " I am an illustrator, I want to create an appealing image based on a text description for commercial purposes.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'gsdf/Counterfeit-V2.5\\', \\'api_call\\': \"pipeline(\\'text-to-image\\', model=\\'gsdf/Counterfeit-V2.5\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'((masterpiece,best quality)),1girl, solo, animal ears, rabbit, barefoot, knees up, dress, sitting, rabbit ears, short sleeves, looking at viewer, grass, short hair, smile, white hair, puffy sleeves, outdoors, puffy short sleeves, bangs, on ground, full body, animal, white dress, sunlight, brown eyes, dappled sunlight, day, depth of field\\', \\'performance\\': {\\'dataset\\': \\'EasyNegative\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Counterfeit-V2.5 is a text-to-image model that generates anime-style images based on text prompts. It has been updated for ease of use and can be used with negative prompts to create high-quality images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 639, "text": " We want to build a Twitter Bot that creates an image based on users' textual requests. Generate an image with an astronaut playing guitar in space using a model.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Image-to-Text\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Captioning\\', \\'api_name\\': \\'microsoft/git-base\\', \\'api_call\\': \"pipeline(\\'image-to-text\\', model=\\'microsoft/git-base\\')\", \\'api_arguments\\': \\'image\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'git_base(image)\\', \\'performance\\': {\\'dataset\\': [\\'COCO\\', \\'Conceptual Captions (CC3M)\\', \\'SBU\\', \\'Visual Genome (VG)\\', \\'Conceptual Captions (CC12M)\\', \\'ALT200M\\'], \\'accuracy\\': \\'Refer to the paper for evaluation results\\'}, \\'description\\': \"GIT (short for GenerativeImage2Text) model, base-sized version. It was introduced in the paper GIT: A Generative Image-to-text Transformer for Vision and Language by Wang et al. and first released in this repository. The model is trained using \\'teacher forcing\\' on a lot of (image, text) pairs. The goal for the model is simply to predict the next text token, giving the image tokens and previous text tokens. This allows the model to be used for tasks like image and video captioning, visual question answering (VQA) on images and videos, and even image classification (by simply conditioning the model on the image and asking it to generate a class for it in text).\"}', metadata={})]", "category": "generic"}
{"question_id": 640, "text": " Generate an image of a beautiful fantasy landscape based on the description provided: a peaceful scene in a lush green forest with a crystal-clear river flowing through it, under a blue sky with fluffy white clouds.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Image-to-Image\\', \\'api_name\\': \\'GreeneryScenery/SheepsControlV3\\', \\'api_call\\': \"pipeline(\\'image-to-image\\', model=\\'GreeneryScenery/SheepsControlV3\\')\", \\'api_arguments\\': {\\'image\\': \\'Path to image file\\', \\'text_guidance\\': \\'Optional text guidance for the model\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'latest\\', \\'torch\\': \\'latest\\'}, \\'example_code\\': [\\'from transformers import pipeline\\', \"model = pipeline(\\'image-to-image\\', model=\\'GreeneryScenery/SheepsControlV3\\')\", \"result = model({\\'image\\': \\'path/to/image.jpg\\', \\'text_guidance\\': \\'Optional text guidance\\'})\"], \\'performance\\': {\\'dataset\\': \\'GreeneryScenery/SheepsControlV3\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'GreeneryScenery/SheepsControlV3 is a model for image-to-image tasks. It can be used to generate images based on the input image and optional text guidance. The model has some limitations, such as the conditioning image not affecting the output image much. Improvements can be made by training for more epochs, using better prompts, and preprocessing the data.\\'}', metadata={})]", "category": "generic"}
{"question_id": 641, "text": " Tell me a text summary and answer a question from an image.\\n###Input: img_url=\\\"https://example.com/image.jpg\\\", question=\\\"What is the main color of the object?\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Image-to-Text\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'kha-white/manga-ocr-base\\', \\'api_call\\': \"pipeline(\\'ocr\\', model=\\'kha-white/manga-ocr-base\\')\", \\'api_arguments\\': \\'image\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'manga109s\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Optical character recognition for Japanese text, with the main focus being Japanese manga. It uses Vision Encoder Decoder framework. Manga OCR can be used as a general purpose printed Japanese OCR, but its main goal was to provide a high quality text recognition, robust against various scenarios specific to manga: both vertical and horizontal text, text with furigana, text overlaid on images, wide variety of fonts and font styles, and low quality images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 642, "text": " My boss wants me to extract captions from images of people in different settings.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Image-to-Text\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Image Captioning\\', \\'api_name\\': \\'blip-image-captioning-large\\', \\'api_call\\': \\'BlipForConditionalGeneration.from_pretrained(Salesforce/blip-image-captioning-large)\\', \\'api_arguments\\': {\\'raw_image\\': \\'Image\\', \\'text\\': \\'Optional Text\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'BlipProcessor, BlipForConditionalGeneration\\', \\'PIL\\': \\'Image\\', \\'requests\\': \\'requests\\'}, \\'example_code\\': {\\'import_requests\\': \\'import requests\\', \\'import_PIL\\': \\'from PIL import Image\\', \\'import_transformers\\': \\'from transformers import BlipProcessor, BlipForConditionalGeneration\\', \\'load_processor\\': \\'processor = BlipProcessor.from_pretrained(Salesforce/blip-image-captioning-large)\\', \\'load_model\\': \\'model = BlipForConditionalGeneration.from_pretrained(Salesforce/blip-image-captioning-large)\\', \\'load_image\\': \"img_url = \\'https://storage.googleapis.com/sfr-vision-language-research/BLIP/demo.jpg\\'\\\\nraw_image = Image.open(requests.get(img_url, stream=True).raw).convert(\\'RGB\\')\", \\'conditional_captioning\\': \\'text = a photography of\\\\ninputs = processor(raw_image, text, return_tensors=pt)\\\\nout = model.generate(**inputs)\\\\nprint(processor.decode(out[0], skip_special_tokens=True))\\', \\'unconditional_captioning\\': \\'inputs = processor(raw_image, return_tensors=pt)\\\\nout = model.generate(**inputs)\\\\nprint(processor.decode(out[0], skip_special_tokens=True))\\'}, \\'performance\\': {\\'dataset\\': \\'COCO\\', \\'accuracy\\': {\\'image-text retrieval\\': \\'+2.7% recall@1\\', \\'image captioning\\': \\'+2.8% CIDEr\\', \\'VQA\\': \\'+1.6% VQA score\\'}}, \\'description\\': \\'BLIP is a Vision-Language Pre-training (VLP) framework that achieves state-of-the-art results on a wide range of vision-language tasks, such as image-text retrieval, image captioning, and VQA. It effectively utilizes noisy web data by bootstrapping the captions, where a captioner generates synthetic captions and a filter removes the noisy ones.\\'}', metadata={})]", "category": "generic"}
{"question_id": 643, "text": " Create an AI system capable of producing short videos based on text input in Persian and English.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'ImRma/Brucelee\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'ImRma/Brucelee\\')\", \\'api_arguments\\': [\\'your_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Hugging Face model for converting Persian and English text into video.\\'}', metadata={})]", "category": "generic"}
{"question_id": 644, "text": " I am organizing a virtual party and want to create a short, autogenerated video based on a text description (e.g., \\\"cats playing with laser pointer\\\"). Can you accomplish this with a text-to-video generation API?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'camenduru/text2-video-zero\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'camenduru/text2-video-zero\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'This model is used for generating videos from text inputs. It is based on the Hugging Face framework and can be used with the transformers library. The model is trained on a variety of text and video datasets, and can be used for tasks such as video summarization, video generation from text prompts, and more.\\'}', metadata={})]", "category": "generic"}
{"question_id": 645, "text": " I need to create a system that can answer questions related to a document provided. The system should use a pre-trained model.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Classification\\', \\'api_name\\': \\'shahrukhx01/question-vs-statement-classifier\\', \\'api_call\\': \"AutoModelForSequenceClassification.from_pretrained(\\'shahrukhx01/question-vs-statement-classifier\\')\", \\'api_arguments\\': {\\'tokenizer\\': \\'AutoTokenizer.from_pretrained(shahrukhx01/question-vs-statement-classifier)\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'from transformers import AutoTokenizer, AutoModelForSequenceClassification\\'}, \\'example_code\\': \\'tokenizer = AutoTokenizer.from_pretrained(shahrukhx01/question-vs-statement-classifier)\\\\nmodel = AutoModelForSequenceClassification.from_pretrained(shahrukhx01/question-vs-statement-classifier)\\', \\'performance\\': {\\'dataset\\': \\'Haystack\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Trained to add the feature for classifying queries between Question Query vs Statement Query using classification in Haystack\\'}', metadata={})]", "category": "generic"}
{"question_id": 646, "text": " We are developing a document management system. We want to extract the relevant information from them through OCR text scanning and answering questions.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 647, "text": " The team is creating a home security system. We are currently looking at understanding the depth of objects in the video stream.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Object Detection\\', \\'api_name\\': \\'keremberke/yolov8s-csgo-player-detection\\', \\'api_call\\': \"YOLO(\\'keremberke/yolov8s-csgo-player-detection\\')\", \\'api_arguments\\': {\\'image\\': \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'}, \\'python_environment_requirements\\': [\\'ultralyticsplus==0.0.23\\', \\'ultralytics==8.0.21\\'], \\'example_code\\': \"from ultralyticsplus import YOLO, render_result\\\\nmodel = YOLO(\\'keremberke/yolov8s-csgo-player-detection\\')\\\\nmodel.overrides[\\'conf\\'] = 0.25\\\\nmodel.overrides[\\'iou\\'] = 0.45\\\\nmodel.overrides[\\'agnostic_nms\\'] = False\\\\nmodel.overrides[\\'max_det\\'] = 1000\\\\nimage = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\\\\nresults = model.predict(image)\\\\nprint(results[0].boxes)\\\\nrender = render_result(model=model, image=image, result=results[0])\\\\nrender.show()\", \\'performance\\': {\\'dataset\\': \\'csgo-object-detection\\', \\'accuracy\\': 0.886}, \\'description\\': \"A YOLOv8 model for detecting Counter-Strike: Global Offensive (CS:GO) players. Supports the labels [\\'ct\\', \\'cthead\\', \\'t\\', \\'thead\\'].\"}', metadata={})]", "category": "generic"}
{"question_id": 648, "text": " I want to know the depth information of an image for a robot navigation project.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Depth Estimation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Depth Estimation\\', \\'api_name\\': \\'glpn-nyu-finetuned-diode\\', \\'api_call\\': \"pipeline(\\'depth-estimation\\', model=\\'sayakpaul/glpn-nyu-finetuned-diode\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'diode-subset\\', \\'accuracy\\': {\\'Loss\\': 0.4359, \\'Rmse\\': 0.4276}}, \\'description\\': \\'This model is a fine-tuned version of vinvino02/glpn-nyu on the diode-subset dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 649, "text": " A start-up is looking to develop a robot for navigation; they require a solution to estimate depth from a single image.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Depth Estimation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Depth Estimation\\', \\'api_name\\': \\'glpn-kitti\\', \\'api_call\\': \"GLPNForDepthEstimation.from_pretrained(\\'vinvino02/glpn-kitti\\')\", \\'api_arguments\\': \\'images, return_tensors\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'from transformers import GLPNFeatureExtractor, GLPNForDepthEstimation\\\\nimport torch\\\\nimport numpy as np\\\\nfrom PIL import Image\\\\nimport requests\\\\nurl = http://images.cocodataset.org/val2017/000000039769.jpg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nfeature_extractor = GLPNFeatureExtractor.from_pretrained(vinvino02/glpn-kitti)\\\\nmodel = GLPNForDepthEstimation.from_pretrained(vinvino02/glpn-kitti)\\\\ninputs = feature_extractor(images=image, return_tensors=pt)\\\\nwith torch.no_grad():\\\\n outputs = model(**inputs)\\\\n predicted_depth = outputs.predicted_depth\\\\nprediction = torch.nn.functional.interpolate(\\\\n predicted_depth.unsqueeze(1),\\\\n size=image.size[::-1],\\\\n mode=bicubic,\\\\n align_corners=False,\\\\n)\\\\noutput = prediction.squeeze().cpu().numpy()\\\\nformatted = (output * 255 / np.max(output)).astype(uint8)\\\\ndepth = Image.fromarray(formatted)\\', \\'performance\\': {\\'dataset\\': \\'KITTI\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Global-Local Path Networks (GLPN) model trained on KITTI for monocular depth estimation. It was introduced in the paper Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth by Kim et al. and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 650, "text": " We are building a smart security system to determine if a person is an adult\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'michellejieli/NSFW_text_classifier\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'michellejieli/NSFW_text_classification\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'classifier(I see you\u2019ve set aside this special time to humiliate yourself in public.)\\', \\'performance\\': {\\'dataset\\': \\'Reddit posts\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'DistilBERT is a transformer model that performs sentiment analysis. I fine-tuned the model on Reddit posts with the purpose of classifying not safe for work (NSFW) content, specifically text that is considered inappropriate and unprofessional. The model predicts 2 classes, which are NSFW or safe for work (SFW). The model is a fine-tuned version of DistilBERT. It was fine-tuned on 14317 Reddit posts pulled from the Reddit API.\\'}', metadata={})]", "category": "generic"}
{"question_id": 651, "text": " Design a program that can help farmers detect diseases in bean crops by analyzing images of the crop leaves.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Image-to-Image\\', \\'api_name\\': \\'GreeneryScenery/SheepsControlV5\\', \\'api_call\\': \"pipeline(\\'image-to-image\\', model=\\'GreeneryScenery/SheepsControlV5\\')\", \\'api_arguments\\': {\\'input_image\\': \\'path/to/image/file\\'}, \\'python_environment_requirements\\': {\\'huggingface_hub\\': \\'>=0.0.17\\', \\'transformers\\': \\'>=4.13.0\\', \\'torch\\': \\'>=1.10.0\\'}, \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'poloclub/diffusiondb\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'SheepsControlV5 is an image-to-image model trained on the poloclub/diffusiondb dataset. It is designed for transforming input images into a different style or representation.\\'}', metadata={})]", "category": "generic"}
{"question_id": 652, "text": " To build an intelligent system to recognize and identify objects in submitted pictures within the household, which includes furniture, electronics and ornaments.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'microsoft/table-transformer-structure-recognition\\', \\'api_call\\': \"pipeline(\\'object-detection\\', model=\\'microsoft/table-transformer-structure-recognition\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'PubTables1M\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Table Transformer (DETR) model trained on PubTables1M for detecting the structure (like rows, columns) in tables.\\'}', metadata={})]", "category": "generic"}
{"question_id": 653, "text": " Give me a solution to find out the license plate in the given car images.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'License Plate Detection\\', \\'api_name\\': \\'keremberke/yolov5s-license-plate\\', \\'api_call\\': \"yolov5.load(\\'keremberke/yolov5s-license-plate\\')\", \\'api_arguments\\': {\\'img\\': \\'image url or path\\', \\'size\\': \\'image resize dimensions\\', \\'augment\\': \\'optional, test time augmentation\\'}, \\'python_environment_requirements\\': \\'pip install -U yolov5\\', \\'example_code\\': [\\'import yolov5\\', \"model = yolov5.load(\\'keremberke/yolov5s-license-plate\\')\", \\'model.conf = 0.25\\', \\'model.iou = 0.45\\', \\'model.agnostic = False\\', \\'model.multi_label = False\\', \\'model.max_det = 1000\\', \"img = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\", \\'results = model(img, size=640)\\', \\'results = model(img, augment=True)\\', \\'predictions = results.pred[0]\\', \\'boxes = predictions[:, :4]\\', \\'scores = predictions[:, 4]\\', \\'categories = predictions[:, 5]\\', \\'results.show()\\', \"results.save(save_dir=\\'results/\\')\"], \\'performance\\': {\\'dataset\\': \\'keremberke/license-plate-object-detection\\', \\'accuracy\\': 0.985}, \\'description\\': \\'A YOLOv5 based license plate detection model trained on a custom dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 654, "text": " I want to build a social media app for outdoor enthusiasts, and I need to identify objects related to outdoor activities in images. Come up with a proper API call to identify these objects.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Object Detection\\', \\'api_name\\': \\'keremberke/yolov8n-csgo-player-detection\\', \\'api_call\\': \"YOLO(\\'keremberke/yolov8n-csgo-player-detection\\')\", \\'api_arguments\\': {\\'image\\': \\'URL or local path to image\\'}, \\'python_environment_requirements\\': \\'pip install ultralyticsplus==0.0.23 ultralytics==8.0.21\\', \\'example_code\\': \"from ultralyticsplus import YOLO, render_result\\\\n\\\\nmodel = YOLO(\\'keremberke/yolov8n-csgo-player-detection\\')\\\\n\\\\nmodel.overrides[\\'conf\\'] = 0.25\\\\nmodel.overrides[\\'iou\\'] = 0.45\\\\nmodel.overrides[\\'agnostic_nms\\'] = False\\\\nmodel.overrides[\\'max_det\\'] = 1000\\\\n\\\\nimage = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\\\\n\\\\nresults = model.predict(image)\\\\n\\\\nprint(results[0].boxes)\\\\nrender = render_result(model=model, image=image, result=results[0])\\\\nrender.show()\", \\'performance\\': {\\'dataset\\': \\'csgo-object-detection\\', \\'accuracy\\': 0.844}, \\'description\\': \"A YOLOv8 model for detecting Counter-Strike: Global Offensive (CS:GO) players with supported labels: [\\'ct\\', \\'cthead\\', \\'t\\', \\'thead\\'].\"}', metadata={})]", "category": "generic"}
{"question_id": 655, "text": " I want an AI that can separate objects in an image based on semantic segmentation.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Segmentation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Semantic Segmentation\\', \\'api_name\\': \\'nvidia/segformer-b5-finetuned-cityscapes-1024-1024\\', \\'api_call\\': \"SegformerForSemanticSegmentation.from_pretrained(\\'nvidia/segformer-b5-finetuned-cityscapes-1024-1024\\')\", \\'api_arguments\\': {\\'images\\': \\'image\\', \\'return_tensors\\': \\'pt\\'}, \\'python_environment_requirements\\': {\\'packages\\': [\\'transformers\\', \\'PIL\\', \\'requests\\']}, \\'example_code\\': \\'from transformers import SegformerFeatureExtractor, SegformerForSemanticSegmentation\\\\nfrom PIL import Image\\\\nimport requests\\\\nfeature_extractor = SegformerFeatureExtractor.from_pretrained(nvidia/segformer-b5-finetuned-cityscapes-1024-1024)\\\\nmodel = SegformerForSemanticSegmentation.from_pretrained(nvidia/segformer-b5-finetuned-cityscapes-1024-1024)\\\\nurl = http://images.cocodataset.org/val2017/000000039769.jpg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\ninputs = feature_extractor(images=image, return_tensors=pt)\\\\noutputs = model(**inputs)\\\\nlogits = outputs.logits\\', \\'performance\\': {\\'dataset\\': \\'CityScapes\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'SegFormer model fine-tuned on CityScapes at resolution 1024x1024. It was introduced in the paper SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers by Xie et al. and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 656, "text": " Our image-processing service has to extract contextual information from images.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'Linaqruf/anything-v3.0\\', \\'api_call\\': \"Text2ImagePipeline(model=\\'Linaqruf/anything-v3.0\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A text-to-image model that generates images from text descriptions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 657, "text": " Our company develops a city planning application. We need to segment streets, buildings, and trees in aerial photographs.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Zero-Shot Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Geolocalization\\', \\'api_name\\': \\'geolocal/StreetCLIP\\', \\'api_call\\': \"CLIPModel.from_pretrained(\\'geolocal/StreetCLIP\\')\", \\'api_arguments\\': {\\'pretrained_model_name_or_path\\': \\'geolocal/StreetCLIP\\'}, \\'python_environment_requirements\\': [\\'transformers\\', \\'PIL\\', \\'requests\\'], \\'example_code\\': \\'from PIL import Image\\\\nimport requests\\\\nfrom transformers import CLIPProcessor, CLIPModel\\\\nmodel = CLIPModel.from_pretrained(geolocal/StreetCLIP)\\\\nprocessor = CLIPProcessor.from_pretrained(geolocal/StreetCLIP)\\\\nurl = https://huggingface.co/geolocal/StreetCLIP/resolve/main/sanfrancisco.jpeg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nchoices = [San Jose, San Diego, Los Angeles, Las Vegas, San Francisco]\\\\ninputs = processor(text=choices, images=image, return_tensors=pt, padding=True)\\\\noutputs = model(**inputs)\\\\nlogits_per_image = outputs.logits_per_image\\\\nprobs = logits_per_image.softmax(dim=1)\\', \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'IM2GPS\\', \\'accuracy\\': {\\'25km\\': 28.3, \\'200km\\': 45.1, \\'750km\\': 74.7, \\'2500km\\': 88.2}}, {\\'name\\': \\'IM2GPS3K\\', \\'accuracy\\': {\\'25km\\': 22.4, \\'200km\\': 37.4, \\'750km\\': 61.3, \\'2500km\\': 80.4}}]}, \\'description\\': \\'StreetCLIP is a robust foundation model for open-domain image geolocalization and other geographic and climate-related tasks. Trained on an original dataset of 1.1 million street-level urban and rural geo-tagged images, it achieves state-of-the-art performance on multiple open-domain image geolocalization benchmarks in zero-shot, outperforming supervised models trained on millions of images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 658, "text": " We are building a product which can identify birds in the images. Design the model which can help us segment the birds in an image.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Paraphrase-based utterance augmentation\\', \\'api_name\\': \\'prithivida/parrot_fluency_model\\', \\'api_call\\': \"pipeline(\\'text-classification\\', model=\\'prithivida/parrot_fluency_model\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"parrot(\\'your input text\\')\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'Parrot is a paraphrase-based utterance augmentation framework purpose-built to accelerate training NLU models. A paraphrase framework is more than just a paraphrasing model.\\'}', metadata={})]", "category": "generic"}
{"question_id": 659, "text": " In this Star Wars movie scene, I want to create a depth estimation for the stormtroopers.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Depth Estimation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Depth Estimation\\', \\'api_name\\': \\'glpn-kitti\\', \\'api_call\\': \"GLPNForDepthEstimation.from_pretrained(\\'vinvino02/glpn-kitti\\')\", \\'api_arguments\\': \\'images, return_tensors\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'from transformers import GLPNFeatureExtractor, GLPNForDepthEstimation\\\\nimport torch\\\\nimport numpy as np\\\\nfrom PIL import Image\\\\nimport requests\\\\nurl = http://images.cocodataset.org/val2017/000000039769.jpg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nfeature_extractor = GLPNFeatureExtractor.from_pretrained(vinvino02/glpn-kitti)\\\\nmodel = GLPNForDepthEstimation.from_pretrained(vinvino02/glpn-kitti)\\\\ninputs = feature_extractor(images=image, return_tensors=pt)\\\\nwith torch.no_grad():\\\\n outputs = model(**inputs)\\\\n predicted_depth = outputs.predicted_depth\\\\nprediction = torch.nn.functional.interpolate(\\\\n predicted_depth.unsqueeze(1),\\\\n size=image.size[::-1],\\\\n mode=bicubic,\\\\n align_corners=False,\\\\n)\\\\noutput = prediction.squeeze().cpu().numpy()\\\\nformatted = (output * 255 / np.max(output)).astype(uint8)\\\\ndepth = Image.fromarray(formatted)\\', \\'performance\\': {\\'dataset\\': \\'KITTI\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Global-Local Path Networks (GLPN) model trained on KITTI for monocular depth estimation. It was introduced in the paper Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth by Kim et al. and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 660, "text": " I am a drone maker that is building a navigation module for my drones. I need to sharpen the images captured from the drone in real-time.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image-to-Image\\', \\'framework\\': \\'Keras\\', \\'functionality\\': \\'Image Deblurring\\', \\'api_name\\': \\'google/maxim-s3-deblurring-gopro\\', \\'api_call\\': \"from_pretrained_keras(\\'google/maxim-s3-deblurring-gopro\\')\", \\'api_arguments\\': [\\'image\\'], \\'python_environment_requirements\\': [\\'huggingface_hub\\', \\'PIL\\', \\'tensorflow\\', \\'numpy\\', \\'requests\\'], \\'example_code\\': \\'from huggingface_hub import from_pretrained_keras\\\\nfrom PIL import Image\\\\nimport tensorflow as tf\\\\nimport numpy as np\\\\nimport requests\\\\nurl = https://github.com/sayakpaul/maxim-tf/raw/main/images/Deblurring/input/1fromGOPR0950.png\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nimage = np.array(image)\\\\nimage = tf.convert_to_tensor(image)\\\\nimage = tf.image.resize(image, (256, 256))\\\\nmodel = from_pretrained_keras(google/maxim-s3-deblurring-gopro)\\\\npredictions = model.predict(tf.expand_dims(image, 0))\\', \\'performance\\': {\\'dataset\\': \\'GoPro\\', \\'accuracy\\': {\\'PSNR\\': 32.86, \\'SSIM\\': 0.961}}, \\'description\\': \\'MAXIM model pre-trained for image deblurring. It was introduced in the paper MAXIM: Multi-Axis MLP for Image Processing by Zhengzhong Tu, Hossein Talebi, Han Zhang, Feng Yang, Peyman Milanfar, Alan Bovik, Yinxiao Li and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 661, "text": " Develop a method to generate images of realistic-looking churches.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Unconditional Image Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Unconditional Image Generation\\', \\'api_name\\': \\'google/ncsnpp-church-256\\', \\'api_call\\': \"DiffusionPipeline.from_pretrained(\\'google/ncsnpp-church-256\\')\", \\'api_arguments\\': \\'model_id\\', \\'python_environment_requirements\\': \\'diffusers\\', \\'example_code\\': \\'!pip install diffusers\\\\nfrom diffusers import DiffusionPipeline\\\\nmodel_id = google/ncsnpp-church-256\\\\nsde_ve = DiffusionPipeline.from_pretrained(model_id)\\\\nimage = sde_ve()[sample]\\\\nimage[0].save(sde_ve_generated_image.png)\\', \\'performance\\': {\\'dataset\\': \\'CIFAR-10\\', \\'accuracy\\': {\\'Inception_score\\': 9.89, \\'FID\\': 2.2, \\'likelihood\\': 2.99}}, \\'description\\': \\'Score-Based Generative Modeling through Stochastic Differential Equations (SDE) for unconditional image generation. This model achieves record-breaking performance on CIFAR-10 and can generate high fidelity images of size 1024 x 1024.\\'}', metadata={})]", "category": "generic"}
{"question_id": 662, "text": " Generate a classical image by using Diffusion Model\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Unconditional Image Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Denoising Diffusion Probabilistic Models (DDPM)\\', \\'api_name\\': \\'google/ddpm-bedroom-256\\', \\'api_call\\': \"DDPMPipeline.from_pretrained(\\'google/ddpm-bedroom-256\\')\", \\'api_arguments\\': \\'None\\', \\'python_environment_requirements\\': \\'diffusers\\', \\'example_code\\': \\'!pip install diffusers\\\\nfrom diffusers import DDPMPipeline, DDIMPipeline, PNDMPipeline\\\\nmodel_id = google/ddpm-bedroom-256\\\\nddpm = DDPMPipeline.from_pretrained(model_id)\\\\nimage = ddpm().images[0]\\\\nimage.save(ddpm_generated_image.png)\\', \\'performance\\': {\\'dataset\\': \\'CIFAR10\\', \\'accuracy\\': {\\'Inception score\\': 9.46, \\'FID score\\': 3.17}}, \\'description\\': \\'We present high quality image synthesis results using diffusion probabilistic models, a class of latent variable models inspired by considerations from nonequilibrium thermodynamics. Our best results are obtained by training on a weighted variational bound designed according to a novel connection between diffusion probabilistic models and denoising score matching with Langevin dynamics, and our models naturally admit a progressive lossy decompression scheme that can be interpreted as a generalization of autoregressive decoding. On the unconditional CIFAR10 dataset, we obtain an Inception score of 9.46 and a state-of-the-art FID score of 3.17. On 256x256 LSUN, we obtain sample quality similar to ProgressiveGAN.\\'}', metadata={})]", "category": "generic"}
{"question_id": 663, "text": " We need a picture that has a nostalgic look in high quality for the cover of our upcoming magazine.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'prompthero/openjourney-v4\\', \\'api_call\\': \"pipeline(\\'text-to-image\\', model=\\'prompthero/openjourney-v4\\')\", \\'api_arguments\\': {\\'text\\': \\'string\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"generate_image(\\'your text here\\')\", \\'performance\\': {\\'dataset\\': \\'Midjourney v4 images\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Openjourney v4 is trained on +124k Midjourney v4 images by PromptHero. It is used for generating images based on text inputs.\\'}', metadata={})]", "category": "generic"}
{"question_id": 664, "text": " We are an e-learning provider who wants to classify the content of a video lecture automatically.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Video Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Video Classification\\', \\'api_name\\': \\'MCG-NJU/videomae-base-short-finetuned-kinetics\\', \\'api_call\\': \"VideoMAEForVideoClassification.from_pretrained(\\'MCG-NJU/videomae-base-short-finetuned-kinetics\\')\", \\'api_arguments\\': [\\'video\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import VideoMAEImageProcessor, VideoMAEForVideoClassification\\\\nimport numpy as np\\\\nimport torch\\\\nvideo = list(np.random.randn(16, 3, 224, 224))\\\\nprocessor = VideoMAEImageProcessor.from_pretrained(\\'MCG-NJU/videomae-base-short-finetuned-kinetics\\')\\\\nmodel = VideoMAEForVideoClassification.from_pretrained(\\'MCG-NJU/videomae-base-short-finetuned-kinetics\\')\\\\ninputs = processor(video, return_tensors=\\'pt\\')\\\\nwith torch.no_grad():\\\\n outputs = model(**inputs)\\\\n logits = outputs.logits\\\\npredicted_class_idx = logits.argmax(-1).item()\\\\nprint(\\'Predicted class:\\', model.config.id2label[predicted_class_idx])\", \\'performance\\': {\\'dataset\\': \\'Kinetics-400\\', \\'accuracy\\': {\\'top-1\\': 79.4, \\'top-5\\': 94.1}}, \\'description\\': \\'VideoMAE model pre-trained for 800 epochs in a self-supervised way and fine-tuned in a supervised way on Kinetics-400. It was introduced in the paper VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training by Tong et al. and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 665, "text": " We need to classify actions of athletes in sports videos. Can you help us to analyze and classify these videos?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Video Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Video Classification\\', \\'api_name\\': \\'MCG-NJU/videomae-base-short-finetuned-kinetics\\', \\'api_call\\': \"VideoMAEForVideoClassification.from_pretrained(\\'MCG-NJU/videomae-base-short-finetuned-kinetics\\')\", \\'api_arguments\\': [\\'video\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import VideoMAEImageProcessor, VideoMAEForVideoClassification\\\\nimport numpy as np\\\\nimport torch\\\\nvideo = list(np.random.randn(16, 3, 224, 224))\\\\nprocessor = VideoMAEImageProcessor.from_pretrained(\\'MCG-NJU/videomae-base-short-finetuned-kinetics\\')\\\\nmodel = VideoMAEForVideoClassification.from_pretrained(\\'MCG-NJU/videomae-base-short-finetuned-kinetics\\')\\\\ninputs = processor(video, return_tensors=\\'pt\\')\\\\nwith torch.no_grad():\\\\n outputs = model(**inputs)\\\\n logits = outputs.logits\\\\npredicted_class_idx = logits.argmax(-1).item()\\\\nprint(\\'Predicted class:\\', model.config.id2label[predicted_class_idx])\", \\'performance\\': {\\'dataset\\': \\'Kinetics-400\\', \\'accuracy\\': {\\'top-1\\': 79.4, \\'top-5\\': 94.1}}, \\'description\\': \\'VideoMAE model pre-trained for 800 epochs in a self-supervised way and fine-tuned in a supervised way on Kinetics-400. It was introduced in the paper VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training by Tong et al. and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 666, "text": " We want to classify images supplied by our users into categories such as cats, dogs, birds, and more.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'abhishek/autotrain-dog-vs-food\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'abhishek/autotrain-dog-vs-food\\')\", \\'api_arguments\\': \\'image_path\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'sasha/dog-food\\', \\'accuracy\\': 0.998}, \\'description\\': \"A pre-trained model for classifying images as either dog or food using Hugging Face\\'s AutoTrain framework.\"}', metadata={})]", "category": "generic"}
{"question_id": 667, "text": " Identify which smartphone brand is featured in an image and predict the intensity of luxury level.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image-to-Image\\', \\'framework\\': \\'Keras\\', \\'functionality\\': \\'Image Deblurring\\', \\'api_name\\': \\'google/maxim-s3-deblurring-gopro\\', \\'api_call\\': \"from_pretrained_keras(\\'google/maxim-s3-deblurring-gopro\\')\", \\'api_arguments\\': [\\'image\\'], \\'python_environment_requirements\\': [\\'huggingface_hub\\', \\'PIL\\', \\'tensorflow\\', \\'numpy\\', \\'requests\\'], \\'example_code\\': \\'from huggingface_hub import from_pretrained_keras\\\\nfrom PIL import Image\\\\nimport tensorflow as tf\\\\nimport numpy as np\\\\nimport requests\\\\nurl = https://github.com/sayakpaul/maxim-tf/raw/main/images/Deblurring/input/1fromGOPR0950.png\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nimage = np.array(image)\\\\nimage = tf.convert_to_tensor(image)\\\\nimage = tf.image.resize(image, (256, 256))\\\\nmodel = from_pretrained_keras(google/maxim-s3-deblurring-gopro)\\\\npredictions = model.predict(tf.expand_dims(image, 0))\\', \\'performance\\': {\\'dataset\\': \\'GoPro\\', \\'accuracy\\': {\\'PSNR\\': 32.86, \\'SSIM\\': 0.961}}, \\'description\\': \\'MAXIM model pre-trained for image deblurring. It was introduced in the paper MAXIM: Multi-Axis MLP for Image Processing by Zhengzhong Tu, Hossein Talebi, Han Zhang, Feng Yang, Peyman Milanfar, Alan Bovik, Yinxiao Li and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 668, "text": " A Chinese social media company has approached us for help in content moderation. We recorded an API to classify images which don't belong to suitable content.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'michellejieli/NSFW_text_classifier\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'michellejieli/NSFW_text_classification\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'classifier(I see you\u2019ve set aside this special time to humiliate yourself in public.)\\', \\'performance\\': {\\'dataset\\': \\'Reddit posts\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'DistilBERT is a transformer model that performs sentiment analysis. I fine-tuned the model on Reddit posts with the purpose of classifying not safe for work (NSFW) content, specifically text that is considered inappropriate and unprofessional. The model predicts 2 classes, which are NSFW or safe for work (SFW). The model is a fine-tuned version of DistilBERT. It was fine-tuned on 14317 Reddit posts pulled from the Reddit API.\\'}', metadata={})]", "category": "generic"}
{"question_id": 669, "text": " We are working on environmental studies and there is a specific image we need to geolocalize. Let's get the probabilities of different cities for the given image.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Zero-Shot Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Geolocalization\\', \\'api_name\\': \\'geolocal/StreetCLIP\\', \\'api_call\\': \"CLIPModel.from_pretrained(\\'geolocal/StreetCLIP\\')\", \\'api_arguments\\': {\\'pretrained_model_name_or_path\\': \\'geolocal/StreetCLIP\\'}, \\'python_environment_requirements\\': [\\'transformers\\', \\'PIL\\', \\'requests\\'], \\'example_code\\': \\'from PIL import Image\\\\nimport requests\\\\nfrom transformers import CLIPProcessor, CLIPModel\\\\nmodel = CLIPModel.from_pretrained(geolocal/StreetCLIP)\\\\nprocessor = CLIPProcessor.from_pretrained(geolocal/StreetCLIP)\\\\nurl = https://huggingface.co/geolocal/StreetCLIP/resolve/main/sanfrancisco.jpeg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nchoices = [San Jose, San Diego, Los Angeles, Las Vegas, San Francisco]\\\\ninputs = processor(text=choices, images=image, return_tensors=pt, padding=True)\\\\noutputs = model(**inputs)\\\\nlogits_per_image = outputs.logits_per_image\\\\nprobs = logits_per_image.softmax(dim=1)\\', \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'IM2GPS\\', \\'accuracy\\': {\\'25km\\': 28.3, \\'200km\\': 45.1, \\'750km\\': 74.7, \\'2500km\\': 88.2}}, {\\'name\\': \\'IM2GPS3K\\', \\'accuracy\\': {\\'25km\\': 22.4, \\'200km\\': 37.4, \\'750km\\': 61.3, \\'2500km\\': 80.4}}]}, \\'description\\': \\'StreetCLIP is a robust foundation model for open-domain image geolocalization and other geographic and climate-related tasks. Trained on an original dataset of 1.1 million street-level urban and rural geo-tagged images, it achieves state-of-the-art performance on multiple open-domain image geolocalization benchmarks in zero-shot, outperforming supervised models trained on millions of images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 670, "text": " Our e-commerce platform needs automatic classification of product images without retraining. Develop a classifier for the e-commerce platform.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'abhishek/autotrain-dog-vs-food\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'abhishek/autotrain-dog-vs-food\\')\", \\'api_arguments\\': \\'image_path\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'sasha/dog-food\\', \\'accuracy\\': 0.998}, \\'description\\': \"A pre-trained model for classifying images as either dog or food using Hugging Face\\'s AutoTrain framework.\"}', metadata={})]", "category": "generic"}
{"question_id": 671, "text": " We are developing a customer support platform for our telecommunication company in Spain. We want to know if they are happy or unhappy with our services, based on the content of their message.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'finiteautomata/beto-sentiment-analysis\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'finiteautomata/beto-sentiment-analysis\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'Hugging Face Transformers library\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'TASS 2020 corpus\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Model trained with TASS 2020 corpus (around ~5k tweets) of several dialects of Spanish. Base model is BETO, a BERT model trained in Spanish. Uses POS, NEG, NEU labels.\\'}', metadata={})]", "category": "generic"}
{"question_id": 672, "text": " I need a function to find the most relevent passage given a question and several candidate passages.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'distilbert-base-uncased-distilled-squad\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=\\'distilbert-base-uncased-distilled-squad\\')\", \\'api_arguments\\': [\\'question\\', \\'context\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nquestion_answerer = pipeline(question-answering, model=\\'distilbert-base-uncased-distilled-squad\\')\\\\ncontext = r\\\\n... Extractive Question Answering is the task of extracting an answer from a text given a question. An example of a\\\\n... question answering dataset is the SQuAD dataset, which is entirely based on that task. If you would like to fine-tune\\\\n... a model on a SQuAD task, you may leverage the examples/pytorch/question-answering/run_squad.py script.\\\\n... \\\\nresult = question_answerer(question=What is a good example of a question answering dataset?, context=context)\\\\nprint(\\\\n... fAnswer: \\'{result[\\'answer\\']}\\', score: {round(result[\\'score\\'], 4)}, start: {result[\\'start\\']}, end: {result[\\'end\\']}\\\\n...)\", \\'performance\\': {\\'dataset\\': \\'SQuAD v1.1\\', \\'accuracy\\': \\'86.9 F1 score\\'}, \\'description\\': \"DistilBERT base uncased distilled SQuAD is a fine-tuned version of DistilBERT-base-uncased for the task of question answering. It has 40% less parameters than bert-base-uncased, runs 60% faster while preserving over 95% of BERT\\'s performances as measured on the GLUE language understanding benchmark.\"}', metadata={})]", "category": "generic"}
{"question_id": 673, "text": " As a community manager, I would like to monitor my forum's comment section for toxic or harmful content. I want to find a solution that can flag these types of comments automatically, so I can address them promptly.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'martin-ha/toxic-comment-model\\', \\'api_call\\': \"pipeline(model=\\'martin-ha/toxic-comment-model\\')\", \\'api_arguments\\': {\\'model_path\\': \\'martin-ha/toxic-comment-model\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import AutoModelForSequenceClassification, AutoTokenizer, TextClassificationPipeline\\\\nmodel_path = martin-ha/toxic-comment-model\\\\ntokenizer = AutoTokenizer.from_pretrained(model_path)\\\\nmodel = AutoModelForSequenceClassification.from_pretrained(model_path)\\\\npipeline = TextClassificationPipeline(model=model, tokenizer=tokenizer)\\\\nprint(pipeline(\\'This is a test text.\\'))\", \\'performance\\': {\\'dataset\\': \\'held-out test set\\', \\'accuracy\\': 0.94, \\'f1-score\\': 0.59}, \\'description\\': \\'This model is a fine-tuned version of the DistilBERT model to classify toxic comments.\\'}', metadata={})]", "category": "generic"}
{"question_id": 674, "text": " I am making a keyword search engine that ranks text passages based on their importance regarding a given keyword.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Text2Text Generation\\', \\'api_name\\': \\'castorini/doc2query-t5-base-msmarco\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'castorini/doc2query-t5-base-msmarco\\')\", \\'api_arguments\\': \\'text, max_length\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'MS MARCO\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'A T5 model trained on the MS MARCO dataset for generating queries from documents.\\'}', metadata={})]", "category": "generic"}
{"question_id": 675, "text": " Identify the type of emotion in a movie review.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Emotion Classification\\', \\'api_name\\': \\'j-hartmann/emotion-english-distilroberta-base\\', \\'api_call\\': \"pipeline(\\'text-classification\\', model=\\'j-hartmann/emotion-english-distilroberta-base\\', return_all_scores=True)\", \\'api_arguments\\': {\\'text\\': \\'string\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'latest\\'}, \\'example_code\\': \\'from transformers import pipeline\\\\nclassifier = pipeline(text-classification, model=j-hartmann/emotion-english-distilroberta-base, return_all_scores=True)\\\\nclassifier(I love this!)\\', \\'performance\\': {\\'dataset\\': \\'Balanced subset from 6 diverse datasets\\', \\'accuracy\\': \\'66%\\'}, \\'description\\': \"This model classifies emotions in English text data. It predicts Ekman\\'s 6 basic emotions, plus a neutral class: anger, disgust, fear, joy, neutral, sadness, and surprise. The model is a fine-tuned checkpoint of DistilRoBERTa-base.\"}', metadata={})]", "category": "generic"}
{"question_id": 676, "text": " We are developing a medical records analysis software that automatically recognizes biomedical entities from physicians' case reports.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Token Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Named Entity Recognition\\', \\'api_name\\': \\'d4data/biomedical-ner-all\\', \\'api_call\\': \"AutoModelForTokenClassification.from_pretrained(\\'d4data/biomedical-ner-all\\')\", \\'api_arguments\\': {\\'model\\': \\'AutoModelForTokenClassification.from_pretrained(d4data/biomedical-ner-all)\\', \\'tokenizer\\': \\'AutoTokenizer.from_pretrained(d4data/biomedical-ner-all)\\', \\'aggregation_strategy\\': \\'simple\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'from transformers import pipeline, AutoTokenizer, AutoModelForTokenClassification\\'}, \\'example_code\\': \\'pipe(The patient reported no recurrence of palpitations at follow-up 6 months after the ablation.)\\', \\'performance\\': {\\'dataset\\': \\'Maccrobat\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'An English Named Entity Recognition model, trained on Maccrobat to recognize the bio-medical entities (107 entities) from a given text corpus (case reports etc.). This model was built on top of distilbert-base-uncased.\\'}', metadata={})]", "category": "generic"}
{"question_id": 677, "text": " We have customer reviews of various software products. We want to extract company names in those reviews.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Classification\\', \\'api_name\\': \\'lvwerra/distilbert-imdb\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'lvwerra/distilbert-imdb\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\', \\'pytorch\\'], \\'example_code\\': \"classifier(\\'I love this movie!\\')\", \\'performance\\': {\\'dataset\\': \\'imdb\\', \\'accuracy\\': 0.928}, \\'description\\': \\'This model is a fine-tuned version of distilbert-base-uncased on the imdb dataset. It is used for sentiment analysis on movie reviews and achieves an accuracy of 0.928 on the evaluation set.\\'}', metadata={})]", "category": "generic"}
{"question_id": 678, "text": " I am working on a Chinese language project and I need to tokenize the sentences for better processing.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Sentence Similarity\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'shibing624/text2vec-base-chinese\\', \\'api_call\\': \"SentenceModel(\\'shibing624/text2vec-base-chinese\\')\", \\'api_arguments\\': [\\'sentences\\'], \\'python_environment_requirements\\': [\\'text2vec\\', \\'transformers\\'], \\'example_code\\': \"from text2vec import SentenceModel\\\\nsentences = [\\'\\', \\'\\']\\\\nmodel = SentenceModel(\\'shibing624/text2vec-base-chinese\\')\\\\nembeddings = model.encode(sentences)\\\\nprint(embeddings)\", \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'ATEC\\', \\'accuracy\\': \\'31.93\\'}, {\\'name\\': \\'BQ\\', \\'accuracy\\': \\'42.67\\'}, {\\'name\\': \\'LCQMC\\', \\'accuracy\\': \\'70.16\\'}, {\\'name\\': \\'PAWSX\\', \\'accuracy\\': \\'17.21\\'}, {\\'name\\': \\'STS-B\\', \\'accuracy\\': \\'79.30\\'}]}, \\'description\\': \\'This is a CoSENT(Cosine Sentence) model: shibing624/text2vec-base-chinese. It maps sentences to a 768 dimensional dense vector space and can be used for tasks like sentence embeddings, text matching or semantic search.\\'}', metadata={})]", "category": "generic"}
{"question_id": 679, "text": " I need a system that extracts all the well-known named entities such as person names, locations, and organizations from news articles.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Token Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Named Entity Recognition\\', \\'api_name\\': \\'dslim/bert-base-NER-uncased\\', \\'api_call\\': \"pipeline(\\'ner\\', model=\\'dslim/bert-base-NER-uncased\\')\", \\'api_arguments\\': {}, \\'python_environment_requirements\\': {\\'transformers\\': \\'>=4.0.0\\'}, \\'example_code\\': \"nlp(\\'My name is John and I live in New York.\\')\", \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A pretrained BERT model for Named Entity Recognition (NER) on uncased text. It can be used to extract entities such as person names, locations, and organizations from text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 680, "text": " I have a list of Olympic Game host cities and their corresponding years. I want to know which year the games were held in Beijing.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Visual Question Answering\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Visual Question Answering\\', \\'api_name\\': \\'sheldonxxxx/OFA_model_weights\\', \\'api_call\\': \"AutoModel.from_pretrained(\\'sheldonxxxx/OFA_model_weights\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'This is an unoffical mirror of the model weights for use with https://github.com/OFA-Sys/OFA. The original link is too slow when downloading from outside of China.\\'}', metadata={})]", "category": "generic"}
{"question_id": 681, "text": " I have a table with data about different types of bards and their magical abilities. Let me see if I can find the best bard given the data in the table.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Table Question Answering\\', \\'framework\\': \\'PyTorch Transformers\\', \\'functionality\\': \\'Table Question Answering\\', \\'api_name\\': \\'table-question-answering-tapas\\', \\'api_call\\': \"pipeline(\\'table-question-answering\\', model=\\'Meena/table-question-answering-tapas\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'This model can be loaded on the Inference API on-demand.\\', \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'SQA (Sequential Question Answering by Microsoft)\\', \\'accuracy\\': None}, {\\'name\\': \\'WTQ (Wiki Table Questions by Stanford University)\\', \\'accuracy\\': None}, {\\'name\\': \\'WikiSQL (by Salesforce)\\', \\'accuracy\\': None}]}, \\'description\\': \\'TAPAS, the model learns an inner representation of the English language used in tables and associated texts, which can then be used to extract features useful for downstream tasks such as answering questions about a table, or determining whether a sentence is entailed or refuted by the contents of a table. It is a BERT-based model specifically designed (and pre-trained) for answering questions about tabular data. TAPAS uses relative position embeddings and has 7 token types that encode tabular structure. It is pre-trained on the masked language modeling (MLM) objective on a large dataset comprising millions of tables from English Wikipedia and corresponding texts.\\'}', metadata={})]", "category": "generic"}
{"question_id": 682, "text": " I want to automate the process of answering questions about historical facts. When given a question and a surrounding context, it should provide an accurate response.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'distilbert-base-uncased-distilled-squad\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=\\'distilbert-base-uncased-distilled-squad\\')\", \\'api_arguments\\': [\\'question\\', \\'context\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nquestion_answerer = pipeline(question-answering, model=\\'distilbert-base-uncased-distilled-squad\\')\\\\ncontext = r\\\\n... Extractive Question Answering is the task of extracting an answer from a text given a question. An example of a\\\\n... question answering dataset is the SQuAD dataset, which is entirely based on that task. If you would like to fine-tune\\\\n... a model on a SQuAD task, you may leverage the examples/pytorch/question-answering/run_squad.py script.\\\\n... \\\\nresult = question_answerer(question=What is a good example of a question answering dataset?, context=context)\\\\nprint(\\\\n... fAnswer: \\'{result[\\'answer\\']}\\', score: {round(result[\\'score\\'], 4)}, start: {result[\\'start\\']}, end: {result[\\'end\\']}\\\\n...)\", \\'performance\\': {\\'dataset\\': \\'SQuAD v1.1\\', \\'accuracy\\': \\'86.9 F1 score\\'}, \\'description\\': \"DistilBERT base uncased distilled SQuAD is a fine-tuned version of DistilBERT-base-uncased for the task of question answering. It has 40% less parameters than bert-base-uncased, runs 60% faster while preserving over 95% of BERT\\'s performances as measured on the GLUE language understanding benchmark.\"}', metadata={})]", "category": "generic"}
{"question_id": 683, "text": " I am a lawyer, I have a lot of text. I need a system which reads context and answers the questions based on the context.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Text2Text Generation\\', \\'api_name\\': \\'castorini/doc2query-t5-base-msmarco\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'castorini/doc2query-t5-base-msmarco\\')\", \\'api_arguments\\': \\'text, max_length\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'MS MARCO\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'A T5 model trained on the MS MARCO dataset for generating queries from documents.\\'}', metadata={})]", "category": "generic"}
{"question_id": 684, "text": " I want to create a question answering script that can help me answer questions about a given passage of text.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'distilbert-base-uncased-distilled-squad\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=\\'distilbert-base-uncased-distilled-squad\\')\", \\'api_arguments\\': [\\'question\\', \\'context\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nquestion_answerer = pipeline(question-answering, model=\\'distilbert-base-uncased-distilled-squad\\')\\\\ncontext = r\\\\n... Extractive Question Answering is the task of extracting an answer from a text given a question. An example of a\\\\n... question answering dataset is the SQuAD dataset, which is entirely based on that task. If you would like to fine-tune\\\\n... a model on a SQuAD task, you may leverage the examples/pytorch/question-answering/run_squad.py script.\\\\n... \\\\nresult = question_answerer(question=What is a good example of a question answering dataset?, context=context)\\\\nprint(\\\\n... fAnswer: \\'{result[\\'answer\\']}\\', score: {round(result[\\'score\\'], 4)}, start: {result[\\'start\\']}, end: {result[\\'end\\']}\\\\n...)\", \\'performance\\': {\\'dataset\\': \\'SQuAD v1.1\\', \\'accuracy\\': \\'86.9 F1 score\\'}, \\'description\\': \"DistilBERT base uncased distilled SQuAD is a fine-tuned version of DistilBERT-base-uncased for the task of question answering. It has 40% less parameters than bert-base-uncased, runs 60% faster while preserving over 95% of BERT\\'s performances as measured on the GLUE language understanding benchmark.\"}', metadata={})]", "category": "generic"}
{"question_id": 685, "text": " To enhance our FAQ bot, we need to extract answers from a given knowledge base text.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'distilbert-base-uncased-distilled-squad\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=\\'distilbert-base-uncased-distilled-squad\\')\", \\'api_arguments\\': [\\'question\\', \\'context\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nquestion_answerer = pipeline(question-answering, model=\\'distilbert-base-uncased-distilled-squad\\')\\\\ncontext = r\\\\n... Extractive Question Answering is the task of extracting an answer from a text given a question. An example of a\\\\n... question answering dataset is the SQuAD dataset, which is entirely based on that task. If you would like to fine-tune\\\\n... a model on a SQuAD task, you may leverage the examples/pytorch/question-answering/run_squad.py script.\\\\n... \\\\nresult = question_answerer(question=What is a good example of a question answering dataset?, context=context)\\\\nprint(\\\\n... fAnswer: \\'{result[\\'answer\\']}\\', score: {round(result[\\'score\\'], 4)}, start: {result[\\'start\\']}, end: {result[\\'end\\']}\\\\n...)\", \\'performance\\': {\\'dataset\\': \\'SQuAD v1.1\\', \\'accuracy\\': \\'86.9 F1 score\\'}, \\'description\\': \"DistilBERT base uncased distilled SQuAD is a fine-tuned version of DistilBERT-base-uncased for the task of question answering. It has 40% less parameters than bert-base-uncased, runs 60% faster while preserving over 95% of BERT\\'s performances as measured on the GLUE language understanding benchmark.\"}', metadata={})]", "category": "generic"}
{"question_id": 686, "text": " Our customer is a Spanish travel agency. They need to classify customer reviews into categories such as 'travel', 'cooking', and 'dancing'.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'finiteautomata/beto-sentiment-analysis\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'finiteautomata/beto-sentiment-analysis\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'Hugging Face Transformers library\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'TASS 2020 corpus\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Model trained with TASS 2020 corpus (around ~5k tweets) of several dialects of Spanish. Base model is BETO, a BERT model trained in Spanish. Uses POS, NEG, NEU labels.\\'}', metadata={})]", "category": "generic"}
{"question_id": 687, "text": " Create a method to determine, as an entertainment recommendation system, which category a text message about a daily activity belongs to.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'michellejieli/NSFW_text_classifier\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'michellejieli/NSFW_text_classification\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'classifier(I see you\u2019ve set aside this special time to humiliate yourself in public.)\\', \\'performance\\': {\\'dataset\\': \\'Reddit posts\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'DistilBERT is a transformer model that performs sentiment analysis. I fine-tuned the model on Reddit posts with the purpose of classifying not safe for work (NSFW) content, specifically text that is considered inappropriate and unprofessional. The model predicts 2 classes, which are NSFW or safe for work (SFW). The model is a fine-tuned version of DistilBERT. It was fine-tuned on 14317 Reddit posts pulled from the Reddit API.\\'}', metadata={})]", "category": "generic"}
{"question_id": 688, "text": " We have written a summary of a new book's plot. Now, we want to ensure if the summary contains conflicting information.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'pszemraj/long-t5-tglobal-base-16384-book-summary\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'pszemraj/long-t5-tglobal-base-16384-book-summary\\')\", \\'api_arguments\\': [\\'long_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nimport torch\\\\nsummarizer = pipeline(\\\\n summarization,\\\\n pszemraj/long-t5-tglobal-base-16384-book-summary,\\\\n device=0 if torch.cuda.is_available() else -1,\\\\n)\\\\nlong_text = Here is a lot of text I don\\'t want to read. Replace me\\\\nresult = summarizer(long_text)\\\\nprint(result[0][summary_text])\", \\'performance\\': {\\'dataset\\': \\'kmfoda/booksum\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.408, \\'ROUGE-2\\': 6.065, \\'ROUGE-L\\': 16.721, \\'ROUGE-LSUM\\': 33.34}}, \\'description\\': \\'A fine-tuned version of google/long-t5-tglobal-base on the kmfoda/booksum dataset, which can be used to summarize long text and generate SparkNotes-esque summaries of arbitrary topics. The model generalizes reasonably well to academic and narrative text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 689, "text": " We have a user comment about a product, and we want to infer if the sentiment of the comment is positive or negative.\\n###Input: \\\"I recently purchased this product and it completely exceeded my expectations! The build quality is top-notch, and I've already recommended it to several friends.\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'bert-base-multilingual-uncased-sentiment\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'nlptown/bert-base-multilingual-uncased-sentiment\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"result = sentiment_pipeline(\\'I love this product!\\')\", \\'performance\\': {\\'dataset\\': [{\\'language\\': \\'English\\', \\'accuracy\\': {\\'exact\\': \\'67%\\', \\'off-by-1\\': \\'95%\\'}}, {\\'language\\': \\'Dutch\\', \\'accuracy\\': {\\'exact\\': \\'57%\\', \\'off-by-1\\': \\'93%\\'}}, {\\'language\\': \\'German\\', \\'accuracy\\': {\\'exact\\': \\'61%\\', \\'off-by-1\\': \\'94%\\'}}, {\\'language\\': \\'French\\', \\'accuracy\\': {\\'exact\\': \\'59%\\', \\'off-by-1\\': \\'94%\\'}}, {\\'language\\': \\'Italian\\', \\'accuracy\\': {\\'exact\\': \\'59%\\', \\'off-by-1\\': \\'95%\\'}}, {\\'language\\': \\'Spanish\\', \\'accuracy\\': {\\'exact\\': \\'58%\\', \\'off-by-1\\': \\'95%\\'}}]}, \\'description\\': \\'This a bert-base-multilingual-uncased model finetuned for sentiment analysis on product reviews in six languages: English, Dutch, German, French, Spanish and Italian. It predicts the sentiment of the review as a number of stars (between 1 and 5).\\'}', metadata={})]", "category": "generic"}
{"question_id": 690, "text": " Extract a conclusion from the following text: \\\"Studies have been shown that owning a dog is good for you. Having a dog can help decrease stress levels, improve your mood, and increase physical activity.\\\"\\n###Input: Studies have been shown that owning a dog is good for you. Having a dog can help decrease stress levels, improve your mood, and increase physical activity.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'michellejieli/emotion_text_classifier\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'michellejieli/emotion_text_classifier\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'classifier(I love this!)\\', \\'performance\\': {\\'dataset\\': [\\'Crowdflower (2016)\\', \\'Emotion Dataset, Elvis et al. (2018)\\', \\'GoEmotions, Demszky et al. (2020)\\', \\'ISEAR, Vikash (2018)\\', \\'MELD, Poria et al. (2019)\\', \\'SemEval-2018, EI-reg, Mohammad et al. (2018)\\', \\'Emotion Lines (Friends)\\'], \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'DistilRoBERTa-base is a transformer model that performs sentiment analysis. I fine-tuned the model on transcripts from the Friends show with the goal of classifying emotions from text data, specifically dialogue from Netflix shows or movies. The model predicts 6 Ekman emotions and a neutral class. These emotions include anger, disgust, fear, joy, neutrality, sadness, and surprise.\\'}', metadata={})]", "category": "generic"}
{"question_id": 691, "text": " Our company needs a versatile NLP model to build a social media manager to generate summaries of lengthy articles for sharing on social media.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Summarization\\', \\'api_name\\': \\'facebook/bart-large-cnn\\', \\'api_call\\': \"pipeline(\\'summarization\\', model=\\'facebook/bart-large-cnn\\')\", \\'api_arguments\\': [\\'ARTICLE\\', \\'max_length\\', \\'min_length\\', \\'do_sample\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'from transformers import pipeline\\\\nsummarizer = pipeline(summarization, model=facebook/bart-large-cnn)\\\\nARTICLE = ...\\\\nprint(summarizer(ARTICLE, max_length=130, min_length=30, do_sample=False))\\', \\'performance\\': {\\'dataset\\': \\'cnn_dailymail\\', \\'accuracy\\': {\\'ROUGE-1\\': 42.949, \\'ROUGE-2\\': 20.815, \\'ROUGE-L\\': 30.619, \\'ROUGE-LSUM\\': 40.038}}, \\'description\\': \\'BART (large-sized model), fine-tuned on CNN Daily Mail. BART is a transformer encoder-encoder (seq2seq) model with a bidirectional (BERT-like) encoder and an autoregressive (GPT-like) decoder. BART is pre-trained by (1) corrupting text with an arbitrary noising function, and (2) learning a model to reconstruct the original text. BART is particularly effective when fine-tuned for text generation (e.g. summarization, translation) but also works well for comprehension tasks (e.g. text classification, question answering). This particular checkpoint has been fine-tuned on CNN Daily Mail, a large collection of text-summary pairs.\\'}', metadata={})]", "category": "generic"}
{"question_id": 692, "text": " We have a real estate website and would like to translate property descriptions in English to French.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Translation\\', \\'api_name\\': \\'opus-mt-fr-en\\', \\'api_call\\': \"pipeline(\\'translation_fr_to_en\\', model=\\'Helsinki-NLP/opus-mt-fr-en\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\', \\'torch\\'], \\'example_code\\': \"translation_pipeline(\\'Bonjour, comment \u00e7a va?\\')\", \\'performance\\': {\\'dataset\\': \\'opus\\', \\'accuracy\\': {\\'BLEU\\': {\\'newsdiscussdev2015-enfr.fr.en\\': 33.1, \\'newsdiscusstest2015-enfr.fr.en\\': 38.7, \\'newssyscomb2009.fr.en\\': 30.3, \\'news-test2008.fr.en\\': 26.2, \\'newstest2009.fr.en\\': 30.2, \\'newstest2010.fr.en\\': 32.2, \\'newstest2011.fr.en\\': 33.0, \\'newstest2012.fr.en\\': 32.8, \\'newstest2013.fr.en\\': 33.9, \\'newstest2014-fren.fr.en\\': 37.8, \\'Tatoeba.fr.en\\': 57.5}}}, \\'description\\': \\'Helsinki-NLP/opus-mt-fr-en is a machine translation model trained to translate from French to English. It is based on the Marian NMT framework and trained on the OPUS dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 693, "text": " My friend wrote me an English text about the upcoming holiday. I need to translate it to Arabic so I can share it with my family.\\n###Input: \\\"My friend is planning a holiday trip for our families. He found a beautiful place with a beach, swimming pool, and a wide range of outdoor activities for kids. There's also a famous seafood restaurant nearby! I think our families will have a great time together.\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Translation\\', \\'api_name\\': \\'Helsinki-NLP/opus-mt-en-ar\\', \\'api_call\\': \"pipeline(\\'translation_en_to_ar\\', model=\\'Helsinki-NLP/opus-mt-en-ar\\')\", \\'api_arguments\\': {\\'text\\': \\'string\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\ntranslation = pipeline(\\'translation_en_to_ar\\', model=\\'Helsinki-NLP/opus-mt-en-ar\\')\\\\ntranslated_text = translation(\\'Hello World\\')\", \\'performance\\': {\\'dataset\\': \\'Tatoeba-test.eng.ara\\', \\'accuracy\\': {\\'BLEU\\': 14.0, \\'chr-F\\': 0.437}}, \\'description\\': \"A Hugging Face Transformers model for English to Arabic translation, trained on the Tatoeba dataset. It uses a transformer architecture and requires a sentence initial language token in the form of \\'>>id<<\\' (id = valid target language ID).\"}', metadata={})]", "category": "generic"}
{"question_id": 694, "text": " We received a customer feedback document which is very lengthy. We need a summarization of it.\\n###Input: The customer support service was excellent. All our concerns were attended to promptly by the friendly and knowledgeable staff. The user interface, however, could use some improvement. Navigating through the platform can be challenging, and it took us quite some time to find the relevant information we needed.\\nAdditionally, some of our team members faced technical issues while using the platform, particularly with the integration of third-party tools. We had to reach out to the support team multiple times to resolve these issues. Overall, while we genuinely appreciate your team's assistance, we expect better performance from the platform itself.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'impira/layoutlm-invoices\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=\\'impira/layoutlm-invoices\\')\", \\'api_arguments\\': \\'question, context\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"qa_pipeline(question=\\'your question\\', context=\\'your document context\\')\", \\'performance\\': {\\'dataset\\': \\'proprietary dataset of invoices, SQuAD2.0, and DocVQA\\', \\'accuracy\\': \\'not provided\\'}, \\'description\\': \\'This is a fine-tuned version of the multi-modal LayoutLM model for the task of question answering on invoices and other documents. It has been fine-tuned on a proprietary dataset of invoices as well as both SQuAD2.0 and DocVQA for general comprehension. Unlike other QA models, which can only extract consecutive tokens (because they predict the start and end of a sequence), this model can predict longer-range, non-consecutive sequences with an additional classifier head.\\'}', metadata={})]", "category": "generic"}
{"question_id": 695, "text": " I would like to give a brief overview of our team meeting to my supervisor, so I need a summary of the conversation.\\n###Input: \\\"Anna: In today's meeting, we discussed increasing marketing budget. Tom: I suggested allocating more funds to social media campaigns. Sarah: I proposed focusing on improving SEO. Anna: We agreed on investing in content creation, too. Tom: The team will revise the strategy and present it next week. Sarah: Let's determine new KPIs for evaluating our progress.\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text2Text Generation\\', \\'api_name\\': \\'t5-efficient-large-nl36_fine_tune_sum_V2\\', \\'api_call\\': \"pipeline(\\'summarization\\', model=\\'Samuel-Fipps/t5-efficient-large-nl36_fine_tune_sum_V2\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'samsum\\', \\'accuracy\\': {\\'ROUGE-1\\': 54.933, \\'ROUGE-2\\': 31.797, \\'ROUGE-L\\': 47.006, \\'ROUGE-LSUM\\': 51.203, \\'loss\\': 1.131, \\'gen_len\\': 23.799}}, {\\'name\\': \\'cnn_dailymail\\', \\'accuracy\\': {\\'ROUGE-1\\': 34.406, \\'ROUGE-2\\': 14.127, \\'ROUGE-L\\': 24.335, \\'ROUGE-LSUM\\': 31.658, \\'loss\\': 2.446, \\'gen_len\\': 45.928}}]}, \\'description\\': \"A T5-based summarization model trained on the Samsum dataset. This model can be used for text-to-text generation tasks such as summarization without adding \\'summarize\\' to the start of the input string. It has been fine-tuned for 10K steps with a batch size of 10.\"}', metadata={})]", "category": "generic"}
{"question_id": 696, "text": " We need to develop a conversational chatbot that can answer users' queries, ask questions, and have a friendly conversation.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'hyunwoongko/blenderbot-9B\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'hyunwoongko/blenderbot-9B\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'Input a message to start chatting with hyunwoongko/blenderbot-9B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high-performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi-turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 697, "text": " My daughter left her laptop logged in, and I found her English essay unfinished. There's a sentence that goes like \\\"In the story, the antagonist represents the <mask> nature of humanity.\\\" Can you help me complete her sentence with an appropriate word?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Fill-Mask\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Fill-Mask\\', \\'api_name\\': \\'cl-tohoku/bert-base-japanese-whole-word-masking\\', \\'api_call\\': \"AutoModelForMaskedLM.from_pretrained(\\'cl-tohoku/bert-base-japanese-whole-word-masking\\')\", \\'api_arguments\\': {\\'model\\': \\'cl-tohoku/bert-base-japanese-whole-word-masking\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'latest\\'}, \\'example_code\\': \"fill_mask(\\'[MASK]\\')\", \\'performance\\': {\\'dataset\\': \\'Japanese Wikipedia\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This is a BERT model pretrained on texts in the Japanese language. This version of the model processes input texts with word-level tokenization based on the IPA dictionary, followed by the WordPiece subword tokenization. Additionally, the model is trained with the whole word masking enabled for the masked language modeling (MLM) objective.\\'}', metadata={})]", "category": "generic"}
{"question_id": 698, "text": " We are developing a program to teach French to English speakers. The program should complete a sentence with a missing word in French.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Audio-to-Audio\\', \\'framework\\': \\'Fairseq\\', \\'functionality\\': \\'speech-to-speech-translation\\', \\'api_name\\': \\'facebook/textless_sm_en_fr\\', \\'api_call\\': \"load_model_ensemble_and_task_from_hf_hub(\\'facebook/textless_sm_en_fr\\')\", \\'api_arguments\\': [\\'input_file\\'], \\'python_environment_requirements\\': [\\'huggingface_hub\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'This model is a speech-to-speech translation model trained by Facebook. It is designed for translating English speech to French speech.\\'}', metadata={})]", "category": "generic"}
{"question_id": 699, "text": " I have a multilingual document, and there is a missing word in the document. Please help me find out what this missing word might be.\\n \n Use this API documentation for reference: [Document(page_content=\"{'domain': 'Natural Language Processing Token Classification', 'framework': 'Transformers', 'functionality': 'punctuation prediction', 'api_name': 'oliverguhr/fullstop-punctuation-multilang-large', 'api_call': 'PunctuationModel()', 'api_arguments': ['text'], 'python_environment_requirements': ['pip install deepmultilingualpunctuation'], 'example_code': 'from deepmultilingualpunctuation import PunctuationModel\\\\nmodel = PunctuationModel()\\\\ntext = My name is Clara and I live in Berkeley California Ist das eine Frage Frau M\u00fcller\\\\nresult = model.restore_punctuation(text)\\\\nprint(result)', 'performance': {'dataset': 'wmt/europarl', 'EN_accuracy': 0.775, 'DE_accuracy': 0.814, 'FR_accuracy': 0.782, 'IT_accuracy': 0.762}, 'description': 'This model predicts the punctuation of English, Italian, French and German texts. It was developed to restore the punctuation of transcribed spoken language and trained on the Europarl Dataset provided by the SEPP-NLG Shared Task. The model restores the following punctuation markers: ., ,, ?, -, :.'}\", metadata={})]", "category": "generic"}
{"question_id": 700, "text": " As a translation company, we are translating messages between co-workers in a multinational company. Translate the message from Hindi to French.\\n###Input: \\\"\\u0906\\u092a\\u0915\\u0940 \\u092a\\u094d\\u0930\\u0947\\u091c\\u093c\\u091f\\u0947\\u0936\\u0928 \\u0915\\u093e \\u0906\\u0927\\u093e\\u0930 \\u0905\\u091a\\u094d\\u091b\\u093e \\u0925\\u093e, \\u0932\\u0947\\u0915\\u093f\\u0928 \\u0921\\u0947\\u091f\\u093e \\u0935\\u093f\\u0936\\u094d\\u0932\\u0947\\u0937\\u0923 \\u092a\\u0930 \\u0927\\u094d\\u092f\\u093e\\u0928 \\u0926\\u0947\\u0928\\u093e \\u091a\\u093e\\u0939\\u093f\\u090f\\u0964\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Translation\\', \\'api_name\\': \\'Helsinki-NLP/opus-mt-en-fr\\', \\'api_call\\': \"translate(\\'input_text\\', model=\\'Helsinki-NLP/opus-mt-en-fr\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'opus\\', \\'accuracy\\': {\\'BLEU\\': {\\'newsdiscussdev2015-enfr.en.fr\\': 33.8, \\'newsdiscusstest2015-enfr.en.fr\\': 40.0, \\'newssyscomb2009.en.fr\\': 29.8, \\'news-test2008.en.fr\\': 27.5, \\'newstest2009.en.fr\\': 29.4, \\'newstest2010.en.fr\\': 32.7, \\'newstest2011.en.fr\\': 34.3, \\'newstest2012.en.fr\\': 31.8, \\'newstest2013.en.fr\\': 33.2, \\'Tatoeba.en.fr\\': 50.5}}}, \\'description\\': \\'Helsinki-NLP/opus-mt-en-fr is a translation model that translates English text to French using the Hugging Face Transformers library. It is based on the OPUS dataset and uses a transformer-align architecture with normalization and SentencePiece pre-processing.\\'}', metadata={})]", "category": "generic"}
{"question_id": 701, "text": " Our task is to complete a given sentence with a missing word. The sentence is from an electronic health record.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentence Correction\\', \\'api_name\\': \\'flexudy/t5-base-multi-sentence-doctor\\', \\'api_call\\': \"AutoModelWithLMHead.from_pretrained(\\'flexudy/t5-base-multi-sentence-doctor\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'from transformers import AutoTokenizer, AutoModelWithLMHead\\\\ntokenizer = AutoTokenizer.from_pretrained(flexudy/t5-base-multi-sentence-doctor)\\\\nmodel = AutoModelWithLMHead.from_pretrained(flexudy/t5-base-multi-sentence-doctor)\\\\ninput_text = repair_sentence: m a medical doct context: {That is my job I a}{or I save lives} </s>\\\\ninput_ids = tokenizer.encode(input_text, return_tensors=pt)\\\\noutputs = model.generate(input_ids, max_length=32, num_beams=1)\\\\nsentence = tokenizer.decode(outputs[0], skip_special_tokens=True, clean_up_tokenization_spaces=True)\\\\nassert sentence == I am a medical doctor.\\', \\'performance\\': {\\'dataset\\': \\'tatoeba\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'Sentence doctor is a T5 model that attempts to correct the errors or mistakes found in sentences. Model works on English, German and French text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 702, "text": " A Dutch friend asked for help in completing a sentence with a missing word. Can you fill in the blank?\\n###Input: \\\"Het is vandaag erg koud, dus vergeet niet je ___ mee te nemen.\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Fill-Mask\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Fill-Mask\\', \\'api_name\\': \\'GroNLP/bert-base-dutch-cased\\', \\'api_call\\': \"AutoModel.from_pretrained(\\'GroNLP/bert-base-dutch-cased\\')\", \\'api_arguments\\': [\\'pretrained_model_name_or_path\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'from transformers import AutoTokenizer, AutoModel, TFAutoModel\\\\ntokenizer = AutoTokenizer.from_pretrained(GroNLP/bert-base-dutch-cased)\\\\nmodel = AutoModel.from_pretrained(GroNLP/bert-base-dutch-cased)\\', \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'CoNLL-2002\\', \\'accuracy\\': \\'90.24\\'}, {\\'name\\': \\'SoNaR-1\\', \\'accuracy\\': \\'84.93\\'}, {\\'name\\': \\'spaCy UD LassySmall\\', \\'accuracy\\': \\'86.10\\'}]}, \\'description\\': \\'BERTje is a Dutch pre-trained BERT model developed at the University of Groningen.\\'}', metadata={})]", "category": "generic"}
{"question_id": 703, "text": " Our school needs to create fill-in-the-blank quizzes for students. Is it possible to generate a fill-in-the-blank question from the following sentence: \\\"The cat chased the mouse and then climbed the tree.\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 704, "text": " We are building a customer support chatbot that needs to find the most related FAQ for a given customer query.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'Zixtrauce/BaekBot\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'Zixtrauce/BaekBot\\')\", \\'api_arguments\\': \\'input_message\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"conversational_pipeline(\\'input_message\\')\", \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'BaekBot is a conversational model based on the GPT-2 architecture for text generation. It can be used for generating human-like responses in a chat-like environment.\\'}', metadata={})]", "category": "generic"}
{"question_id": 705, "text": " We want to make a system that reads our email messages out loud in different voices.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'SYSPIN/Marathi_Male_TTS\\', \\'api_call\\': \"api.load(\\'ESPnet/espnet_model_zoo:SYSPIN/Marathi_Male_TTS\\').\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'huggingface_hub\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Marathi Male Text-to-Speech model using ESPnet framework.\\'}', metadata={})]", "category": "generic"}
{"question_id": 706, "text": " We are working on a language learning app for Chinese. We need to give audio examples for each lesson.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'CTranslate2\\', \\'functionality\\': \\'Automatic Speech Recognition\\', \\'api_name\\': \\'guillaumekln/faster-whisper-large-v2\\', \\'api_call\\': \"WhisperModel(\\'large-v2\\')\", \\'api_arguments\\': [\\'audio.mp3\\'], \\'python_environment_requirements\\': [\\'faster_whisper\\'], \\'example_code\\': \\'from faster_whisper import WhisperModel\\\\nmodel = WhisperModel(large-v2)\\\\nsegments, info = model.transcribe(audio.mp3)\\\\nfor segment in segments:\\\\n print([%.2fs -&gt; %.2fs] %s % (segment.start, segment.end, segment.text))\\', \\'performance\\': {\\'dataset\\': \\'99 languages\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Whisper large-v2 model for CTranslate2. This model can be used in CTranslate2 or projets based on CTranslate2 such as faster-whisper.\\'}', metadata={})]", "category": "generic"}
{"question_id": 707, "text": " We are building an AI chatbot that reads out user messages using synthesized human-like speech. For this purpose, we need to convert text messages to audio. \\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'hyunwoongko/blenderbot-9B\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'hyunwoongko/blenderbot-9B\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'Input a message to start chatting with hyunwoongko/blenderbot-9B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high-performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi-turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 708, "text": " I want to identify when people are speaking in an audio file.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'pyannote.audio\\', \\'functionality\\': \\'Speaker diarization\\', \\'api_name\\': \\'johnislarry/cloned-pyannote-speaker-diarization-endpoint\\', \\'api_call\\': \"Pipeline.from_pretrained(\\'pyannote/speaker-diarization@2.1\\',use_auth_token=\\'ACCESS_TOKEN_GOES_HERE\\')\", \\'api_arguments\\': [\\'num_speakers\\', \\'min_speakers\\', \\'max_speakers\\', \\'segmentation_onset\\'], \\'python_environment_requirements\\': \\'pyannote.audio 2.0\\', \\'example_code\\': {\\'load_pipeline\\': \\'from pyannote.audio import Pipeline\\\\npipeline = Pipeline.from_pretrained(pyannote/speaker-diarization@2022.07)\\', \\'apply_pipeline\\': \\'diarization = pipeline(audio.wav)\\', \\'save_output\\': \\'with open(audio.rttm, w) as rttm:\\\\n  diarization.write_rttm(rttm)\\'}, \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'AISHELL-4\\', \\'accuracy\\': {\\'DER%\\': 14.61, \\'FA%\\': 3.31, \\'Miss%\\': 4.35, \\'Conf%\\': 6.95}}, {\\'name\\': \\'AMI Mix-Headset only_words\\', \\'accuracy\\': {\\'DER%\\': 18.21, \\'FA%\\': 3.28, \\'Miss%\\': 11.07, \\'Conf%\\': 3.87}}, {\\'name\\': \\'AMI Array1-01 only_words\\', \\'accuracy\\': {\\'DER%\\': 29.0, \\'FA%\\': 2.71, \\'Miss%\\': 21.61, \\'Conf%\\': 4.68}}, {\\'name\\': \\'CALLHOME Part2\\', \\'accuracy\\': {\\'DER%\\': 30.24, \\'FA%\\': 3.71, \\'Miss%\\': 16.86, \\'Conf%\\': 9.66}}, {\\'name\\': \\'DIHARD 3 Full\\', \\'accuracy\\': {\\'DER%\\': 20.99, \\'FA%\\': 4.25, \\'Miss%\\': 10.74, \\'Conf%\\': 6.0}}, {\\'name\\': \\'REPERE Phase 2\\', \\'accuracy\\': {\\'DER%\\': 12.62, \\'FA%\\': 1.55, \\'Miss%\\': 3.3, \\'Conf%\\': 7.76}}, {\\'name\\': \\'VoxConverse v0.0.2\\', \\'accuracy\\': {\\'DER%\\': 12.76, \\'FA%\\': 3.45, \\'Miss%\\': 3.85, \\'Conf%\\': 5.46}}]}, \\'description\\': \\'This API provides speaker diarization functionality using the pyannote.audio framework. It is capable of processing audio files and outputting speaker diarization results in RTTM format. The API supports providing the number of speakers, minimum and maximum number of speakers, and adjusting the segmentation onset threshold.\\'}', metadata={})]", "category": "generic"}
{"question_id": 709, "text": " I have a collection of audio recordings from different sources, and I want to convert them into text to make an archive.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'mio/Artoria\\', \\'api_call\\': \"pipeline(\\'text-to-speech\\', model=\\'mio/Artoria\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline; tts = pipeline(\\'text-to-speech\\', model=\\'mio/Artoria\\'); tts(\\'s\\')\", \\'performance\\': {\\'dataset\\': \\'fate\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This model was trained by mio using fate recipe in espnet. It is a text-to-speech model that can convert text input into speech output.\\'}', metadata={})]", "category": "generic"}
{"question_id": 710, "text": " A company developing an application for transcribing customer service calls requires a model that can understand spoken language.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'CTranslate2\\', \\'functionality\\': \\'Automatic Speech Recognition\\', \\'api_name\\': \\'guillaumekln/faster-whisper-large-v2\\', \\'api_call\\': \"WhisperModel(\\'large-v2\\')\", \\'api_arguments\\': [\\'audio.mp3\\'], \\'python_environment_requirements\\': [\\'faster_whisper\\'], \\'example_code\\': \\'from faster_whisper import WhisperModel\\\\nmodel = WhisperModel(large-v2)\\\\nsegments, info = model.transcribe(audio.mp3)\\\\nfor segment in segments:\\\\n print([%.2fs -&gt; %.2fs] %s % (segment.start, segment.end, segment.text))\\', \\'performance\\': {\\'dataset\\': \\'99 languages\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Whisper large-v2 model for CTranslate2. This model can be used in CTranslate2 or projets based on CTranslate2 such as faster-whisper.\\'}', metadata={})]", "category": "generic"}
{"question_id": 711, "text": " We are building an audio assistant. Apply noise suppression to our new voice commands.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Voice Activity Detection, Speech-to-Noise Ratio, and C50 Room Acoustics Estimation\\', \\'api_name\\': \\'pyannote/brouhaha\\', \\'api_call\\': \"Model.from_pretrained(\\'pyannote/brouhaha\\', use_auth_token=\\'ACCESS_TOKEN_GOES_HERE\\')\", \\'api_arguments\\': [\\'audio.wav\\'], \\'python_environment_requirements\\': [\\'pyannote-audio\\', \\'brouhaha-vad\\'], \\'example_code\\': [\\'from pyannote.audio import Model\\', \\'model = Model.from_pretrained(pyannote/brouhaha, use_auth_token=ACCESS_TOKEN_GOES_HERE)\\', \\'from pyannote.audio import Inference\\', \\'inference = Inference(model)\\', \\'output = inference(audio.wav)\\', \\'for frame, (vad, snr, c50) in output:\\', \\'  t = frame.middle\\', \\'  print(f{t:8.3f} vad={100*vad:.0f}% snr={snr:.0f} c50={c50:.0f})\\'], \\'performance\\': {\\'dataset\\': \\'LibriSpeech, AudioSet, EchoThief, MIT-Acoustical-Reverberation-Scene\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Brouhaha is a joint voice activity detection, speech-to-noise ratio, and C50 room acoustics estimation model. It is based on the PyTorch framework and uses the pyannote.audio library.\\'}', metadata={})]", "category": "generic"}
{"question_id": 712, "text": " We have an audio recording with overlapping speakers talking. We need to remove the overlaps and make it a clear single speaker recording.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'pyannote.audio\\', \\'functionality\\': \\'Speaker diarization\\', \\'api_name\\': \\'johnislarry/cloned-pyannote-speaker-diarization-endpoint\\', \\'api_call\\': \"Pipeline.from_pretrained(\\'pyannote/speaker-diarization@2.1\\',use_auth_token=\\'ACCESS_TOKEN_GOES_HERE\\')\", \\'api_arguments\\': [\\'num_speakers\\', \\'min_speakers\\', \\'max_speakers\\', \\'segmentation_onset\\'], \\'python_environment_requirements\\': \\'pyannote.audio 2.0\\', \\'example_code\\': {\\'load_pipeline\\': \\'from pyannote.audio import Pipeline\\\\npipeline = Pipeline.from_pretrained(pyannote/speaker-diarization@2022.07)\\', \\'apply_pipeline\\': \\'diarization = pipeline(audio.wav)\\', \\'save_output\\': \\'with open(audio.rttm, w) as rttm:\\\\n  diarization.write_rttm(rttm)\\'}, \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'AISHELL-4\\', \\'accuracy\\': {\\'DER%\\': 14.61, \\'FA%\\': 3.31, \\'Miss%\\': 4.35, \\'Conf%\\': 6.95}}, {\\'name\\': \\'AMI Mix-Headset only_words\\', \\'accuracy\\': {\\'DER%\\': 18.21, \\'FA%\\': 3.28, \\'Miss%\\': 11.07, \\'Conf%\\': 3.87}}, {\\'name\\': \\'AMI Array1-01 only_words\\', \\'accuracy\\': {\\'DER%\\': 29.0, \\'FA%\\': 2.71, \\'Miss%\\': 21.61, \\'Conf%\\': 4.68}}, {\\'name\\': \\'CALLHOME Part2\\', \\'accuracy\\': {\\'DER%\\': 30.24, \\'FA%\\': 3.71, \\'Miss%\\': 16.86, \\'Conf%\\': 9.66}}, {\\'name\\': \\'DIHARD 3 Full\\', \\'accuracy\\': {\\'DER%\\': 20.99, \\'FA%\\': 4.25, \\'Miss%\\': 10.74, \\'Conf%\\': 6.0}}, {\\'name\\': \\'REPERE Phase 2\\', \\'accuracy\\': {\\'DER%\\': 12.62, \\'FA%\\': 1.55, \\'Miss%\\': 3.3, \\'Conf%\\': 7.76}}, {\\'name\\': \\'VoxConverse v0.0.2\\', \\'accuracy\\': {\\'DER%\\': 12.76, \\'FA%\\': 3.45, \\'Miss%\\': 3.85, \\'Conf%\\': 5.46}}]}, \\'description\\': \\'This API provides speaker diarization functionality using the pyannote.audio framework. It is capable of processing audio files and outputting speaker diarization results in RTTM format. The API supports providing the number of speakers, minimum and maximum number of speakers, and adjusting the segmentation onset threshold.\\'}', metadata={})]", "category": "generic"}
{"question_id": 713, "text": " We have just built an app for hosting podcasts, and we need a speech enhancement model to clean noise from the audio recordings submitted by our users.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'CTranslate2\\', \\'functionality\\': \\'Automatic Speech Recognition\\', \\'api_name\\': \\'guillaumekln/faster-whisper-large-v2\\', \\'api_call\\': \"WhisperModel(\\'large-v2\\')\", \\'api_arguments\\': [\\'audio.mp3\\'], \\'python_environment_requirements\\': [\\'faster_whisper\\'], \\'example_code\\': \\'from faster_whisper import WhisperModel\\\\nmodel = WhisperModel(large-v2)\\\\nsegments, info = model.transcribe(audio.mp3)\\\\nfor segment in segments:\\\\n print([%.2fs -&gt; %.2fs] %s % (segment.start, segment.end, segment.text))\\', \\'performance\\': {\\'dataset\\': \\'99 languages\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Whisper large-v2 model for CTranslate2. This model can be used in CTranslate2 or projets based on CTranslate2 such as faster-whisper.\\'}', metadata={})]", "category": "generic"}
{"question_id": 714, "text": " My coworker sent me a voice message in Spanish. Translate this audio message to English, so that I can understand it.\\n###Input: spanish_voice_message.wav\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'CTranslate2\\', \\'functionality\\': \\'Automatic Speech Recognition\\', \\'api_name\\': \\'guillaumekln/faster-whisper-large-v2\\', \\'api_call\\': \"WhisperModel(\\'large-v2\\')\", \\'api_arguments\\': [\\'audio.mp3\\'], \\'python_environment_requirements\\': [\\'faster_whisper\\'], \\'example_code\\': \\'from faster_whisper import WhisperModel\\\\nmodel = WhisperModel(large-v2)\\\\nsegments, info = model.transcribe(audio.mp3)\\\\nfor segment in segments:\\\\n print([%.2fs -&gt; %.2fs] %s % (segment.start, segment.end, segment.text))\\', \\'performance\\': {\\'dataset\\': \\'99 languages\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Whisper large-v2 model for CTranslate2. This model can be used in CTranslate2 or projets based on CTranslate2 such as faster-whisper.\\'}', metadata={})]", "category": "generic"}
{"question_id": 715, "text": " We are developing a virtual assistant and need to integrate a speech enhancement feature in it.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'SYSPIN/Marathi_Male_TTS\\', \\'api_call\\': \"api.load(\\'ESPnet/espnet_model_zoo:SYSPIN/Marathi_Male_TTS\\').\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'huggingface_hub\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Marathi Male Text-to-Speech model using ESPnet framework.\\'}', metadata={})]", "category": "generic"}
{"question_id": 716, "text": " A kid's educational toy company is including a voice assistant that recognizes simple voice commands from children. They need an example on how to recognize spoken digits.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Voice Activity Detection\\', \\'api_name\\': \\'d4data/Indian-voice-cloning\\', \\'api_call\\': \"pipeline(\\'voice-activity-detection\\', model=\\'d4data/Indian-voice-cloning\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A model for detecting voice activity in Indian languages.\\'}', metadata={})]", "category": "generic"}
{"question_id": 717, "text": " Our company needs to analyze customer phone call recordings and identify specific numbers mentioned by customers during the call.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Voice Activity Detection\\', \\'api_name\\': \\'d4data/Indian-voice-cloning\\', \\'api_call\\': \"pipeline(\\'voice-activity-detection\\', model=\\'d4data/Indian-voice-cloning\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A model for detecting voice activity in Indian languages.\\'}', metadata={})]", "category": "generic"}
{"question_id": 718, "text": " You are building a virtual global tour guide that can identify languages from the audio of people speaking. Use a model to identify which language is being spoken.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'CTranslate2\\', \\'functionality\\': \\'Automatic Speech Recognition\\', \\'api_name\\': \\'guillaumekln/faster-whisper-large-v2\\', \\'api_call\\': \"WhisperModel(\\'large-v2\\')\", \\'api_arguments\\': [\\'audio.mp3\\'], \\'python_environment_requirements\\': [\\'faster_whisper\\'], \\'example_code\\': \\'from faster_whisper import WhisperModel\\\\nmodel = WhisperModel(large-v2)\\\\nsegments, info = model.transcribe(audio.mp3)\\\\nfor segment in segments:\\\\n print([%.2fs -&gt; %.2fs] %s % (segment.start, segment.end, segment.text))\\', \\'performance\\': {\\'dataset\\': \\'99 languages\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Whisper large-v2 model for CTranslate2. This model can be used in CTranslate2 or projets based on CTranslate2 such as faster-whisper.\\'}', metadata={})]", "category": "generic"}
{"question_id": 719, "text": " I am working as the head of customer service for a Spanish speaking market. I want to know the sentiment of my customers on their last call with our support agents.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'finiteautomata/beto-sentiment-analysis\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'finiteautomata/beto-sentiment-analysis\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'Hugging Face Transformers library\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'TASS 2020 corpus\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Model trained with TASS 2020 corpus (around ~5k tweets) of several dialects of Spanish. Base model is BETO, a BERT model trained in Spanish. Uses POS, NEG, NEU labels.\\'}', metadata={})]", "category": "generic"}
{"question_id": 720, "text": " I have a recording of a meeting from which I need to extract the parts where people are talking and eliminate the silent parts.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'CTranslate2\\', \\'functionality\\': \\'Automatic Speech Recognition\\', \\'api_name\\': \\'guillaumekln/faster-whisper-large-v2\\', \\'api_call\\': \"WhisperModel(\\'large-v2\\')\", \\'api_arguments\\': [\\'audio.mp3\\'], \\'python_environment_requirements\\': [\\'faster_whisper\\'], \\'example_code\\': \\'from faster_whisper import WhisperModel\\\\nmodel = WhisperModel(large-v2)\\\\nsegments, info = model.transcribe(audio.mp3)\\\\nfor segment in segments:\\\\n print([%.2fs -&gt; %.2fs] %s % (segment.start, segment.end, segment.text))\\', \\'performance\\': {\\'dataset\\': \\'99 languages\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Whisper large-v2 model for CTranslate2. This model can be used in CTranslate2 or projets based on CTranslate2 such as faster-whisper.\\'}', metadata={})]", "category": "generic"}
{"question_id": 721, "text": " We are a company producing wine. Based on the chemical properties of our wine data, we need to analyze the quality of our products and determine whether they are good or bad.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Scikit-learn\\', \\'functionality\\': \\'Wine Quality classification\\', \\'api_name\\': \\'julien-c/wine-quality\\', \\'api_call\\': \"joblib.load(cached_download(hf_hub_url(\\'julien-c/wine-quality\\', \\'winequality-red.csv\\')))\", \\'api_arguments\\': [\\'X\\'], \\'python_environment_requirements\\': [\\'huggingface_hub\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \\'from huggingface_hub import hf_hub_url, cached_download\\\\nimport joblib\\\\nimport pandas as pd\\\\nREPO_ID = julien-c/wine-quality\\\\nFILENAME = sklearn_model.joblib\\\\nmodel = joblib.load(cached_download(\\\\n hf_hub_url(REPO_ID, FILENAME)\\\\n))\\\\ndata_file = cached_download(\\\\n hf_hub_url(REPO_ID, winequality-red.csv)\\\\n)\\\\nwinedf = pd.read_csv(data_file, sep=;)\\\\nX = winedf.drop([quality], axis=1)\\\\nY = winedf[quality]\\\\nprint(X[:3])\\\\nlabels = model.predict(X[:3])\\\\nmodel.score(X, Y)\\', \\'performance\\': {\\'dataset\\': \\'julien-c/wine-quality\\', \\'accuracy\\': 0.6616635397123202}, \\'description\\': \\'A Simple Example of Scikit-learn Pipeline for Wine Quality classification. Inspired by https://towardsdatascience.com/a-simple-example-of-pipeline-in-machine-learning-with-scikit-learn-e726ffbb6976 by Saptashwa Bhattacharyya.\\'}', metadata={})]", "category": "generic"}
{"question_id": 722, "text": " Our team is working on a podcast app, and we want to automatically generate transcripts with timestamps for each speaker.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'imdanboy/jets\\', \\'api_call\\': \"pipeline(\\'text-to-speech\\', model=\\'imdanboy/jets\\')\", \\'api_arguments\\': None, \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"from transformers import pipeline; tts = pipeline(\\'text-to-speech\\', model=\\'imdanboy/jets\\'); tts(\\'Hello world\\')\", \\'performance\\': {\\'dataset\\': \\'ljspeech\\', \\'accuracy\\': None}, \\'description\\': \\'This model was trained by imdanboy using ljspeech recipe in espnet.\\'}', metadata={})]", "category": "generic"}
{"question_id": 723, "text": " I have a large dataset with both numerical and categorical features related to customer behavior. I want to classify them into different segments for targeted marketing. How do I use a TabTransformer for this task?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Keras\\', \\'functionality\\': \\'Structured data learning with TabTransformer\\', \\'api_name\\': \\'keras-io/tab_transformer\\', \\'api_call\\': \\'TabTransformer.from_config()\\', \\'api_arguments\\': \\'N/A\\', \\'python_environment_requirements\\': \\'Hugging Face\\', \\'example_code\\': \\'N/A\\', \\'performance\\': {\\'dataset\\': \\'United States Census Income Dataset\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \"This model uses self-attention based Transformers structure followed by multiple feed forward layers to serve supervised and semi-supervised learning tasks on tabular data. The model\\'s inputs can contain both numerical and categorical features. Categorical features are encoded into embedding vectors before being fed into a stack of Transformer blocks. The contextual embeddings of the categorical features after the final Transformer layer are concatenated with the input numerical features and fed into a final MLP block. A SoftMax function is applied at the end of the model.\"}', metadata={})]", "category": "generic"}
{"question_id": 724, "text": " Determine if an employee's annual salary meets or exceeds $50000.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Keras\\', \\'functionality\\': \\'Binary Classification\\', \\'api_name\\': \\'TF_Decision_Trees\\', \\'api_call\\': \\'TF_Decision_Trees(input_features, target)\\', \\'api_arguments\\': [\\'input_features\\', \\'target\\'], \\'python_environment_requirements\\': [\\'tensorflow >= 7.0\\'], \\'example_code\\': \\'https://github.com/tdubon/TF-GB-Forest/blob/c0cf4c7e3e29d819b996cfe4eecc1f2728115e52/TFDecisionTrees_Final.ipynb\\', \\'performance\\': {\\'dataset\\': \\'Census-Income Data Set\\', \\'accuracy\\': 96.57}, \\'description\\': \"Use TensorFlow\\'s Gradient Boosted Trees model in binary classification of structured data. Build a decision forests model by specifying the input feature usage. Implement a custom Binary Target encoder as a Keras Preprocessing layer to encode the categorical features with respect to their target value co-occurrences, and then use the encoded features to build a decision forests model. The model is trained on the US Census Income Dataset containing approximately 300k instances with 41 numerical and categorical variables. The task is to determine whether a person makes over 50k a year.\"}', metadata={})]", "category": "generic"}
{"question_id": 725, "text": " Identify whether a newly-setup chemical plant is exceeding carbon emission limits based on a CSV file containing data collected.\\n###Input: data.csv\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Joblib\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'pachi107/autotrain-in-class-test-1780161764\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data.csv\\'], \\'python_environment_requirements\\': [\\'json\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(data.csv)\\\\ndata = data[features]\\\\ndata.columns = [feat_ + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'Validation Metrics\\', \\'accuracy\\': 0.974}, \\'description\\': \\'A binary classification model for predicting CO2 emissions based on tabular data. Trained using AutoTrain with a model ID of 1780161764.\\'}', metadata={})]", "category": "generic"}
{"question_id": 726, "text": " I am working in a bank, I want to estimate the mortgage for a given housing using the housing's features.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Keras\\', \\'functionality\\': \\'Binary Classification\\', \\'api_name\\': \\'TF_Decision_Trees\\', \\'api_call\\': \\'TF_Decision_Trees(input_features, target)\\', \\'api_arguments\\': [\\'input_features\\', \\'target\\'], \\'python_environment_requirements\\': [\\'tensorflow >= 7.0\\'], \\'example_code\\': \\'https://github.com/tdubon/TF-GB-Forest/blob/c0cf4c7e3e29d819b996cfe4eecc1f2728115e52/TFDecisionTrees_Final.ipynb\\', \\'performance\\': {\\'dataset\\': \\'Census-Income Data Set\\', \\'accuracy\\': 96.57}, \\'description\\': \"Use TensorFlow\\'s Gradient Boosted Trees model in binary classification of structured data. Build a decision forests model by specifying the input feature usage. Implement a custom Binary Target encoder as a Keras Preprocessing layer to encode the categorical features with respect to their target value co-occurrences, and then use the encoded features to build a decision forests model. The model is trained on the US Census Income Dataset containing approximately 300k instances with 41 numerical and categorical variables. The task is to determine whether a person makes over 50k a year.\"}', metadata={})]", "category": "generic"}
{"question_id": 727, "text": " We are building a robot for hopping in a controlled environment. Train it to perform a hop using Decision Transformers.\\n**Input:<noinput>**\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Reinforcement Learning\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'decision-transformer-gym-hopper-medium\\', \\'api_call\\': \"AutoModel.from_pretrained(\\'edbeeching/decision-transformer-gym-hopper-medium\\')\", \\'api_arguments\\': {\\'mean\\': [1.311279, -0.08469521, -0.5382719, -0.07201576, 0.04932366, 2.1066856, -0.15017354, 0.00878345, -0.2848186, -0.18540096, -0.28461286], \\'std\\': [0.17790751, 0.05444621, 0.21297139, 0.14530419, 0.6124444, 0.85174465, 1.4515252, 0.6751696, 1.536239, 1.6160746, 5.6072536]}, \\'python_environment_requirements\\': {\\'transformers\\': \\'latest\\'}, \\'example_code\\': \\'See our Blog Post, Colab notebook or Example Script for usage.\\', \\'performance\\': {\\'dataset\\': \\'Gym Hopper environment\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Decision Transformer model trained on medium trajectories sampled from the Gym Hopper environment.\\'}', metadata={})]", "category": "generic"}
{"question_id": 728, "text": " We are building a robot for elderly care. The robot should be able to understand what's happening and take appropriate actions based on the elderly's current activities.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'hyunwoongko/blenderbot-9B\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'hyunwoongko/blenderbot-9B\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'Input a message to start chatting with hyunwoongko/blenderbot-9B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high-performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi-turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 729, "text": " To build a source code recommendation engine, we need to extract features from text and code using Transformer models.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Program Synthesis\\', \\'api_name\\': \\'Salesforce/codegen-350M-multi\\', \\'api_call\\': \"AutoTokenizer.from_pretrained(\\'Salesforce/codegen-350M-multi\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'from transformers import AutoTokenizer, AutoModelForCausalLM\\\\ntokenizer = AutoTokenizer.from_pretrained(Salesforce/codegen-350M-multi)\\\\nmodel = AutoModelForCausalLM.from_pretrained(Salesforce/codegen-350M-multi)\\\\ntext = def hello_world():\\\\ninput_ids = tokenizer(text, return_tensors=pt).input_ids\\\\ngenerated_ids = model.generate(input_ids, max_length=128)\\\\nprint(tokenizer.decode(generated_ids[0], skip_special_tokens=True))\\', \\'performance\\': {\\'dataset\\': \\'HumanEval and MTPB\\', \\'accuracy\\': \\'Refer to the paper for accuracy details\\'}, \\'description\\': \\'CodeGen is a family of autoregressive language models for program synthesis. The checkpoint included in this repository is denoted as CodeGen-Multi 350M, where Multi means the model is initialized with CodeGen-NL 350M and further pre-trained on a dataset of multiple programming languages, and 350M refers to the number of trainable parameters. The model is capable of extracting features from given natural language and programming language texts, and calculating the likelihood of them. It is best at program synthesis, generating executable code given English prompts, and can complete partially-generated code as well.\\'}', metadata={})]", "category": "generic"}
{"question_id": 730, "text": " Our team needs to prepare a feature matrix based on the given source code's programming constructs and comments for data analysis.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'tabular regression\\', \\'api_name\\': \\'farouk97/autotrain-test7-2644pc-linearregr-38619101723\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'farouk97/autotrain-data-test7-2644pc-linearregr\\', \\'accuracy\\': {\\'Loss\\': 0.145, \\'R2\\': 0.0, \\'MSE\\': 0.021, \\'MAE\\': 0.099, \\'RMSLE\\': 0.101}}, \\'description\\': \\'A tabular regression model trained using AutoTrain to predict CO2 emissions (in grams).\\'}', metadata={})]", "category": "generic"}
{"question_id": 731, "text": " We are developing an interactive story app that would display a picture of each character as mentioned in the story given to our GPT-3 model.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'prompthero/openjourney-v4\\', \\'api_call\\': \"pipeline(\\'text-to-image\\', model=\\'prompthero/openjourney-v4\\')\", \\'api_arguments\\': {\\'text\\': \\'string\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"generate_image(\\'your text here\\')\", \\'performance\\': {\\'dataset\\': \\'Midjourney v4 images\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Openjourney v4 is trained on +124k Midjourney v4 images by PromptHero. It is used for generating images based on text inputs.\\'}', metadata={})]", "category": "generic"}
{"question_id": 732, "text": " As a landscape architect, I want to generate a description of an image of a park I've designed to use for promotional purposes.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'prompthero/openjourney-v4\\', \\'api_call\\': \"pipeline(\\'text-to-image\\', model=\\'prompthero/openjourney-v4\\')\", \\'api_arguments\\': {\\'text\\': \\'string\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"generate_image(\\'your text here\\')\", \\'performance\\': {\\'dataset\\': \\'Midjourney v4 images\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Openjourney v4 is trained on +124k Midjourney v4 images by PromptHero. It is used for generating images based on text inputs.\\'}', metadata={})]", "category": "generic"}
{"question_id": 733, "text": " I need to build a solution that generates textual descriptions for images in my database.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Image-to-Text\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Captioning\\', \\'api_name\\': \\'microsoft/git-base\\', \\'api_call\\': \"pipeline(\\'image-to-text\\', model=\\'microsoft/git-base\\')\", \\'api_arguments\\': \\'image\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'git_base(image)\\', \\'performance\\': {\\'dataset\\': [\\'COCO\\', \\'Conceptual Captions (CC3M)\\', \\'SBU\\', \\'Visual Genome (VG)\\', \\'Conceptual Captions (CC12M)\\', \\'ALT200M\\'], \\'accuracy\\': \\'Refer to the paper for evaluation results\\'}, \\'description\\': \"GIT (short for GenerativeImage2Text) model, base-sized version. It was introduced in the paper GIT: A Generative Image-to-text Transformer for Vision and Language by Wang et al. and first released in this repository. The model is trained using \\'teacher forcing\\' on a lot of (image, text) pairs. The goal for the model is simply to predict the next text token, giving the image tokens and previous text tokens. This allows the model to be used for tasks like image and video captioning, visual question answering (VQA) on images and videos, and even image classification (by simply conditioning the model on the image and asking it to generate a class for it in text).\"}', metadata={})]", "category": "generic"}
{"question_id": 734, "text": " We are a company managing hotel bookings. We need to answer our customer's questions regarding rental rates from our pricing document.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'impira/layoutlm-invoices\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=\\'impira/layoutlm-invoices\\')\", \\'api_arguments\\': \\'question, context\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"qa_pipeline(question=\\'your question\\', context=\\'your document context\\')\", \\'performance\\': {\\'dataset\\': \\'proprietary dataset of invoices, SQuAD2.0, and DocVQA\\', \\'accuracy\\': \\'not provided\\'}, \\'description\\': \\'This is a fine-tuned version of the multi-modal LayoutLM model for the task of question answering on invoices and other documents. It has been fine-tuned on a proprietary dataset of invoices as well as both SQuAD2.0 and DocVQA for general comprehension. Unlike other QA models, which can only extract consecutive tokens (because they predict the start and end of a sequence), this model can predict longer-range, non-consecutive sequences with an additional classifier head.\\'}', metadata={})]", "category": "generic"}
{"question_id": 735, "text": " Provide me a pipeline for extracting relevant information from an insurance policy document.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'impira/layoutlm-invoices\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=\\'impira/layoutlm-invoices\\')\", \\'api_arguments\\': \\'question, context\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"qa_pipeline(question=\\'your question\\', context=\\'your document context\\')\", \\'performance\\': {\\'dataset\\': \\'proprietary dataset of invoices, SQuAD2.0, and DocVQA\\', \\'accuracy\\': \\'not provided\\'}, \\'description\\': \\'This is a fine-tuned version of the multi-modal LayoutLM model for the task of question answering on invoices and other documents. It has been fine-tuned on a proprietary dataset of invoices as well as both SQuAD2.0 and DocVQA for general comprehension. Unlike other QA models, which can only extract consecutive tokens (because they predict the start and end of a sequence), this model can predict longer-range, non-consecutive sequences with an additional classifier head.\\'}', metadata={})]", "category": "generic"}
{"question_id": 736, "text": " Create an accurate system to estimate the distance between the camera and the objects in a photo captured by a drone flying over a terrain.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Depth Estimation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Depth Estimation\\', \\'api_name\\': \\'glpn-kitti\\', \\'api_call\\': \"GLPNForDepthEstimation.from_pretrained(\\'vinvino02/glpn-kitti\\')\", \\'api_arguments\\': \\'images, return_tensors\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'from transformers import GLPNFeatureExtractor, GLPNForDepthEstimation\\\\nimport torch\\\\nimport numpy as np\\\\nfrom PIL import Image\\\\nimport requests\\\\nurl = http://images.cocodataset.org/val2017/000000039769.jpg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nfeature_extractor = GLPNFeatureExtractor.from_pretrained(vinvino02/glpn-kitti)\\\\nmodel = GLPNForDepthEstimation.from_pretrained(vinvino02/glpn-kitti)\\\\ninputs = feature_extractor(images=image, return_tensors=pt)\\\\nwith torch.no_grad():\\\\n outputs = model(**inputs)\\\\n predicted_depth = outputs.predicted_depth\\\\nprediction = torch.nn.functional.interpolate(\\\\n predicted_depth.unsqueeze(1),\\\\n size=image.size[::-1],\\\\n mode=bicubic,\\\\n align_corners=False,\\\\n)\\\\noutput = prediction.squeeze().cpu().numpy()\\\\nformatted = (output * 255 / np.max(output)).astype(uint8)\\\\ndepth = Image.fromarray(formatted)\\', \\'performance\\': {\\'dataset\\': \\'KITTI\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Global-Local Path Networks (GLPN) model trained on KITTI for monocular depth estimation. It was introduced in the paper Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth by Kim et al. and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 737, "text": " We are a team of architects and civil engineers looking to estimate the depth of elements in architectural designs from 2D images of the structures. We want to implement a depth estimation model that will transform these images into depictions of depth.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Depth Estimation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'dpt-large-redesign\\', \\'api_call\\': \"AutoModelForDepthEstimation.from_pretrained(\\'nielsr/dpt-large-redesign\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'torch\\', \\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A depth estimation model based on the DPT architecture.\\'}', metadata={})]", "category": "generic"}
{"question_id": 738, "text": " I have an image on my computer named \\\"eye.jpg\\\". I want to use a model to predict whether this image indicates diabetic retinopathy.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'martinezomg/vit-base-patch16-224-diabetic-retinopathy\\', \\'api_call\\': \"pipeline(\\'image-classification\\', \\'martinezomg/vit-base-patch16-224-diabetic-retinopathy\\')\", \\'api_arguments\\': {\\'model_name\\': \\'martinezomg/vit-base-patch16-224-diabetic-retinopathy\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'4.28.1\\', \\'pytorch\\': \\'2.0.0+cu118\\', \\'datasets\\': \\'2.11.0\\', \\'tokenizers\\': \\'0.13.3\\'}, \\'example_code\\': \"from transformers import pipeline\\\\nimage_classifier = pipeline(\\'image-classification\\', \\'martinezomg/vit-base-patch16-224-diabetic-retinopathy\\')\\\\nresult = image_classifier(\\'path/to/image.jpg\\')\", \\'performance\\': {\\'dataset\\': \\'None\\', \\'accuracy\\': 0.7744}, \\'description\\': \\'This model is a fine-tuned version of google/vit-base-patch16-224 on the None dataset. It is designed for image classification tasks, specifically for diabetic retinopathy detection.\\'}', metadata={})]", "category": "generic"}
{"question_id": 739, "text": " We need to build a model that recognizes objects in images. Build a model using the Vision Transformer (ViT) for this purpose.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'google/vit-base-patch16-384\\', \\'api_call\\': \"ViTForImageClassification.from_pretrained(\\'google/vit-base-patch16-384\\')\", \\'api_arguments\\': {\\'pretrained_model_name_or_path\\': \\'google/vit-base-patch16-384\\'}, \\'python_environment_requirements\\': [\\'transformers\\', \\'PIL\\', \\'requests\\'], \\'example_code\\': \"from transformers import ViTFeatureExtractor, ViTForImageClassification\\\\nfrom PIL import Image\\\\nimport requests\\\\nurl = \\'http://images.cocodataset.org/val2017/000000039769.jpg\\'\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nfeature_extractor = ViTFeatureExtractor.from_pretrained(\\'google/vit-base-patch16-384\\')\\\\nmodel = ViTForImageClassification.from_pretrained(\\'google/vit-base-patch16-384\\')\\\\ninputs = feature_extractor(images=image, return_tensors=pt)\\\\noutputs = model(**inputs)\\\\nlogits = outputs.logits\\\\npredicted_class_idx = logits.argmax(-1).item()\\\\nprint(Predicted class:, model.config.id2label[predicted_class_idx])\", \\'performance\\': {\\'dataset\\': \\'ImageNet\\', \\'accuracy\\': \\'Refer to tables 2 and 5 of the original paper\\'}, \\'description\\': \\'Vision Transformer (ViT) model pre-trained on ImageNet-21k (14 million images, 21,843 classes) at resolution 224x224, and fine-tuned on ImageNet 2012 (1 million images, 1,000 classes) at resolution 384x384. It was introduced in the paper An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale by Dosovitskiy et al. Images are presented to the model as a sequence of fixed-size patches (resolution 16x16), which are linearly embedded. One also adds a [CLS] token to the beginning of a sequence to use it for classification tasks. One also adds absolute position embeddings before feeding the sequence to the layers of the Transformer encoder.\\'}', metadata={})]", "category": "generic"}
{"question_id": 740, "text": " My pet store website needs a tool to recognize different dog breeds from user uploaded images.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'abhishek/autotrain-dog-vs-food\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'abhishek/autotrain-dog-vs-food\\')\", \\'api_arguments\\': \\'image_path\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'sasha/dog-food\\', \\'accuracy\\': 0.998}, \\'description\\': \"A pre-trained model for classifying images as either dog or food using Hugging Face\\'s AutoTrain framework.\"}', metadata={})]", "category": "generic"}
{"question_id": 741, "text": " As a programmer, I want to classify different animal species based on their images. Help me determine the type of species. \\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'abhishek/autotrain-dog-vs-food\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'abhishek/autotrain-dog-vs-food\\')\", \\'api_arguments\\': \\'image_path\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'sasha/dog-food\\', \\'accuracy\\': 0.998}, \\'description\\': \"A pre-trained model for classifying images as either dog or food using Hugging Face\\'s AutoTrain framework.\"}', metadata={})]", "category": "generic"}
{"question_id": 742, "text": " We are a robotics company specializing in object detection for logistics. We need to detect objects in warehouses using an advanced object detection model.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Object Detection\\', \\'api_name\\': \\'keremberke/yolov8m-forklift-detection\\', \\'api_call\\': \"YOLO(\\'keremberke/yolov8m-forklift-detection\\')\", \\'api_arguments\\': {\\'image\\': \\'URL or local path to the image\\'}, \\'python_environment_requirements\\': [\\'ultralyticsplus==0.0.23\\', \\'ultralytics==8.0.21\\'], \\'example_code\\': [\\'from ultralyticsplus import YOLO, render_result\\', \"model = YOLO(\\'keremberke/yolov8m-forklift-detection\\')\", \"model.overrides[\\'conf\\'] = 0.25\", \"model.overrides[\\'iou\\'] = 0.45\", \"model.overrides[\\'agnostic_nms\\'] = False\", \"model.overrides[\\'max_det\\'] = 1000\", \"image = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\", \\'results = model.predict(image)\\', \\'print(results[0].boxes)\\', \\'render = render_result(model=model, image=image, result=results[0])\\', \\'render.show()\\'], \\'performance\\': {\\'dataset\\': \\'forklift-object-detection\\', \\'accuracy\\': 0.846}, \\'description\\': \\'A YOLOv8 model for detecting forklifts and persons in images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 743, "text": " An IoT device collects images from different locations. Create a model to detect objects in these images to analyze the surroundings.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Object Detection\\', \\'api_name\\': \\'keremberke/yolov8m-forklift-detection\\', \\'api_call\\': \"YOLO(\\'keremberke/yolov8m-forklift-detection\\')\", \\'api_arguments\\': {\\'image\\': \\'URL or local path to the image\\'}, \\'python_environment_requirements\\': [\\'ultralyticsplus==0.0.23\\', \\'ultralytics==8.0.21\\'], \\'example_code\\': [\\'from ultralyticsplus import YOLO, render_result\\', \"model = YOLO(\\'keremberke/yolov8m-forklift-detection\\')\", \"model.overrides[\\'conf\\'] = 0.25\", \"model.overrides[\\'iou\\'] = 0.45\", \"model.overrides[\\'agnostic_nms\\'] = False\", \"model.overrides[\\'max_det\\'] = 1000\", \"image = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\", \\'results = model.predict(image)\\', \\'print(results[0].boxes)\\', \\'render = render_result(model=model, image=image, result=results[0])\\', \\'render.show()\\'], \\'performance\\': {\\'dataset\\': \\'forklift-object-detection\\', \\'accuracy\\': 0.846}, \\'description\\': \\'A YOLOv8 model for detecting forklifts and persons in images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 744, "text": " I'm building a drone surveillance system for detecting airplanes in the sky. What can I use to achieve the required object detection?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Object Detection\\', \\'api_name\\': \\'keremberke/yolov8m-plane-detection\\', \\'api_call\\': \"YOLO(\\'keremberke/yolov8m-plane-detection\\')\", \\'api_arguments\\': {\\'image\\': \\'URL or local path to the image\\'}, \\'python_environment_requirements\\': [\\'pip install ultralyticsplus==0.0.23 ultralytics==8.0.21\\'], \\'example_code\\': [\\'from ultralyticsplus import YOLO, render_result\\', \"model = YOLO(\\'keremberke/yolov8m-plane-detection\\')\", \"model.overrides[\\'conf\\'] = 0.25\", \"model.overrides[\\'iou\\'] = 0.45\", \"model.overrides[\\'agnostic_nms\\'] = False\", \"model.overrides[\\'max_det\\'] = 1000\", \"image = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\", \\'results = model.predict(image)\\', \\'print(results[0].boxes)\\', \\'render = render_result(model=model, image=image, result=results[0])\\', \\'render.show()\\'], \\'performance\\': {\\'dataset\\': \\'plane-detection\\', \\'accuracy\\': \\'0.995\\'}, \\'description\\': \\'A YOLOv8 model for plane detection trained on the keremberke/plane-detection dataset. The model is capable of detecting planes in images with high accuracy.\\'}', metadata={})]", "category": "generic"}
{"question_id": 745, "text": " We are building an IoT device to monitor apartment corridors for security. Please detect objects in the image and notify if anything abnormal is detected.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Object Detection\\', \\'api_name\\': \\'keremberke/yolov8n-csgo-player-detection\\', \\'api_call\\': \"YOLO(\\'keremberke/yolov8n-csgo-player-detection\\')\", \\'api_arguments\\': {\\'image\\': \\'URL or local path to image\\'}, \\'python_environment_requirements\\': \\'pip install ultralyticsplus==0.0.23 ultralytics==8.0.21\\', \\'example_code\\': \"from ultralyticsplus import YOLO, render_result\\\\n\\\\nmodel = YOLO(\\'keremberke/yolov8n-csgo-player-detection\\')\\\\n\\\\nmodel.overrides[\\'conf\\'] = 0.25\\\\nmodel.overrides[\\'iou\\'] = 0.45\\\\nmodel.overrides[\\'agnostic_nms\\'] = False\\\\nmodel.overrides[\\'max_det\\'] = 1000\\\\n\\\\nimage = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\\\\n\\\\nresults = model.predict(image)\\\\n\\\\nprint(results[0].boxes)\\\\nrender = render_result(model=model, image=image, result=results[0])\\\\nrender.show()\", \\'performance\\': {\\'dataset\\': \\'csgo-object-detection\\', \\'accuracy\\': 0.844}, \\'description\\': \"A YOLOv8 model for detecting Counter-Strike: Global Offensive (CS:GO) players with supported labels: [\\'ct\\', \\'cthead\\', \\'t\\', \\'thead\\'].\"}', metadata={})]", "category": "generic"}
{"question_id": 746, "text": " Design an application for a smart drone that can detect and segment objects within the area it's flying.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Object Detection\\', \\'api_name\\': \\'keremberke/yolov8m-plane-detection\\', \\'api_call\\': \"YOLO(\\'keremberke/yolov8m-plane-detection\\')\", \\'api_arguments\\': {\\'image\\': \\'URL or local path to the image\\'}, \\'python_environment_requirements\\': [\\'pip install ultralyticsplus==0.0.23 ultralytics==8.0.21\\'], \\'example_code\\': [\\'from ultralyticsplus import YOLO, render_result\\', \"model = YOLO(\\'keremberke/yolov8m-plane-detection\\')\", \"model.overrides[\\'conf\\'] = 0.25\", \"model.overrides[\\'iou\\'] = 0.45\", \"model.overrides[\\'agnostic_nms\\'] = False\", \"model.overrides[\\'max_det\\'] = 1000\", \"image = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\", \\'results = model.predict(image)\\', \\'print(results[0].boxes)\\', \\'render = render_result(model=model, image=image, result=results[0])\\', \\'render.show()\\'], \\'performance\\': {\\'dataset\\': \\'plane-detection\\', \\'accuracy\\': \\'0.995\\'}, \\'description\\': \\'A YOLOv8 model for plane detection trained on the keremberke/plane-detection dataset. The model is capable of detecting planes in images with high accuracy.\\'}', metadata={})]", "category": "generic"}
{"question_id": 747, "text": " I want to build a tool to recognize urban landscapes and identify different objects in the image.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Zero-Shot Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Geolocalization\\', \\'api_name\\': \\'geolocal/StreetCLIP\\', \\'api_call\\': \"CLIPModel.from_pretrained(\\'geolocal/StreetCLIP\\')\", \\'api_arguments\\': {\\'pretrained_model_name_or_path\\': \\'geolocal/StreetCLIP\\'}, \\'python_environment_requirements\\': [\\'transformers\\', \\'PIL\\', \\'requests\\'], \\'example_code\\': \\'from PIL import Image\\\\nimport requests\\\\nfrom transformers import CLIPProcessor, CLIPModel\\\\nmodel = CLIPModel.from_pretrained(geolocal/StreetCLIP)\\\\nprocessor = CLIPProcessor.from_pretrained(geolocal/StreetCLIP)\\\\nurl = https://huggingface.co/geolocal/StreetCLIP/resolve/main/sanfrancisco.jpeg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nchoices = [San Jose, San Diego, Los Angeles, Las Vegas, San Francisco]\\\\ninputs = processor(text=choices, images=image, return_tensors=pt, padding=True)\\\\noutputs = model(**inputs)\\\\nlogits_per_image = outputs.logits_per_image\\\\nprobs = logits_per_image.softmax(dim=1)\\', \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'IM2GPS\\', \\'accuracy\\': {\\'25km\\': 28.3, \\'200km\\': 45.1, \\'750km\\': 74.7, \\'2500km\\': 88.2}}, {\\'name\\': \\'IM2GPS3K\\', \\'accuracy\\': {\\'25km\\': 22.4, \\'200km\\': 37.4, \\'750km\\': 61.3, \\'2500km\\': 80.4}}]}, \\'description\\': \\'StreetCLIP is a robust foundation model for open-domain image geolocalization and other geographic and climate-related tasks. Trained on an original dataset of 1.1 million street-level urban and rural geo-tagged images, it achieves state-of-the-art performance on multiple open-domain image geolocalization benchmarks in zero-shot, outperforming supervised models trained on millions of images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 748, "text": " Show me how to detect defects of PCB boards from an image in real-time.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Segmentation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Segmentation\\', \\'api_name\\': \\'keremberke/yolov8s-pcb-defect-segmentation\\', \\'api_call\\': \"YOLO(\\'keremberke/yolov8s-pcb-defect-segmentation\\')\", \\'api_arguments\\': {\\'image\\': \\'URL or local path to image\\'}, \\'python_environment_requirements\\': [\\'ultralyticsplus==0.0.23\\', \\'ultralytics==8.0.21\\'], \\'example_code\\': \"from ultralyticsplus import YOLO, render_result\\\\n\\\\nmodel = YOLO(\\'keremberke/yolov8s-pcb-defect-segmentation\\')\\\\nmodel.overrides[\\'conf\\'] = 0.25\\\\nmodel.overrides[\\'iou\\'] = 0.45\\\\nmodel.overrides[\\'agnostic_nms\\'] = False\\\\nmodel.overrides[\\'max_det\\'] = 1000\\\\n\\\\nimage = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\\\\nresults = model.predict(image)\\\\nprint(results[0].boxes)\\\\nprint(results[0].masks)\\\\nrender = render_result(model=model, image=image, result=results[0])\\\\nrender.show()\", \\'performance\\': {\\'dataset\\': \\'pcb-defect-segmentation\\', \\'accuracy\\': {\\'mAP@0.5(box)\\': 0.515, \\'mAP@0.5(mask)\\': 0.491}}, \\'description\\': \\'YOLOv8s model for PCB defect segmentation. The model is trained to detect and segment PCB defects such as Dry_joint, Incorrect_installation, PCB_damage, and Short_circuit.\\'}', metadata={})]", "category": "generic"}
{"question_id": 749, "text": " I want to assess the condition of roads in a city by analyzing drone footage. Create an image segmentation model to identify potholes in images.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Segmentation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Segmentation\\', \\'api_name\\': \\'keremberke/yolov8s-pothole-segmentation\\', \\'api_call\\': \"YOLO(\\'keremberke/yolov8s-pothole-segmentation\\')\", \\'api_arguments\\': {\\'image\\': \\'URL or local path to the image\\'}, \\'python_environment_requirements\\': {\\'ultralyticsplus\\': \\'0.0.23\\', \\'ultralytics\\': \\'8.0.21\\'}, \\'example_code\\': \"from ultralyticsplus import YOLO, render_result\\\\nmodel = YOLO(\\'keremberke/yolov8s-pothole-segmentation\\')\\\\nmodel.overrides[\\'conf\\'] = 0.25\\\\nmodel.overrides[\\'iou\\'] = 0.45\\\\nmodel.overrides[\\'agnostic_nms\\'] = False\\\\nmodel.overrides[\\'max_det\\'] = 1000\\\\nimage = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\\\\nresults = model.predict(image)\\\\nprint(results[0].boxes)\\\\nprint(results[0].masks)\\\\nrender = render_result(model=model, image=image, result=results[0])\\\\nrender.show()\", \\'performance\\': {\\'dataset\\': \\'pothole-segmentation\\', \\'accuracy\\': {\\'mAP@0.5(box)\\': 0.928, \\'mAP@0.5(mask)\\': 0.928}}, \\'description\\': \\'A YOLOv8 model for pothole segmentation. This model detects potholes in images and outputs bounding boxes and masks for the detected potholes.\\'}', metadata={})]", "category": "generic"}
{"question_id": 750, "text": " We are building a road maintenance reporting application. We need to use images to identify and segment road potholes.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Segmentation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Segmentation\\', \\'api_name\\': \\'keremberke/yolov8s-pothole-segmentation\\', \\'api_call\\': \"YOLO(\\'keremberke/yolov8s-pothole-segmentation\\')\", \\'api_arguments\\': {\\'image\\': \\'URL or local path to the image\\'}, \\'python_environment_requirements\\': {\\'ultralyticsplus\\': \\'0.0.23\\', \\'ultralytics\\': \\'8.0.21\\'}, \\'example_code\\': \"from ultralyticsplus import YOLO, render_result\\\\nmodel = YOLO(\\'keremberke/yolov8s-pothole-segmentation\\')\\\\nmodel.overrides[\\'conf\\'] = 0.25\\\\nmodel.overrides[\\'iou\\'] = 0.45\\\\nmodel.overrides[\\'agnostic_nms\\'] = False\\\\nmodel.overrides[\\'max_det\\'] = 1000\\\\nimage = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\\\\nresults = model.predict(image)\\\\nprint(results[0].boxes)\\\\nprint(results[0].masks)\\\\nrender = render_result(model=model, image=image, result=results[0])\\\\nrender.show()\", \\'performance\\': {\\'dataset\\': \\'pothole-segmentation\\', \\'accuracy\\': {\\'mAP@0.5(box)\\': 0.928, \\'mAP@0.5(mask)\\': 0.928}}, \\'description\\': \\'A YOLOv8 model for pothole segmentation. This model detects potholes in images and outputs bounding boxes and masks for the detected potholes.\\'}', metadata={})]", "category": "generic"}
{"question_id": 751, "text": " The marketing team needs different variations of a product image to use in advertising and promotional materials.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Image Variations\\', \\'api_name\\': \\'lambdalabs/sd-image-variations-diffusers\\', \\'api_call\\': \"StableDiffusionImageVariationPipeline.from_pretrained(\\'lambdalabs/sd-image-variations-diffusers\\', revision=\\'v2.0\\')\", \\'api_arguments\\': {\\'revision\\': \\'v2.0\\'}, \\'python_environment_requirements\\': \\'Diffusers >=0.8.0\\', \\'example_code\\': \\'from diffusers import StableDiffusionImageVariationPipeline\\\\nfrom PIL import Image\\\\ndevice = cuda:0\\\\nsd_pipe = StableDiffusionImageVariationPipeline.from_pretrained(\\\\n lambdalabs/sd-image-variations-diffusers,\\\\n revision=v2.0,\\\\n)\\\\nsd_pipe = sd_pipe.to(device)\\\\nim = Image.open(path/to/image.jpg)\\\\ntform = transforms.Compose([\\\\n transforms.ToTensor(),\\\\n transforms.Resize(\\\\n  (224, 224),\\\\n  interpolation=transforms.InterpolationMode.BICUBIC,\\\\n  antialias=False,\\\\n ),\\\\n transforms.Normalize(\\\\n  [0.48145466, 0.4578275, 0.40821073],\\\\n  [0.26862954, 0.26130258, 0.27577711]),\\\\n])\\\\ninp = tform(im).to(device).unsqueeze(0)\\\\nout = sd_pipe(inp, guidance_scale=3)\\\\nout[images][0].save(result.jpg)\\', \\'performance\\': {\\'dataset\\': \\'ChristophSchuhmann/improved_aesthetics_6plus\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'This version of Stable Diffusion has been fine tuned from CompVis/stable-diffusion-v1-4-original to accept CLIP image embedding rather than text embeddings. This allows the creation of image variations similar to DALLE-2 using Stable Diffusion.\\'}', metadata={})]", "category": "generic"}
{"question_id": 752, "text": " A movie studio needs to estimate the human pose of an actor from an image for an upcoming film project.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Human Pose Estimation\\', \\'api_name\\': \\'lllyasviel/sd-controlnet-openpose\\', \\'api_call\\': \"ControlNetModel.from_pretrained(\\'lllyasviel/sd-controlnet-openpose\\')\", \\'api_arguments\\': {\\'text\\': \\'chef in the kitchen\\', \\'image\\': \\'image\\', \\'num_inference_steps\\': 20}, \\'python_environment_requirements\\': {\\'diffusers\\': \\'pip install diffusers\\', \\'transformers\\': \\'pip install transformers\\', \\'accelerate\\': \\'pip install accelerate\\', \\'controlnet_aux\\': \\'pip install controlnet_aux\\'}, \\'example_code\\': \"from PIL import Image\\\\nfrom diffusers import StableDiffusionControlNetPipeline, ControlNetModel, UniPCMultistepScheduler\\\\nimport torch\\\\nfrom controlnet_aux import OpenposeDetector\\\\nfrom diffusers.utils import load_image\\\\nopenpose = OpenposeDetector.from_pretrained(\\'lllyasviel/ControlNet\\')\\\\nimage = load_image(https://huggingface.co/lllyasviel/sd-controlnet-openpose/resolve/main/images/pose.png)\\\\nimage = openpose(image)\\\\ncontrolnet = ControlNetModel.from_pretrained(\\\\n lllyasviel/sd-controlnet-openpose, torch_dtype=torch.float16\\\\n)\\\\npipe = StableDiffusionControlNetPipeline.from_pretrained(\\\\n runwayml/stable-diffusion-v1-5, controlnet=controlnet, safety_checker=None, torch_dtype=torch.float16\\\\n)\\\\npipe.scheduler = UniPCMultistepScheduler.from_config(pipe.scheduler.config)\\\\npipe.enable_xformers_memory_efficient_attention()\\\\npipe.enable_model_cpu_offload()\\\\nimage = pipe(chef in the kitchen, image, num_inference_steps=20).images[0]\\\\nimage.save(\\'images/chef_pose_out.png\\')\", \\'performance\\': {\\'dataset\\': \\'200k pose-image, caption pairs\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'ControlNet is a neural network structure to control diffusion models by adding extra conditions. This checkpoint corresponds to the ControlNet conditioned on Human Pose Estimation. It can be used in combination with Stable Diffusion.\\'}', metadata={})]", "category": "generic"}
{"question_id": 753, "text": " Create a program to determine the depth map from an input image of a street filled with people.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Depth Estimation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Monocular Depth Estimation\\', \\'api_name\\': \\'Intel/dpt-large\\', \\'api_call\\': \"DPTForDepthEstimation.from_pretrained(\\'Intel/dpt-large\\')\", \\'api_arguments\\': {\\'pretrained_model_name_or_path\\': \\'Intel/dpt-large\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'from transformers import DPTImageProcessor, DPTForDepthEstimation\\\\nimport torch\\\\nimport numpy as np\\\\nfrom PIL import Image\\\\nimport requests\\\\nurl = http://images.cocodataset.org/val2017/000000039769.jpg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nprocessor = DPTImageProcessor.from_pretrained(Intel/dpt-large)\\\\nmodel = DPTForDepthEstimation.from_pretrained(Intel/dpt-large)\\\\ninputs = processor(images=image, return_tensors=pt)\\\\nwith torch.no_grad():\\\\n outputs = model(**inputs)\\\\n predicted_depth = outputs.predicted_depth\\\\nprediction = torch.nn.functional.interpolate(\\\\n predicted_depth.unsqueeze(1),\\\\n size=image.size[::-1],\\\\n mode=bicubic,\\\\n align_corners=False,\\\\n)\\\\noutput = prediction.squeeze().cpu().numpy()\\\\nformatted = (output * 255 / np.max(output)).astype(uint8)\\\\ndepth = Image.fromarray(formatted)\\', \\'performance\\': {\\'dataset\\': \\'MIX 6\\', \\'accuracy\\': \\'10.82\\'}, \\'description\\': \\'Dense Prediction Transformer (DPT) model trained on 1.4 million images for monocular depth estimation. Introduced in the paper Vision Transformers for Dense Prediction by Ranftl et al. (2021). DPT uses the Vision Transformer (ViT) as backbone and adds a neck + head on top for monocular depth estimation.\\'}', metadata={})]", "category": "generic"}
{"question_id": 754, "text": " I need to create a new piece of art to add to my digital gallery that resembles WikiArt images.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Unconditional Image Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Unconditional Image Generation\\', \\'api_name\\': \\'johnowhitaker/sd-class-wikiart-from-bedrooms\\', \\'api_call\\': \"DDPMPipeline.from_pretrained(\\'johnowhitaker/sd-class-wikiart-from-bedrooms\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'diffusers\\', \\'example_code\\': \"from diffusers import DDPMPipeline\\\\npipeline = DDPMPipeline.from_pretrained(\\'johnowhitaker/sd-class-wikiart-from-bedrooms\\')\\\\nimage = pipeline().images[0]\\\\nimage\", \\'performance\\': {\\'dataset\\': \\'https://huggingface.co/datasets/huggan/wikiart\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This model is a diffusion model initialized from https://huggingface.co/google/ddpm-bedroom-256 and trained for 5000 steps on https://huggingface.co/datasets/huggan/wikiart.\\'}', metadata={})]", "category": "generic"}
{"question_id": 755, "text": " Our team is designing a butterfly-themed stationery set for children. We need to generate images of cute butterflies to use in the design.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Unconditional Image Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Unconditional Image Generation\\', \\'api_name\\': \\'ntrant7/sd-class-butterflies-32\\', \\'api_call\\': \"DDPMPipeline.from_pretrained(\\'ntrant7/sd-class-butterflies-32\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'diffusers\\'], \\'example_code\\': \"from diffusers import DDPMPipeline\\\\npipeline = DDPMPipeline.from_pretrained(\\'ntrant7/sd-class-butterflies-32\\')\\\\nimage = pipeline().images[0]\\\\nimage\", \\'performance\\': {\\'dataset\\': \\'Not specified\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'This model is a diffusion model for unconditional image generation of cute butterflies.\\'}', metadata={})]", "category": "generic"}
{"question_id": 756, "text": " Design a mobile application that suggests pictures of butterflies whenever the user attempts to change their theme.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Unconditional Image Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Unconditional Image Generation\\', \\'api_name\\': \\'ntrant7/sd-class-butterflies-32\\', \\'api_call\\': \"DDPMPipeline.from_pretrained(\\'ntrant7/sd-class-butterflies-32\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'diffusers\\'], \\'example_code\\': \"from diffusers import DDPMPipeline\\\\npipeline = DDPMPipeline.from_pretrained(\\'ntrant7/sd-class-butterflies-32\\')\\\\nimage = pipeline().images[0]\\\\nimage\", \\'performance\\': {\\'dataset\\': \\'Not specified\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'This model is a diffusion model for unconditional image generation of cute butterflies.\\'}', metadata={})]", "category": "generic"}
{"question_id": 757, "text": " We need to generate some vintage images for a promotion poster.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'vintedois-diffusion-v0-1\\', \\'api_call\\': \"pipeline(\\'text-to-image\\', model=\\'22h/vintedois-diffusion-v0-1\\')\", \\'api_arguments\\': [\\'prompt\\', \\'CFG Scale\\', \\'Scheduler\\', \\'Steps\\', \\'Seed\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"text2img(\\'photo of an old man in a jungle, looking at the camera\\', CFG Scale=7.5, Scheduler=\\'diffusers.EulerAncestralDiscreteScheduler\\', Steps=30, Seed=44)\", \\'performance\\': {\\'dataset\\': \\'large amount of high quality images\\', \\'accuracy\\': \\'not specified\\'}, \\'description\\': \\'Vintedois (22h) Diffusion model trained by Predogl and piEsposito with open weights, configs and prompts. This model generates beautiful images without a lot of prompt engineering. It can also generate high fidelity faces with a little amount of steps.\\'}', metadata={})]", "category": "generic"}
{"question_id": 758, "text": " Generate images of cute butterflies using myunus1/diffmodels_galaxies_scratchbook model.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Unconditional Image Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Diffusers\\', \\'api_name\\': \\'myunus1/diffmodels_galaxies_scratchbook\\', \\'api_call\\': \"DDPMPipeline.from_pretrained(\\'myunus1/diffmodels_galaxies_scratchbook\\')\", \\'api_arguments\\': {\\'from_pretrained\\': \\'myunus1/diffmodels_galaxies_scratchbook\\'}, \\'python_environment_requirements\\': {\\'package\\': \\'diffusers\\', \\'import\\': \\'from diffusers import DDPMPipeline\\'}, \\'example_code\\': {\\'initialize_pipeline\\': \"pipeline = DDPMPipeline.from_pretrained(\\'myunus1/diffmodels_galaxies_scratchbook\\')\", \\'generate_image\\': \\'image = pipeline().images[0]\\', \\'display_image\\': \\'image\\'}, \\'performance\\': {\\'dataset\\': \\'Not provided\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This model is a diffusion model for unconditional image generation of cute \ud83e\udd8b.\\'}', metadata={})]", "category": "generic"}
{"question_id": 759, "text": " As part of the online platform's development, we need an AI that, given input text about a video, estimates the content of the video and judges its category.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'ImRma/Brucelee\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'ImRma/Brucelee\\')\", \\'api_arguments\\': [\\'your_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Hugging Face model for converting Persian and English text into video.\\'}', metadata={})]", "category": "generic"}
{"question_id": 760, "text": " We need to develop a model to classify sports clips by identifying the type of sports being played in the video.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Video Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Feature Extraction\\', \\'api_name\\': \\'microsoft/xclip-base-patch16-zero-shot\\', \\'api_call\\': \"XClipModel.from_pretrained(\\'microsoft/xclip-base-patch16-zero-shot\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'For code examples, we refer to the documentation.\\', \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'HMDB-51\\', \\'accuracy\\': 44.6}, {\\'name\\': \\'UCF-101\\', \\'accuracy\\': 72.0}, {\\'name\\': \\'Kinetics-600\\', \\'accuracy\\': 65.2}]}, \\'description\\': \\'X-CLIP is a minimal extension of CLIP for general video-language understanding. The model is trained in a contrastive way on (video, text) pairs. This allows the model to be used for tasks like zero-shot, few-shot or fully supervised video classification and video-text retrieval.\\'}', metadata={})]", "category": "generic"}
{"question_id": 761, "text": " We are designing an application for professional athletes. They need a tool to categorize their exercises based on videos.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Video Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Video Classification\\', \\'api_name\\': \\'MCG-NJU/videomae-base-short-finetuned-kinetics\\', \\'api_call\\': \"VideoMAEForVideoClassification.from_pretrained(\\'MCG-NJU/videomae-base-short-finetuned-kinetics\\')\", \\'api_arguments\\': [\\'video\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import VideoMAEImageProcessor, VideoMAEForVideoClassification\\\\nimport numpy as np\\\\nimport torch\\\\nvideo = list(np.random.randn(16, 3, 224, 224))\\\\nprocessor = VideoMAEImageProcessor.from_pretrained(\\'MCG-NJU/videomae-base-short-finetuned-kinetics\\')\\\\nmodel = VideoMAEForVideoClassification.from_pretrained(\\'MCG-NJU/videomae-base-short-finetuned-kinetics\\')\\\\ninputs = processor(video, return_tensors=\\'pt\\')\\\\nwith torch.no_grad():\\\\n outputs = model(**inputs)\\\\n logits = outputs.logits\\\\npredicted_class_idx = logits.argmax(-1).item()\\\\nprint(\\'Predicted class:\\', model.config.id2label[predicted_class_idx])\", \\'performance\\': {\\'dataset\\': \\'Kinetics-400\\', \\'accuracy\\': {\\'top-1\\': 79.4, \\'top-5\\': 94.1}}, \\'description\\': \\'VideoMAE model pre-trained for 800 epochs in a self-supervised way and fine-tuned in a supervised way on Kinetics-400. It was introduced in the paper VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training by Tong et al. and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 762, "text": " We are building a video analysis tool that can automatically detect the main action happening in a given video clip.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'chavinlo/TempoFunk\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'chavinlo/TempoFunk\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Text-to-Video model using Hugging Face Transformers library. Model is capable of generating video content based on the input text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 763, "text": " I would like to create an application that identifies animals in Chinese language image captions. Specifically, we want to know if a picture includes a cat or a dog.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Zero-Shot Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Zero-Shot Image Classification\\', \\'api_name\\': \\'OFA-Sys/chinese-clip-vit-large-patch14-336px\\', \\'api_call\\': \"ChineseCLIPModel.from_pretrained(\\'OFA-Sys/chinese-clip-vit-large-patch14-336px\\')\", \\'api_arguments\\': {\\'images\\': \\'image\\', \\'text\\': \\'texts\\', \\'return_tensors\\': \\'pt\\', \\'padding\\': \\'True\\'}, \\'python_environment_requirements\\': [\\'PIL\\', \\'requests\\', \\'transformers\\'], \\'example_code\\': \\'from PIL import Image\\\\nimport requests\\\\nfrom transformers import ChineseCLIPProcessor, ChineseCLIPModel\\\\nmodel = ChineseCLIPModel.from_pretrained(OFA-Sys/chinese-clip-vit-large-patch14-336px)\\\\nprocessor = ChineseCLIPProcessor.from_pretrained(OFA-Sys/chinese-clip-vit-large-patch14-336px)\\\\nurl = https://clip-cn-beijing.oss-cn-beijing.aliyuncs.com/pokemon.jpeg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\ntexts = []\\\\ninputs = processor(images=image, return_tensors=pt)\\\\nimage_features = model.get_image_features(**inputs)\\\\nimage_features = image_features / image_features.norm(p=2, dim=-1, keepdim=True)\\\\ninputs = processor(text=texts, padding=True, return_tensors=pt)\\\\ntext_features = model.get_text_features(**inputs)\\\\ntext_features = text_features / text_features.norm(p=2, dim=-1, keepdim=True)\\\\ninputs = processor(text=texts, images=image, return_tensors=pt, padding=True)\\\\noutputs = model(**inputs)\\\\nlogits_per_image = outputs.logits_per_image\\\\nprobs = logits_per_image.softmax(dim=1)\\', \\'performance\\': {\\'dataset\\': {\\'CIFAR10\\': 96.0, \\'CIFAR100\\': 79.75, \\'DTD\\': 51.2, \\'EuroSAT\\': 52.0, \\'FER\\': 55.1, \\'FGVC\\': 26.2, \\'KITTI\\': 49.9, \\'MNIST\\': 79.4, \\'PC\\': 63.5, \\'VOC\\': 84.9}, \\'accuracy\\': \\'various\\'}, \\'description\\': \\'Chinese CLIP is a simple implementation of CLIP on a large-scale dataset of around 200 million Chinese image-text pairs. It uses ViT-L/14@336px as the image encoder and RoBERTa-wwm-base as the text encoder.\\'}', metadata={})]", "category": "generic"}
{"question_id": 764, "text": " I need a portable system to perform quick sentiment analysis of customer reviews.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'cardiffnlp/twitter-xlm-roberta-base-sentiment\\', \\'api_call\\': \"pipeline(sentiment-analysis, model=\\'cardiffnlp/twitter-xlm-roberta-base-sentiment\\')\", \\'api_arguments\\': [\\'model_path\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nmodel_path = cardiffnlp/twitter-xlm-roberta-base-sentiment\\\\nsentiment_task = pipeline(sentiment-analysis, model=model_path, tokenizer=model_path)\\\\nsentiment_task(T\\'estimo!)\", \\'performance\\': {\\'dataset\\': \\'Twitter\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This is a multilingual XLM-roBERTa-base model trained on ~198M tweets and finetuned for sentiment analysis. The sentiment fine-tuning was done on 8 languages (Ar, En, Fr, De, Hi, It, Sp, Pt) but it can be used for more languages (see paper for details).\\'}', metadata={})]", "category": "generic"}
{"question_id": 765, "text": " A language model has identified a recent book review and provided a summary of its content. Perform sentiment analysis on the summary to evaluate the reviewer's feelings.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'cardiffnlp/twitter-xlm-roberta-base-sentiment\\', \\'api_call\\': \"pipeline(sentiment-analysis, model=\\'cardiffnlp/twitter-xlm-roberta-base-sentiment\\')\", \\'api_arguments\\': [\\'model_path\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nmodel_path = cardiffnlp/twitter-xlm-roberta-base-sentiment\\\\nsentiment_task = pipeline(sentiment-analysis, model=model_path, tokenizer=model_path)\\\\nsentiment_task(T\\'estimo!)\", \\'performance\\': {\\'dataset\\': \\'Twitter\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This is a multilingual XLM-roBERTa-base model trained on ~198M tweets and finetuned for sentiment analysis. The sentiment fine-tuning was done on 8 languages (Ar, En, Fr, De, Hi, It, Sp, Pt) but it can be used for more languages (see paper for details).\\'}', metadata={})]", "category": "generic"}
{"question_id": 766, "text": " We have movie reviews on our website, and we need to display whether the review is positive or negative.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Classification\\', \\'api_name\\': \\'lvwerra/distilbert-imdb\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'lvwerra/distilbert-imdb\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\', \\'pytorch\\'], \\'example_code\\': \"classifier(\\'I love this movie!\\')\", \\'performance\\': {\\'dataset\\': \\'imdb\\', \\'accuracy\\': 0.928}, \\'description\\': \\'This model is a fine-tuned version of distilbert-base-uncased on the imdb dataset. It is used for sentiment analysis on movie reviews and achieves an accuracy of 0.928 on the evaluation set.\\'}', metadata={})]", "category": "generic"}
{"question_id": 767, "text": " I wish to automatically classify a given text's emotion. What Transformers model should I use?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'michellejieli/emotion_text_classifier\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'michellejieli/emotion_text_classifier\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'classifier(I love this!)\\', \\'performance\\': {\\'dataset\\': [\\'Crowdflower (2016)\\', \\'Emotion Dataset, Elvis et al. (2018)\\', \\'GoEmotions, Demszky et al. (2020)\\', \\'ISEAR, Vikash (2018)\\', \\'MELD, Poria et al. (2019)\\', \\'SemEval-2018, EI-reg, Mohammad et al. (2018)\\', \\'Emotion Lines (Friends)\\'], \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'DistilRoBERTa-base is a transformer model that performs sentiment analysis. I fine-tuned the model on transcripts from the Friends show with the goal of classifying emotions from text data, specifically dialogue from Netflix shows or movies. The model predicts 6 Ekman emotions and a neutral class. These emotions include anger, disgust, fear, joy, neutrality, sadness, and surprise.\\'}', metadata={})]", "category": "generic"}
{"question_id": 768, "text": " Develop a tool to analyze restaurant reviews from Yelp for positive or negative sentiments.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'bert-base-multilingual-uncased-sentiment\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'nlptown/bert-base-multilingual-uncased-sentiment\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"result = sentiment_pipeline(\\'I love this product!\\')\", \\'performance\\': {\\'dataset\\': [{\\'language\\': \\'English\\', \\'accuracy\\': {\\'exact\\': \\'67%\\', \\'off-by-1\\': \\'95%\\'}}, {\\'language\\': \\'Dutch\\', \\'accuracy\\': {\\'exact\\': \\'57%\\', \\'off-by-1\\': \\'93%\\'}}, {\\'language\\': \\'German\\', \\'accuracy\\': {\\'exact\\': \\'61%\\', \\'off-by-1\\': \\'94%\\'}}, {\\'language\\': \\'French\\', \\'accuracy\\': {\\'exact\\': \\'59%\\', \\'off-by-1\\': \\'94%\\'}}, {\\'language\\': \\'Italian\\', \\'accuracy\\': {\\'exact\\': \\'59%\\', \\'off-by-1\\': \\'95%\\'}}, {\\'language\\': \\'Spanish\\', \\'accuracy\\': {\\'exact\\': \\'58%\\', \\'off-by-1\\': \\'95%\\'}}]}, \\'description\\': \\'This a bert-base-multilingual-uncased model finetuned for sentiment analysis on product reviews in six languages: English, Dutch, German, French, Spanish and Italian. It predicts the sentiment of the review as a number of stars (between 1 and 5).\\'}', metadata={})]", "category": "generic"}
{"question_id": 769, "text": " We're creating a chatbot that will detect the user's emotion. I want to start by implementing the basic functionality of emotion detection from the user's responses.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Emotion Classification\\', \\'api_name\\': \\'j-hartmann/emotion-english-distilroberta-base\\', \\'api_call\\': \"pipeline(\\'text-classification\\', model=\\'j-hartmann/emotion-english-distilroberta-base\\', return_all_scores=True)\", \\'api_arguments\\': {\\'text\\': \\'string\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'latest\\'}, \\'example_code\\': \\'from transformers import pipeline\\\\nclassifier = pipeline(text-classification, model=j-hartmann/emotion-english-distilroberta-base, return_all_scores=True)\\\\nclassifier(I love this!)\\', \\'performance\\': {\\'dataset\\': \\'Balanced subset from 6 diverse datasets\\', \\'accuracy\\': \\'66%\\'}, \\'description\\': \"This model classifies emotions in English text data. It predicts Ekman\\'s 6 basic emotions, plus a neutral class: anger, disgust, fear, joy, neutral, sadness, and surprise. The model is a fine-tuned checkpoint of DistilRoBERTa-base.\"}', metadata={})]", "category": "generic"}
{"question_id": 770, "text": " I am a journalist. I am writing an article about European start-ups. I need to see all the dates and company names in the text.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'text2text-generation\\', \\'api_name\\': \\'google/pegasus-newsroom\\', \\'api_call\\': \"pipeline(\\'summarization\\', model=\\'google/pegasus-newsroom\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'This model can be loaded on the Inference API on-demand.\\', \\'performance\\': {\\'dataset\\': \\'newsroom\\', \\'accuracy\\': \\'45.98/34.20/42.18\\'}, \\'description\\': \\'PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization by Jingqing Zhang, Yao Zhao, Mohammad Saleh, and Peter J. Liu. The model is trained on both C4 and HugeNews datasets and is designed for summarization tasks.\\'}', metadata={})]", "category": "generic"}
{"question_id": 771, "text": " Create a tool to extract entities from news articles to help journalists research important figures and organizations.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'it5-base-news-summarization\\', \\'api_call\\': \"pipeline(\\'summarization\\', model=\\'it5/it5-base-news-summarization\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'newsum(Dal 31 maggio \u00e8 infine partita la piattaforma ITsART, a pi\u00f9 di un anno da quando \u2013 durante il primo lockdown \u2013 il ministro della Cultura Dario Franceschini ne aveva parlato come di \u00abuna sorta di Netflix della cultura\u00bb, pensata per \u00aboffrire a tutto il mondo la cultura italiana a pagamento\u00bb. \u00c8 presto per dare giudizi definitivi sulla piattaforma, e di certo sar\u00e0 difficile farlo anche pi\u00f9 avanti senza numeri precisi. Al momento, l\u2019unica cosa che si pu\u00f2 fare \u00e8 guardare com\u2019\u00e8 fatto il sito, contare quanti contenuti ci sono (circa 700 \u201ctitoli\u201d, tra film, documentari, spettacoli teatrali e musicali e altri eventi) e provare a dare un giudizio sul loro valore e sulla loro variet\u00e0. Intanto, una cosa notata da pi\u00f9 parti \u00e8 che diversi contenuti di ITsART sono a pagamento sulla piattaforma sebbene altrove, per esempio su RaiPlay, siano invece disponibili gratuitamente.)\\', \\'performance\\': {\\'dataset\\': \\'NewsSum-IT\\', \\'accuracy\\': {\\'Rouge1\\': 0.339, \\'Rouge2\\': 0.16, \\'RougeL\\': 0.263}}, \\'description\\': \\'IT5 Base model fine-tuned on news summarization on the Fanpage and Il Post corpora for Italian Language Understanding and Generation.\\'}', metadata={})]", "category": "generic"}
{"question_id": 772, "text": " A famous writer is working on a novel. He needs your help to predict the punctuation marks needed in his written draft.\\n \n Use this API documentation for reference: [Document(page_content=\"{'domain': 'Natural Language Processing Token Classification', 'framework': 'Transformers', 'functionality': 'punctuation prediction', 'api_name': 'oliverguhr/fullstop-punctuation-multilang-large', 'api_call': 'PunctuationModel()', 'api_arguments': ['text'], 'python_environment_requirements': ['pip install deepmultilingualpunctuation'], 'example_code': 'from deepmultilingualpunctuation import PunctuationModel\\\\nmodel = PunctuationModel()\\\\ntext = My name is Clara and I live in Berkeley California Ist das eine Frage Frau M\u00fcller\\\\nresult = model.restore_punctuation(text)\\\\nprint(result)', 'performance': {'dataset': 'wmt/europarl', 'EN_accuracy': 0.775, 'DE_accuracy': 0.814, 'FR_accuracy': 0.782, 'IT_accuracy': 0.762}, 'description': 'This model predicts the punctuation of English, Italian, French and German texts. It was developed to restore the punctuation of transcribed spoken language and trained on the Europarl Dataset provided by the SEPP-NLG Shared Task. The model restores the following punctuation markers: ., ,, ?, -, :.'}\", metadata={})]", "category": "generic"}
{"question_id": 773, "text": " There's a collection of texts that we want to be able to analyze for their part-of-speech tags to better understand the structure of the sentences within the texts.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'finiteautomata/beto-sentiment-analysis\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'finiteautomata/beto-sentiment-analysis\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'Hugging Face Transformers library\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'TASS 2020 corpus\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Model trained with TASS 2020 corpus (around ~5k tweets) of several dialects of Spanish. Base model is BETO, a BERT model trained in Spanish. Uses POS, NEG, NEU labels.\\'}', metadata={})]", "category": "generic"}
{"question_id": 774, "text": " I am working on a project where I want to make predictions from my data that is stored in structured tables. Find a pre-trained model for table question answering.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Table Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Table Question Answering\\', \\'api_name\\': \\'google/tapas-mini-finetuned-wtq\\', \\'api_call\\': \"AutoModelForTableQuestionAnswering.from_pretrained(\\'google/tapas-mini-finetuned-wtq\\')\", \\'api_arguments\\': \"tokenizer = AutoTokenizer.from_pretrained(\\'google/tapas-mini-finetuned-wtq\\'); model = AutoModelForTableQuestionAnswering.from_pretrained(\\'google/tapas-mini-finetuned-wtq\\')\", \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"from transformers import pipeline\\\\n\\\\nnlp = pipeline(\\'table-question-answering\\', model=\\'google/tapas-mini-finetuned-wtq\\', tokenizer=\\'google/tapas-mini-finetuned-wtq\\')\", \\'performance\\': {\\'dataset\\': \\'wikitablequestions\\', \\'accuracy\\': 0.2854}, \\'description\\': \\'TAPAS mini model fine-tuned on WikiTable Questions (WTQ). It is pretrained on a large corpus of English data from Wikipedia and can be used for answering questions related to a table.\\'}', metadata={})]", "category": "generic"}
{"question_id": 775, "text": " I have a table in CSV format and a query related to it. Could you obtain an answer for my query?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Table Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Table Question Answering\\', \\'api_name\\': \\'google/tapas-small-finetuned-sqa\\', \\'api_call\\': \"pipeline(\\'table-question-answering\\', model=\\'google/tapas-small-finetuned-sqa\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'msr_sqa\\', \\'accuracy\\': 0.6155}, \\'description\\': \\'TAPAS small model fine-tuned on Sequential Question Answering (SQA). It uses relative position embeddings (i.e. resetting the position index at every cell of the table).\\'}', metadata={})]", "category": "generic"}
{"question_id": 776, "text": " Develop a tool that helps me get answers to questions related to a specific text.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'distilbert-base-uncased-distilled-squad\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=\\'distilbert-base-uncased-distilled-squad\\')\", \\'api_arguments\\': [\\'question\\', \\'context\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\nquestion_answerer = pipeline(question-answering, model=\\'distilbert-base-uncased-distilled-squad\\')\\\\ncontext = r\\\\n... Extractive Question Answering is the task of extracting an answer from a text given a question. An example of a\\\\n... question answering dataset is the SQuAD dataset, which is entirely based on that task. If you would like to fine-tune\\\\n... a model on a SQuAD task, you may leverage the examples/pytorch/question-answering/run_squad.py script.\\\\n... \\\\nresult = question_answerer(question=What is a good example of a question answering dataset?, context=context)\\\\nprint(\\\\n... fAnswer: \\'{result[\\'answer\\']}\\', score: {round(result[\\'score\\'], 4)}, start: {result[\\'start\\']}, end: {result[\\'end\\']}\\\\n...)\", \\'performance\\': {\\'dataset\\': \\'SQuAD v1.1\\', \\'accuracy\\': \\'86.9 F1 score\\'}, \\'description\\': \"DistilBERT base uncased distilled SQuAD is a fine-tuned version of DistilBERT-base-uncased for the task of question answering. It has 40% less parameters than bert-base-uncased, runs 60% faster while preserving over 95% of BERT\\'s performances as measured on the GLUE language understanding benchmark.\"}', metadata={})]", "category": "generic"}
{"question_id": 777, "text": " A healthcare professional wants to get quick answers to COVID-19 related questions from the latest research articles.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'deepset/roberta-base-squad2-covid\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=RobertaForQuestionAnswering.from_pretrained(\\'deepset/roberta-base-squad2-covid\\'), tokenizer=RobertaTokenizer.from_pretrained(\\'deepset/roberta-base-squad2-covid\\'))\", \\'api_arguments\\': {\\'model_name\\': \\'deepset/roberta-base-squad2-covid\\', \\'tokenizer\\': \\'deepset/roberta-base-squad2-covid\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': {\\'QA_input\\': {\\'question\\': \\'Why is model conversion important?\\', \\'context\\': \\'The option to convert models between FARM and transformers gives freedom to the user and let people easily switch between frameworks.\\'}, \\'res\\': \\'nlp(QA_input)\\'}, \\'performance\\': {\\'dataset\\': \\'squad_v2\\', \\'accuracy\\': {\\'XVAL_EM\\': 0.17890995260663506, \\'XVAL_f1\\': 0.49925444207319924, \\'XVAL_top_3_recall\\': 0.8021327014218009}}, \\'description\\': \\'This model is a Roberta-based model fine-tuned on SQuAD-style CORD-19 annotations for the task of extractive question answering in the context of COVID-19. It can be used with the Hugging Face Transformers library for question answering tasks.\\'}', metadata={})]", "category": "generic"}
{"question_id": 778, "text": " I have this app for sharing cooking recipes. Users upload photos and ask questions about the showcased recipe. I need to automatically answer their questions based on the recipe image provided.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Image-to-Text\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'promptcap-coco-vqa\\', \\'api_call\\': \"PromptCap(\\'vqascore/promptcap-coco-vqa\\')\", \\'api_arguments\\': {\\'prompt\\': \\'string\\', \\'image\\': \\'string\\'}, \\'python_environment_requirements\\': \\'pip install promptcap\\', \\'example_code\\': [\\'import torch\\', \\'from promptcap import PromptCap\\', \\'model = PromptCap(vqascore/promptcap-coco-vqa)\\', \\'if torch.cuda.is_available():\\', \\'  model.cuda()\\', \\'prompt = please describe this image according to the given question: what piece of clothing is this boy putting on?\\', \\'image = glove_boy.jpeg\\', \\'print(model.caption(prompt, image))\\'], \\'performance\\': {\\'dataset\\': {\\'coco\\': {\\'accuracy\\': \\'150 CIDEr\\'}, \\'OK-VQA\\': {\\'accuracy\\': \\'60.4%\\'}, \\'A-OKVQA\\': {\\'accuracy\\': \\'59.6%\\'}}}, \\'description\\': \\'PromptCap is a captioning model that can be controlled by natural language instruction. The instruction may contain a question that the user is interested in. It achieves SOTA performance on COCO captioning (150 CIDEr) and knowledge-based VQA tasks when paired with GPT-3 (60.4% on OK-VQA and 59.6% on A-OKVQA).\\'}', metadata={})]", "category": "generic"}
{"question_id": 779, "text": " The company's legal team is working on a case. They need a highly accurate tool to extract answers from a large set of legal documents. Develop a tool for this purpose.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 780, "text": " We have received a customer inquiry. Help us categorize the inquiry into one of the following categories: \\\"sales\\\", \\\"technical support\\\", or \\\"billing\\\".\\n###Input: \\\"I am experiencing difficulty with the installation process of your software.\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Question Answering\\', \\'api_name\\': \\'impira/layoutlm-invoices\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=\\'impira/layoutlm-invoices\\')\", \\'api_arguments\\': \\'question, context\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"qa_pipeline(question=\\'your question\\', context=\\'your document context\\')\", \\'performance\\': {\\'dataset\\': \\'proprietary dataset of invoices, SQuAD2.0, and DocVQA\\', \\'accuracy\\': \\'not provided\\'}, \\'description\\': \\'This is a fine-tuned version of the multi-modal LayoutLM model for the task of question answering on invoices and other documents. It has been fine-tuned on a proprietary dataset of invoices as well as both SQuAD2.0 and DocVQA for general comprehension. Unlike other QA models, which can only extract consecutive tokens (because they predict the start and end of a sequence), this model can predict longer-range, non-consecutive sequences with an additional classifier head.\\'}', metadata={})]", "category": "generic"}
{"question_id": 781, "text": " Imagine you have a news aggregator platform, and you want it to separate news articles into categories like Politics, Sports, Technology, Business, and Entertainment. Implement the classification using this API.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'michellejieli/NSFW_text_classifier\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'michellejieli/NSFW_text_classification\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'classifier(I see you\u2019ve set aside this special time to humiliate yourself in public.)\\', \\'performance\\': {\\'dataset\\': \\'Reddit posts\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'DistilBERT is a transformer model that performs sentiment analysis. I fine-tuned the model on Reddit posts with the purpose of classifying not safe for work (NSFW) content, specifically text that is considered inappropriate and unprofessional. The model predicts 2 classes, which are NSFW or safe for work (SFW). The model is a fine-tuned version of DistilBERT. It was fine-tuned on 14317 Reddit posts pulled from the Reddit API.\\'}', metadata={})]", "category": "generic"}
{"question_id": 782, "text": " I need to classify German news articles into categories like crime, tragedy, and theft. The classifier should be able to understand German as well.\\n \n Use this API documentation for reference: [Document(page_content=\"{'domain': 'Natural Language Processing Text Classification', 'framework': 'Hugging Face Transformers', 'functionality': 'German Sentiment Classification', 'api_name': 'oliverguhr/german-sentiment-bert', 'api_call': 'SentimentModel()', 'api_arguments': ['texts'], 'python_environment_requirements': 'pip install germansentiment', 'example_code': ['from germansentiment import SentimentModel', 'model = SentimentModel()', 'texts = [', ' Mit keinem guten Ergebniss,Das ist gar nicht mal so gut,', ' Total awesome!,nicht so schlecht wie erwartet,', ' Der Test verlief positiv.,Sie fhrt ein grnes Auto.]', 'result = model.predict_sentiment(texts)', 'print(result)'], 'performance': {'dataset': ['holidaycheck', 'scare', 'filmstarts', 'germeval', 'PotTS', 'emotions', 'sb10k', 'Leipzig Wikipedia Corpus 2016', 'all'], 'accuracy': [0.9568, 0.9418, 0.9021, 0.7536, 0.678, 0.9649, 0.7376, 0.9967, 0.9639]}, 'description': 'This model was trained for sentiment classification of German language texts. The model uses the Googles Bert architecture and was trained on 1.834 million German-language samples. The training data contains texts from various domains like Twitter, Facebook and movie, app and hotel reviews.'}\", metadata={})]", "category": "generic"}
{"question_id": 783, "text": " The company has just received a document written in French, and they need it translated into English.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Translation\\', \\'api_name\\': \\'opus-mt-fr-en\\', \\'api_call\\': \"pipeline(\\'translation_fr_to_en\\', model=\\'Helsinki-NLP/opus-mt-fr-en\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\', \\'torch\\'], \\'example_code\\': \"translation_pipeline(\\'Bonjour, comment \u00e7a va?\\')\", \\'performance\\': {\\'dataset\\': \\'opus\\', \\'accuracy\\': {\\'BLEU\\': {\\'newsdiscussdev2015-enfr.fr.en\\': 33.1, \\'newsdiscusstest2015-enfr.fr.en\\': 38.7, \\'newssyscomb2009.fr.en\\': 30.3, \\'news-test2008.fr.en\\': 26.2, \\'newstest2009.fr.en\\': 30.2, \\'newstest2010.fr.en\\': 32.2, \\'newstest2011.fr.en\\': 33.0, \\'newstest2012.fr.en\\': 32.8, \\'newstest2013.fr.en\\': 33.9, \\'newstest2014-fren.fr.en\\': 37.8, \\'Tatoeba.fr.en\\': 57.5}}}, \\'description\\': \\'Helsinki-NLP/opus-mt-fr-en is a machine translation model trained to translate from French to English. It is based on the Marian NMT framework and trained on the OPUS dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 784, "text": " A business collaboration project requires staff to translate Russian documents into English.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Translation\\', \\'api_name\\': \\'Helsinki-NLP/opus-mt-es-en\\', \\'api_call\\': \"pipeline(\\'translation_es_to_en\\', model=\\'Helsinki-NLP/opus-mt-es-en\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\ntranslation = pipeline(\\'translation_es_to_en\\', model=\\'Helsinki-NLP/opus-mt-es-en\\')(\\'Hola, \u00bfc\u00f3mo est\u00e1s?\\')\", \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'newssyscomb2009-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 30.6, \\'chr-F\\': 0.57}}, {\\'name\\': \\'news-test2008-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 27.9, \\'chr-F\\': 0.553}}, {\\'name\\': \\'newstest2009-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 30.4, \\'chr-F\\': 0.572}}, {\\'name\\': \\'newstest2010-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 36.1, \\'chr-F\\': 0.614}}, {\\'name\\': \\'newstest2011-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 34.2, \\'chr-F\\': 0.599}}, {\\'name\\': \\'newstest2012-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 37.9, \\'chr-F\\': 0.624}}, {\\'name\\': \\'newstest2013-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 35.3, \\'chr-F\\': 0.609}}, {\\'name\\': \\'Tatoeba-test.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 59.6, \\'chr-F\\': 0.739}}]}, \\'description\\': \\'Helsinki-NLP/opus-mt-es-en is a machine translation model trained to translate from Spanish to English using the Hugging Face Transformers library. The model is based on the Marian framework and was trained on the OPUS dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 785, "text": " In this fast world, a user spends very little time on reading news articles and requires summary of the news articles.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'it5-base-news-summarization\\', \\'api_call\\': \"pipeline(\\'summarization\\', model=\\'it5/it5-base-news-summarization\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'newsum(Dal 31 maggio \u00e8 infine partita la piattaforma ITsART, a pi\u00f9 di un anno da quando \u2013 durante il primo lockdown \u2013 il ministro della Cultura Dario Franceschini ne aveva parlato come di \u00abuna sorta di Netflix della cultura\u00bb, pensata per \u00aboffrire a tutto il mondo la cultura italiana a pagamento\u00bb. \u00c8 presto per dare giudizi definitivi sulla piattaforma, e di certo sar\u00e0 difficile farlo anche pi\u00f9 avanti senza numeri precisi. Al momento, l\u2019unica cosa che si pu\u00f2 fare \u00e8 guardare com\u2019\u00e8 fatto il sito, contare quanti contenuti ci sono (circa 700 \u201ctitoli\u201d, tra film, documentari, spettacoli teatrali e musicali e altri eventi) e provare a dare un giudizio sul loro valore e sulla loro variet\u00e0. Intanto, una cosa notata da pi\u00f9 parti \u00e8 che diversi contenuti di ITsART sono a pagamento sulla piattaforma sebbene altrove, per esempio su RaiPlay, siano invece disponibili gratuitamente.)\\', \\'performance\\': {\\'dataset\\': \\'NewsSum-IT\\', \\'accuracy\\': {\\'Rouge1\\': 0.339, \\'Rouge2\\': 0.16, \\'RougeL\\': 0.263}}, \\'description\\': \\'IT5 Base model fine-tuned on news summarization on the Fanpage and Il Post corpora for Italian Language Understanding and Generation.\\'}', metadata={})]", "category": "generic"}
{"question_id": 786, "text": " We need to summarize a scientific article. The input must include all the important points discussed in the article, and the result should be a concise abstraction of the content.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Summarization\\', \\'api_name\\': \\'distilbart-cnn-12-6-samsum\\', \\'api_call\\': \"pipeline(\\'summarization\\', model=\\'philschmid/distilbart-cnn-12-6-samsum\\')\", \\'api_arguments\\': {\\'model\\': \\'philschmid/distilbart-cnn-12-6-samsum\\'}, \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"from transformers import pipeline\\\\nsummarizer = pipeline(summarization, model=philschmid/distilbart-cnn-12-6-samsum)\\\\nconversation = \\'\\'\\'Jeff: Can I train a \ud83e\udd17 Transformers model on Amazon SageMaker? \\\\nPhilipp: Sure you can use the new Hugging Face Deep Learning Container. \\\\nJeff: ok.\\\\nJeff: and how can I get started? \\\\nJeff: where can I find documentation? \\\\nPhilipp: ok, ok you can find everything here. https://huggingface.co/blog/the-partnership-amazon-sagemaker-and-hugging-face <br />\\\\n\\'\\'\\'\\\\nsummarizer(conversation)\", \\'performance\\': {\\'dataset\\': \\'samsum\\', \\'accuracy\\': {\\'ROUGE-1\\': 41.09, \\'ROUGE-2\\': 20.746, \\'ROUGE-L\\': 31.595, \\'ROUGE-LSUM\\': 38.339}}, \\'description\\': \\'This model is a DistilBART-based text summarization model trained on the SAMsum dataset. It can be used to generate summaries of conversational text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 787, "text": " Now I need to create a summary of my chat with my friend last night.\\n###Input: conversation = '''Hannah: Hey, do you have Betty's number?\\nAmanda: Lemme check\\nAmanda: Sorry, can't find it.\\nAmanda: Ask Larry\\nAmanda: He called her last time we were at the park together\\nHannah: I don't know him well\\nAmanda: Don't be shy, he's very nice\\nHannah: If you say so..\\nHannah: I'd rather you texted him\\nAmanda: Just text him \\ud83d\\ude42\\nHannah: Urgh.. Alright\\nHannah: Bye\\nAmanda: Bye bye\\n'''\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'lidiya/bart-large-xsum-samsum\\', \\'api_call\\': \"pipeline(\\'summarization\\', model=\\'lidiya/bart-large-xsum-samsum\\')\", \\'api_arguments\\': \\'conversation\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"from transformers import pipeline\\\\nsummarizer = pipeline(summarization, model=lidiya/bart-large-xsum-samsum)\\\\nconversation = \\'\\'\\'Hannah: Hey, do you have Betty\\'s number?\\\\nAmanda: Lemme check\\\\nAmanda: Sorry, can\\'t find it.\\\\nAmanda: Ask Larry\\\\nAmanda: He called her last time we were at the park together\\\\nHannah: I don\\'t know him well\\\\nAmanda: Don\\'t be shy, he\\'s very nice\\\\nHannah: If you say so..\\\\nHannah: I\\'d rather you texted him\\\\nAmanda: Just text him \ud83d\ude42\\\\nHannah: Urgh.. Alright\\\\nHannah: Bye\\\\nAmanda: Bye bye <br />\\\\n\\'\\'\\'\\\\nsummarizer(conversation)\", \\'performance\\': {\\'dataset\\': \\'SAMSum Corpus: A Human-annotated Dialogue Dataset for Abstractive Summarization\\', \\'accuracy\\': {\\'rouge1\\': 53.306, \\'rouge2\\': 28.355, \\'rougeL\\': 44.095}}, \\'description\\': \\'This model was obtained by fine-tuning facebook/bart-large-xsum on Samsum dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 788, "text": " As a news agency, we need a summarized version of a recent article about YouTube's new policy on vaccine misinformation.\\n###Input: Videos that say approved vaccines are dangerous and cause autism, cancer or infertility are among those that will be taken down, the company said. The policy includes the termination of accounts of anti-vaccine influencers. Tech giants have been criticised for not doing more to counter false health information on their sites. In July, US President Joe Biden said social media platforms were largely responsible for people's scepticism in getting vaccinated by spreading misinformation, and appealed for them to address the issue. YouTube, which is owned by Google, said 130,000 videos were removed from its platform since last year, when it implemented a ban on content spreading misinformation about Covid vaccines. In a blog post, the company said it had seen false claims about Covid jabs spill over into misinformation about vaccines in general. The new policy covers long-approved vaccines, such as those against measles or hepatitis B. We're expanding our medical misinformation policies on YouTube with new guidelines on currently administered vaccines that are approved and confirmed to be safe and effective by local health authorities and the WHO, the post said, referring to the World Health Organization.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'text2text-generation\\', \\'api_name\\': \\'csebuetnlp/mT5_multilingual_XLSum\\', \\'api_call\\': \"AutoModelForSeq2SeqLM.from_pretrained(\\'csebuetnlp/mT5_multilingual_XLSum\\')\", \\'api_arguments\\': [\\'model_name\\'], \\'python_environment_requirements\\': [\\'transformers==4.11.0.dev0\\'], \\'example_code\\': \"import re\\\\nfrom transformers import AutoTokenizer, AutoModelForSeq2SeqLM\\\\nWHITESPACE_HANDLER = lambda k: re.sub(\\'\\\\\\\\s+\\', \\' \\', re.sub(\\'\\\\\\\\n+\\', \\' \\', k.strip()))\\\\narticle_text = Videos that say approved vaccines are dangerous and cause autism, cancer or infertility are among those that will be taken down, the company said. The policy includes the termination of accounts of anti-vaccine influencers. Tech giants have been criticised for not doing more to counter false health information on their sites. In July, US President Joe Biden said social media platforms were largely responsible for people\\'s scepticism in getting vaccinated by spreading misinformation, and appealed for them to address the issue. YouTube, which is owned by Google, said 130,000 videos were removed from its platform since last year, when it implemented a ban on content spreading misinformation about Covid vaccines. In a blog post, the company said it had seen false claims about Covid jabs spill over into misinformation about vaccines in general. The new policy covers long-approved vaccines, such as those against measles or hepatitis B. We\\'re expanding our medical misinformation policies on YouTube with new guidelines on currently administered vaccines that are approved and confirmed to be safe and effective by local health authorities and the WHO, the post said, referring to the World Health Organization.\\\\nmodel_name = csebuetnlp/mT5_multilingual_XLSum\\\\ntokenizer = AutoTokenizer.from_pretrained(model_name)\\\\nmodel = AutoModelForSeq2SeqLM.from_pretrained(model_name)\\\\ninput_ids = tokenizer(\\\\n [WHITESPACE_HANDLER(article_text)],\\\\n return_tensors=pt,\\\\n padding=max_length,\\\\n truncation=True,\\\\n max_length=512\\\\n)[input_ids]\\\\noutput_ids = model.generate(\\\\n input_ids=input_ids,\\\\n max_length=84,\\\\n no_repeat_ngram_size=2,\\\\n num_beams=4\\\\n)[0]\\\\nsummary = tokenizer.decode(\\\\n output_ids,\\\\n skip_special_tokens=True,\\\\n clean_up_tokenization_spaces=False\\\\n)\\\\nprint(summary)\", \\'performance\\': {\\'dataset\\': \\'xsum\\', \\'accuracy\\': {\\'ROUGE-1\\': 36.5, \\'ROUGE-2\\': 13.934, \\'ROUGE-L\\': 28.988, \\'ROUGE-LSUM\\': 28.996, \\'loss\\': 2.067, \\'gen_len\\': 26.973}}, \\'description\\': \\'This repository contains the mT5 checkpoint finetuned on the 45 languages of XL-Sum dataset. It is a multilingual abstractive summarization model that supports text-to-text generation for 43 languages.\\'}', metadata={})]", "category": "generic"}
{"question_id": 789, "text": " Let's build a Chatbot that can solve daily life problems of a research Assistant.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Generation\\', \\'api_name\\': \\'mywateriswet/ShuanBot\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'mywateriswet/ShuanBot\\')\", \\'api_arguments\\': \\'message\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"response = chatbot(\\'What is your name?\\')\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'ShuanBot is a conversational chatbot model based on the GPT-2 architecture. It can be used for generating human-like responses in a chat context.\\'}', metadata={})]", "category": "generic"}
{"question_id": 790, "text": " We need to integrate the personal assistant we're building with the capability of having conversations with people, sometimes answering general knowledge questions.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'hyunwoongko/blenderbot-9B\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'hyunwoongko/blenderbot-9B\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'Input a message to start chatting with hyunwoongko/blenderbot-9B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high-performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi-turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 791, "text": " I want to create a chatbot for discussing gardening topics and tips, which can refer to external information while generating responses.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Generation\\', \\'api_name\\': \\'mywateriswet/ShuanBot\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'mywateriswet/ShuanBot\\')\", \\'api_arguments\\': \\'message\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"response = chatbot(\\'What is your name?\\')\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'ShuanBot is a conversational chatbot model based on the GPT-2 architecture. It can be used for generating human-like responses in a chat context.\\'}', metadata={})]", "category": "generic"}
{"question_id": 792, "text": " We're trying to help out a friend who's developing an application for composing text. He is trying to create a bot that comes up with creative ideas for your paragraph.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Paraphrasing\\', \\'api_name\\': \\'prithivida/parrot_paraphraser_on_T5\\', \\'api_call\\': \"Parrot(model_tag=\\'prithivida/parrot_paraphraser_on_T5\\', use_gpu=False)\", \\'api_arguments\\': [\\'input_phrase\\', \\'diversity_ranker\\', \\'do_diverse\\', \\'max_return_phrases\\', \\'max_length\\', \\'adequacy_threshold\\', \\'fluency_threshold\\'], \\'python_environment_requirements\\': [\\'torch\\', \\'transformers\\'], \\'example_code\\': \\'from parrot import Parrot\\\\nimport torch\\\\nimport warnings\\\\nwarnings.filterwarnings(ignore)\\\\n\\\\nparrot = Parrot(model_tag=prithivida/parrot_paraphraser_on_T5, use_gpu=False)\\\\nphrases = [Can you recommed some upscale restaurants in Newyork?,\\\\n What are the famous places we should not miss in Russia?\\\\n]\\\\nfor phrase in phrases:\\\\n print(-*100)\\\\n print(Input_phrase: , phrase)\\\\n print(-*100)\\\\n para_phrases = parrot.augment(input_phrase=phrase)\\\\n for para_phrase in para_phrases:\\\\n  print(para_phrase)\\', \\'performance\\': {\\'dataset\\': \\'Not mentioned\\', \\'accuracy\\': \\'Not mentioned\\'}, \\'description\\': \\'Parrot is a paraphrase based utterance augmentation framework purpose built to accelerate training NLU models. It offers knobs to control Adequacy, Fluency, and Diversity as per your needs. It mainly focuses on augmenting texts typed-into or spoken-to conversational interfaces for building robust NLU models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 793, "text": " We are creating an AI newsletter application that generates summaries of news articles. We need the AI to generate a brief summary for a given article.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'mrm8488/t5-base-finetuned-summarize-news\\', \\'api_call\\': \"AutoModelWithLMHead.from_pretrained(\\'mrm8488/t5-base-finetuned-summarize-news\\')\", \\'api_arguments\\': [\\'text\\', \\'max_length\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'from transformers import AutoModelWithLMHead, AutoTokenizer\\\\ntokenizer = AutoTokenizer.from_pretrained(mrm8488/t5-base-finetuned-summarize-news)\\\\nmodel = AutoModelWithLMHead.from_pretrained(mrm8488/t5-base-finetuned-summarize-news)\\\\ndef summarize(text, max_length=150):\\\\n input_ids = tokenizer.encode(text, return_tensors=pt, add_special_tokens=True)\\\\ngenerated_ids = model.generate(input_ids=input_ids, num_beams=2, max_length=max_length, repetition_penalty=2.5, length_penalty=1.0, early_stopping=True)\\\\npreds = [tokenizer.decode(g, skip_special_tokens=True, clean_up_tokenization_spaces=True) for g in generated_ids]\\\\nreturn preds[0]\\', \\'performance\\': {\\'dataset\\': \\'News Summary\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \"Google\\'s T5 base fine-tuned on News Summary dataset for summarization downstream task. The dataset consists of 4515 examples and contains Author_name, Headlines, Url of Article, Short text, Complete Article. Time period ranges from February to August 2017.\"}', metadata={})]", "category": "generic"}
{"question_id": 794, "text": " I want a code completion tool to assist with finishing my incomplete Python code.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Paraphrasing\\', \\'api_name\\': \\'prithivida/parrot_paraphraser_on_T5\\', \\'api_call\\': \"Parrot(model_tag=\\'prithivida/parrot_paraphraser_on_T5\\', use_gpu=False)\", \\'api_arguments\\': [\\'input_phrase\\', \\'diversity_ranker\\', \\'do_diverse\\', \\'max_return_phrases\\', \\'max_length\\', \\'adequacy_threshold\\', \\'fluency_threshold\\'], \\'python_environment_requirements\\': [\\'torch\\', \\'transformers\\'], \\'example_code\\': \\'from parrot import Parrot\\\\nimport torch\\\\nimport warnings\\\\nwarnings.filterwarnings(ignore)\\\\n\\\\nparrot = Parrot(model_tag=prithivida/parrot_paraphraser_on_T5, use_gpu=False)\\\\nphrases = [Can you recommed some upscale restaurants in Newyork?,\\\\n What are the famous places we should not miss in Russia?\\\\n]\\\\nfor phrase in phrases:\\\\n print(-*100)\\\\n print(Input_phrase: , phrase)\\\\n print(-*100)\\\\n para_phrases = parrot.augment(input_phrase=phrase)\\\\n for para_phrase in para_phrases:\\\\n  print(para_phrase)\\', \\'performance\\': {\\'dataset\\': \\'Not mentioned\\', \\'accuracy\\': \\'Not mentioned\\'}, \\'description\\': \\'Parrot is a paraphrase based utterance augmentation framework purpose built to accelerate training NLU models. It offers knobs to control Adequacy, Fluency, and Diversity as per your needs. It mainly focuses on augmenting texts typed-into or spoken-to conversational interfaces for building robust NLU models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 795, "text": " We are making an AI copywriter for marketing content. Help me to provide content for a product relating to eco-friendly kitchenware.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'hyunwoongko/blenderbot-9B\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'hyunwoongko/blenderbot-9B\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'Input a message to start chatting with hyunwoongko/blenderbot-9B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high-performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi-turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 796, "text": " I am an astronaut in space, writing my diary every day. I need a summary of my diary before sharing it with my family.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Summarization\\', \\'api_name\\': \\'distilbart-cnn-12-6-samsum\\', \\'api_call\\': \"pipeline(\\'summarization\\', model=\\'philschmid/distilbart-cnn-12-6-samsum\\')\", \\'api_arguments\\': {\\'model\\': \\'philschmid/distilbart-cnn-12-6-samsum\\'}, \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"from transformers import pipeline\\\\nsummarizer = pipeline(summarization, model=philschmid/distilbart-cnn-12-6-samsum)\\\\nconversation = \\'\\'\\'Jeff: Can I train a \ud83e\udd17 Transformers model on Amazon SageMaker? \\\\nPhilipp: Sure you can use the new Hugging Face Deep Learning Container. \\\\nJeff: ok.\\\\nJeff: and how can I get started? \\\\nJeff: where can I find documentation? \\\\nPhilipp: ok, ok you can find everything here. https://huggingface.co/blog/the-partnership-amazon-sagemaker-and-hugging-face <br />\\\\n\\'\\'\\'\\\\nsummarizer(conversation)\", \\'performance\\': {\\'dataset\\': \\'samsum\\', \\'accuracy\\': {\\'ROUGE-1\\': 41.09, \\'ROUGE-2\\': 20.746, \\'ROUGE-L\\': 31.595, \\'ROUGE-LSUM\\': 38.339}}, \\'description\\': \\'This model is a DistilBART-based text summarization model trained on the SAMsum dataset. It can be used to generate summaries of conversational text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 797, "text": " I have some paragraphs that I want to use as a source for generating questions, based on the information present in them.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Text2Text Generation\\', \\'api_name\\': \\'t5_sentence_paraphraser\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'ramsrigouthamg/t5_sentence_paraphraser\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A T5 model for paraphrasing sentences\\'}', metadata={})]", "category": "generic"}
{"question_id": 798, "text": " I am a novel writer. I plan to write some paragraphs, but I cannot find the exact word placeholder or missing word. Can you help me make a tool to complete the given text below?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Paraphrasing\\', \\'api_name\\': \\'prithivida/parrot_paraphraser_on_T5\\', \\'api_call\\': \"Parrot(model_tag=\\'prithivida/parrot_paraphraser_on_T5\\', use_gpu=False)\", \\'api_arguments\\': [\\'input_phrase\\', \\'diversity_ranker\\', \\'do_diverse\\', \\'max_return_phrases\\', \\'max_length\\', \\'adequacy_threshold\\', \\'fluency_threshold\\'], \\'python_environment_requirements\\': [\\'torch\\', \\'transformers\\'], \\'example_code\\': \\'from parrot import Parrot\\\\nimport torch\\\\nimport warnings\\\\nwarnings.filterwarnings(ignore)\\\\n\\\\nparrot = Parrot(model_tag=prithivida/parrot_paraphraser_on_T5, use_gpu=False)\\\\nphrases = [Can you recommed some upscale restaurants in Newyork?,\\\\n What are the famous places we should not miss in Russia?\\\\n]\\\\nfor phrase in phrases:\\\\n print(-*100)\\\\n print(Input_phrase: , phrase)\\\\n print(-*100)\\\\n para_phrases = parrot.augment(input_phrase=phrase)\\\\n for para_phrase in para_phrases:\\\\n  print(para_phrase)\\', \\'performance\\': {\\'dataset\\': \\'Not mentioned\\', \\'accuracy\\': \\'Not mentioned\\'}, \\'description\\': \\'Parrot is a paraphrase based utterance augmentation framework purpose built to accelerate training NLU models. It offers knobs to control Adequacy, Fluency, and Diversity as per your needs. It mainly focuses on augmenting texts typed-into or spoken-to conversational interfaces for building robust NLU models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 799, "text": " I am working in a dating chatapp development team. We want to generate sentences to make it more interactive.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'hyunwoongko/blenderbot-9B\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'hyunwoongko/blenderbot-9B\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'Input a message to start chatting with hyunwoongko/blenderbot-9B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high-performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi-turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 801, "text": " I am building a recommendation engine to group news articles. I need a way to determine the similarity between two sentences.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Sentence Similarity\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Sentence Transformers\\', \\'api_name\\': \\'sentence-transformers/distiluse-base-multilingual-cased-v2\\', \\'api_call\\': \"SentenceTransformer(\\'sentence-transformers/distiluse-base-multilingual-cased-v2\\')\", \\'api_arguments\\': [\\'sentences\\'], \\'python_environment_requirements\\': \\'pip install -U sentence-transformers\\', \\'example_code\\': \"from sentence_transformers import SentenceTransformer\\\\nsentences = [This is an example sentence, Each sentence is converted]\\\\nmodel = SentenceTransformer(\\'sentence-transformers/distiluse-base-multilingual-cased-v2\\')\\\\nembeddings = model.encode(sentences)\\\\nprint(embeddings)\", \\'performance\\': {\\'dataset\\': \\'https://seb.sbert.net\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This is a sentence-transformers model: It maps sentences & paragraphs to a 512 dimensional dense vector space and can be used for tasks like clustering or semantic search.\\'}', metadata={})]", "category": "generic"}
{"question_id": 802, "text": " The company needs a tool to analyze customers' reviews about their products. We need to find out which ones are positive, neutral, or negative.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'bert-base-multilingual-uncased-sentiment\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'nlptown/bert-base-multilingual-uncased-sentiment\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"result = sentiment_pipeline(\\'I love this product!\\')\", \\'performance\\': {\\'dataset\\': [{\\'language\\': \\'English\\', \\'accuracy\\': {\\'exact\\': \\'67%\\', \\'off-by-1\\': \\'95%\\'}}, {\\'language\\': \\'Dutch\\', \\'accuracy\\': {\\'exact\\': \\'57%\\', \\'off-by-1\\': \\'93%\\'}}, {\\'language\\': \\'German\\', \\'accuracy\\': {\\'exact\\': \\'61%\\', \\'off-by-1\\': \\'94%\\'}}, {\\'language\\': \\'French\\', \\'accuracy\\': {\\'exact\\': \\'59%\\', \\'off-by-1\\': \\'94%\\'}}, {\\'language\\': \\'Italian\\', \\'accuracy\\': {\\'exact\\': \\'59%\\', \\'off-by-1\\': \\'95%\\'}}, {\\'language\\': \\'Spanish\\', \\'accuracy\\': {\\'exact\\': \\'58%\\', \\'off-by-1\\': \\'95%\\'}}]}, \\'description\\': \\'This a bert-base-multilingual-uncased model finetuned for sentiment analysis on product reviews in six languages: English, Dutch, German, French, Spanish and Italian. It predicts the sentiment of the review as a number of stars (between 1 and 5).\\'}', metadata={})]", "category": "generic"}
{"question_id": 803, "text": " I am working on a project that ranks answers to questions based on their relevance. Can you help me find the most relevant answer to a specific question by using sentence similarity?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Sentence Similarity\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Sentence Transformers\\', \\'api_name\\': \\'sentence-transformers/multi-qa-mpnet-base-dot-v1\\', \\'api_call\\': \"SentenceTransformer(\\'sentence-transformers/multi-qa-mpnet-base-dot-v1\\')\", \\'api_arguments\\': [\\'query\\', \\'docs\\'], \\'python_environment_requirements\\': \\'pip install -U sentence-transformers\\', \\'example_code\\': \"from sentence_transformers import SentenceTransformer, util\\\\nquery = How many people live in London?\\\\ndocs = [Around 9 Million people live in London, London is known for its financial district]\\\\nmodel = SentenceTransformer(\\'sentence-transformers/multi-qa-mpnet-base-dot-v1\\')\\\\nquery_emb = model.encode(query)\\\\ndoc_emb = model.encode(docs)\\\\nscores = util.dot_score(query_emb, doc_emb)[0].cpu().tolist()\\\\ndoc_score_pairs = list(zip(docs, scores))\\\\ndoc_score_pairs = sorted(doc_score_pairs, key=lambda x: x[1], reverse=True)\\\\nfor doc, score in doc_score_pairs:\\\\n print(score, doc)\", \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'WikiAnswers\\', \\'accuracy\\': 77427422}, {\\'name\\': \\'PAQ\\', \\'accuracy\\': 64371441}, {\\'name\\': \\'Stack Exchange\\', \\'accuracy\\': 25316456}]}, \\'description\\': \\'This is a sentence-transformers model that maps sentences & paragraphs to a 768 dimensional dense vector space and was designed for semantic search. It has been trained on 215M (question, answer) pairs from diverse sources.\\'}', metadata={})]", "category": "generic"}
{"question_id": 804, "text": " Assist a client producing an audiobook in Japanese. They need a solution to convert their text script into spoken Japanese.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'kan-bayashi_jvs_tts_finetune_jvs001_jsut_vits_raw_phn_jaconv_pyopenjta-truncated-178804\\', \\'api_call\\': \"AutoModelForCausalLM.from_pretrained(\\'espnet/kan-bayashi_jvs_tts_finetune_jvs001_jsut_vits_raw_phn_jaconv_pyopenjta-truncated-178804\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Japanese text-to-speech model trained using the ESPnet framework. It is designed to convert text input into natural-sounding speech.\\'}', metadata={})]", "category": "generic"}
{"question_id": 805, "text": " We have a product for the German-speaking audience, and we need to generate some sample voiceovers for the promotional videos.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'speechbrain\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'padmalcom/tts-tacotron2-german\\', \\'api_call\\': \"Tacotron2.from_hparams(source=\\'padmalcom/tts-tacotron2-german\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'pip install speechbrain\\'], \\'example_code\\': [\\'import torchaudio\\', \\'from speechbrain.pretrained import Tacotron2\\', \\'from speechbrain.pretrained import HIFIGAN\\', \\'tacotron2 = Tacotron2.from_hparams(source=padmalcom/tts-tacotron2-german, savedir=tmpdir_tts)\\', \\'hifi_gan = HIFIGAN.from_hparams(source=speechbrain/tts-hifigan-ljspeech, savedir=tmpdir_vocoder)\\', \\'mel_output, mel_length, alignment = tacotron2.encode_text(Die Sonne schien den ganzen Tag.)\\', \\'waveforms = hifi_gan.decode_batch(mel_output)\\', \"torchaudio.save(\\'example_TTS.wav\\',waveforms.squeeze(1), 22050)\"], \\'performance\\': {\\'dataset\\': \\'custom german dataset\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Text-to-Speech (TTS) with Tacotron2 trained on a custom german dataset with 12 days voice using speechbrain. Trained for 39 epochs (english speechbrain models are trained for 750 epochs) so there is room for improvement and the model is most likely to be updated soon. The hifigan vocoder can fortunately be used language-independently.\\'}', metadata={})]", "category": "generic"}
{"question_id": 806, "text": " I work as a transcriptionist, and I have an hour-long recording of a meeting. I need to identify and segment the speech of various speakers in the recording.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'pyannote.audio\\', \\'functionality\\': \\'Speaker diarization\\', \\'api_name\\': \\'johnislarry/cloned-pyannote-speaker-diarization-endpoint\\', \\'api_call\\': \"Pipeline.from_pretrained(\\'pyannote/speaker-diarization@2.1\\',use_auth_token=\\'ACCESS_TOKEN_GOES_HERE\\')\", \\'api_arguments\\': [\\'num_speakers\\', \\'min_speakers\\', \\'max_speakers\\', \\'segmentation_onset\\'], \\'python_environment_requirements\\': \\'pyannote.audio 2.0\\', \\'example_code\\': {\\'load_pipeline\\': \\'from pyannote.audio import Pipeline\\\\npipeline = Pipeline.from_pretrained(pyannote/speaker-diarization@2022.07)\\', \\'apply_pipeline\\': \\'diarization = pipeline(audio.wav)\\', \\'save_output\\': \\'with open(audio.rttm, w) as rttm:\\\\n  diarization.write_rttm(rttm)\\'}, \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'AISHELL-4\\', \\'accuracy\\': {\\'DER%\\': 14.61, \\'FA%\\': 3.31, \\'Miss%\\': 4.35, \\'Conf%\\': 6.95}}, {\\'name\\': \\'AMI Mix-Headset only_words\\', \\'accuracy\\': {\\'DER%\\': 18.21, \\'FA%\\': 3.28, \\'Miss%\\': 11.07, \\'Conf%\\': 3.87}}, {\\'name\\': \\'AMI Array1-01 only_words\\', \\'accuracy\\': {\\'DER%\\': 29.0, \\'FA%\\': 2.71, \\'Miss%\\': 21.61, \\'Conf%\\': 4.68}}, {\\'name\\': \\'CALLHOME Part2\\', \\'accuracy\\': {\\'DER%\\': 30.24, \\'FA%\\': 3.71, \\'Miss%\\': 16.86, \\'Conf%\\': 9.66}}, {\\'name\\': \\'DIHARD 3 Full\\', \\'accuracy\\': {\\'DER%\\': 20.99, \\'FA%\\': 4.25, \\'Miss%\\': 10.74, \\'Conf%\\': 6.0}}, {\\'name\\': \\'REPERE Phase 2\\', \\'accuracy\\': {\\'DER%\\': 12.62, \\'FA%\\': 1.55, \\'Miss%\\': 3.3, \\'Conf%\\': 7.76}}, {\\'name\\': \\'VoxConverse v0.0.2\\', \\'accuracy\\': {\\'DER%\\': 12.76, \\'FA%\\': 3.45, \\'Miss%\\': 3.85, \\'Conf%\\': 5.46}}]}, \\'description\\': \\'This API provides speaker diarization functionality using the pyannote.audio framework. It is capable of processing audio files and outputting speaker diarization results in RTTM format. The API supports providing the number of speakers, minimum and maximum number of speakers, and adjusting the segmentation onset threshold.\\'}', metadata={})]", "category": "generic"}
{"question_id": 807, "text": " We are a company focusing on transcription services for Arabic speakers. We need to convert their speech to text.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'Fairseq\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'tts_transformer-ar-cv7\\', \\'api_call\\': \"load_model_ensemble_and_task_from_hf_hub(\\'facebook/tts_transformer-ar-cv7\\')\", \\'api_arguments\\': {\\'text\\': \\'input text\\'}, \\'python_environment_requirements\\': [\\'fairseq\\', \\'IPython\\'], \\'example_code\\': \\'from fairseq.checkpoint_utils import load_model_ensemble_and_task_from_hf_hub\\\\nfrom fairseq.models.text_to_speech.hub_interface import TTSHubInterface\\\\nimport IPython.display as ipd\\\\nmodels, cfg, task = load_model_ensemble_and_task_from_hf_hub(\\\\n facebook/tts_transformer-ar-cv7,\\\\n arg_overrides={vocoder: hifigan, fp16: False}\\\\n)\\\\nmodel = models[0]\\\\nTTSHubInterface.update_cfg_with_data_cfg(cfg, task.data_cfg)\\\\ngenerator = task.build_generator(model, cfg)\\\\ntext = \u0645\u0631\u062d\u0628\u064b\u0627 \u060c \u0647\u0630\u0627 \u0627\u062e\u062a\u0628\u0627\u0631 \u062a\u0634\u063a\u064a\u0644.\\\\nsample = TTSHubInterface.get_model_input(task, text)\\\\nwav, rate = TTSHubInterface.get_prediction(task, model, generator, sample)\\\\nipd.Audio(wav, rate=rate)\\', \\'performance\\': {\\'dataset\\': \\'common_voice\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'Transformer text-to-speech model for Arabic language with a single-speaker male voice, trained on Common Voice v7 dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 808, "text": " An audio file is recorded in a conference and we need the text version of the conversation for record-keeping purposes.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Audio-to-Audio\\', \\'framework\\': \\'Fairseq\\', \\'functionality\\': \\'speech-to-speech-translation\\', \\'api_name\\': \\'facebook/textless_sm_ro_en\\', \\'api_call\\': \"pipeline(\\'audio-to-audio\\', model=\\'facebook/textless_sm_ro_en\\')\", \\'api_arguments\\': \\'audio file or recording\\', \\'python_environment_requirements\\': \\'fairseq, huggingface_hub\\', \\'example_code\\': \\'https://huggingface.co/facebook/textless_sm_cs_en\\', \\'performance\\': {\\'dataset\\': \\'unknown\\', \\'accuracy\\': \\'unknown\\'}, \\'description\\': \\'A speech-to-speech translation model for Romanian to English developed by Facebook AI\\'}', metadata={})]", "category": "generic"}
{"question_id": 809, "text": " Provide a way to translate spoken English to spoken Hokkien for an audio file.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'Fairseq\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'unit_hifigan_HK_layer12.km2500_frame_TAT-TTS\\', \\'api_call\\': \"load_model_ensemble_and_task_from_hf_hub(\\'facebook/unit_hifigan_HK_layer12.km2500_frame_TAT-TT\\')\", \\'api_arguments\\': {\\'unit\\': \\'Text input for the TTS model\\'}, \\'python_environment_requirements\\': [\\'fairseq\\', \\'huggingface_hub\\', \\'torchaudio\\'], \\'example_code\\': \"import json\\\\nimport os\\\\nfrom pathlib import Path\\\\nimport IPython.display as ipd\\\\nfrom fairseq import hub_utils\\\\nfrom fairseq.checkpoint_utils import load_model_ensemble_and_task_from_hf_hub\\\\nfrom fairseq.models.speech_to_text.hub_interface import S2THubInterface\\\\nfrom fairseq.models.text_to_speech import CodeHiFiGANVocoder\\\\nfrom fairseq.models.text_to_speech.hub_interface import VocoderHubInterface\\\\nfrom huggingface_hub import snapshot_download\\\\nimport torchaudio\\\\ncache_dir = os.getenv(HUGGINGFACE_HUB_CACHE)\\\\nlibrary_name = fairseq\\\\ncache_dir = (\\\\n cache_dir or (Path.home() / .cache / library_name).as_posix()\\\\n)\\\\ncache_dir = snapshot_download(\\\\n ffacebook/unit_hifigan_HK_layer12.km2500_frame_TAT-TTS, cache_dir=cache_dir, library_name=library_name\\\\n)\\\\nx = hub_utils.from_pretrained(\\\\n cache_dir,\\\\n model.pt,\\\\n .,\\\\n archive_map=CodeHiFiGANVocoder.hub_models(),\\\\n config_yaml=config.json,\\\\n fp16=False,\\\\n is_vocoder=True,\\\\n)\\\\nwith open(f{x[\\'args\\'][\\'data\\']}/config.json) as f:\\\\n vocoder_cfg = json.load(f)\\\\nassert (\\\\n len(x[args][model_path]) == 1\\\\n), Too many vocoder models in the input\\\\nvocoder = CodeHiFiGANVocoder(x[args][model_path][0], vocoder_cfg)\\\\ntts_model = VocoderHubInterface(vocoder_cfg, vocoder)\\\\ntts_sample = tts_model.get_model_input(unit)\\\\nwav, sr = tts_model.get_prediction(tts_sample)\\\\nipd.Audio(wav, rate=sr)\", \\'performance\\': {\\'dataset\\': \\'TAT-TTS\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Hokkien unit HiFiGAN based vocoder from fairseq. Trained with TAT-TTS data with 4 speakers in Taiwanese Hokkien accent.\\'}', metadata={})]", "category": "generic"}
{"question_id": 810, "text": " We want to separate the speaker sources from the original audio file to filter the noise.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Audio-to-Audio\\', \\'framework\\': \\'SpeechBrain\\', \\'functionality\\': \\'Audio Source Separation\\', \\'api_name\\': \\'sepformer-wsj02mix\\', \\'api_call\\': \"separator.from_hparams(source=\\'speechbrain/sepformer-wsj02mix\\')\", \\'api_arguments\\': [\\'path\\'], \\'python_environment_requirements\\': [\\'pip install speechbrain\\'], \\'example_code\\': [\\'from speechbrain.pretrained import SepformerSeparation as separator\\', \\'import torchaudio\\', \"model = separator.from_hparams(source=speechbrain/sepformer-wsj02mix, savedir=\\'pretrained_models/sepformer-wsj02mix\\')\", \"est_sources = model.separate_file(path=\\'speechbrain/sepformer-wsj02mix/test_mixture.wav\\')\", \\'torchaudio.save(source1hat.wav, est_sources[:, :, 0].detach().cpu(), 8000)\\', \\'torchaudio.save(source2hat.wav, est_sources[:, :, 1].detach().cpu(), 8000)\\'], \\'performance\\': {\\'dataset\\': \\'WSJ0-2Mix\\', \\'accuracy\\': \\'22.4 dB\\'}, \\'description\\': \\'This repository provides all the necessary tools to perform audio source separation with a SepFormer model, implemented with SpeechBrain, and pretrained on WSJ0-2Mix dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 811, "text": " Our goal is to analyze the emotions expressed in a user's recorded message.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Audio Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'padmalcom/wav2vec2-large-emotion-detection-german\\', \\'api_call\\': \"pipeline(\\'audio-classification\\', model=\\'padmalcom/wav2vec2-large-emotion-detection-german\\')\", \\'api_arguments\\': \\'audio_file\\', \\'python_environment_requirements\\': \\'transformers, torch\\', \\'example_code\\': \"from transformers import pipeline\\\\naudio_classifier = pipeline(\\'audio-classification\\', model=\\'padmalcom/wav2vec2-large-emotion-detection-german\\')\\\\nresult = audio_classifier(audio_file)\", \\'performance\\': {\\'dataset\\': \\'emo-DB\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This wav2vec2 based emotion detection model is trained on the emo-DB dataset. It can classify emotions in German audio files into seven classes: anger, boredom, disgust, fear, happiness, sadness, and neutral.\\'}', metadata={})]", "category": "generic"}
{"question_id": 812, "text": " I want to make a conference call app which differentiates between the times when the user is speaking and when there is no voice activity. \\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Voice Activity Detection\\', \\'api_name\\': \\'d4data/Indian-voice-cloning\\', \\'api_call\\': \"pipeline(\\'voice-activity-detection\\', model=\\'d4data/Indian-voice-cloning\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A model for detecting voice activity in Indian languages.\\'}', metadata={})]", "category": "generic"}
{"question_id": 813, "text": " We want to analyze a conference call recording to identify the speakers and the segments of the conversation they participated in.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'pyannote.audio\\', \\'functionality\\': \\'Speaker diarization\\', \\'api_name\\': \\'johnislarry/cloned-pyannote-speaker-diarization-endpoint\\', \\'api_call\\': \"Pipeline.from_pretrained(\\'pyannote/speaker-diarization@2.1\\',use_auth_token=\\'ACCESS_TOKEN_GOES_HERE\\')\", \\'api_arguments\\': [\\'num_speakers\\', \\'min_speakers\\', \\'max_speakers\\', \\'segmentation_onset\\'], \\'python_environment_requirements\\': \\'pyannote.audio 2.0\\', \\'example_code\\': {\\'load_pipeline\\': \\'from pyannote.audio import Pipeline\\\\npipeline = Pipeline.from_pretrained(pyannote/speaker-diarization@2022.07)\\', \\'apply_pipeline\\': \\'diarization = pipeline(audio.wav)\\', \\'save_output\\': \\'with open(audio.rttm, w) as rttm:\\\\n  diarization.write_rttm(rttm)\\'}, \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'AISHELL-4\\', \\'accuracy\\': {\\'DER%\\': 14.61, \\'FA%\\': 3.31, \\'Miss%\\': 4.35, \\'Conf%\\': 6.95}}, {\\'name\\': \\'AMI Mix-Headset only_words\\', \\'accuracy\\': {\\'DER%\\': 18.21, \\'FA%\\': 3.28, \\'Miss%\\': 11.07, \\'Conf%\\': 3.87}}, {\\'name\\': \\'AMI Array1-01 only_words\\', \\'accuracy\\': {\\'DER%\\': 29.0, \\'FA%\\': 2.71, \\'Miss%\\': 21.61, \\'Conf%\\': 4.68}}, {\\'name\\': \\'CALLHOME Part2\\', \\'accuracy\\': {\\'DER%\\': 30.24, \\'FA%\\': 3.71, \\'Miss%\\': 16.86, \\'Conf%\\': 9.66}}, {\\'name\\': \\'DIHARD 3 Full\\', \\'accuracy\\': {\\'DER%\\': 20.99, \\'FA%\\': 4.25, \\'Miss%\\': 10.74, \\'Conf%\\': 6.0}}, {\\'name\\': \\'REPERE Phase 2\\', \\'accuracy\\': {\\'DER%\\': 12.62, \\'FA%\\': 1.55, \\'Miss%\\': 3.3, \\'Conf%\\': 7.76}}, {\\'name\\': \\'VoxConverse v0.0.2\\', \\'accuracy\\': {\\'DER%\\': 12.76, \\'FA%\\': 3.45, \\'Miss%\\': 3.85, \\'Conf%\\': 5.46}}]}, \\'description\\': \\'This API provides speaker diarization functionality using the pyannote.audio framework. It is capable of processing audio files and outputting speaker diarization results in RTTM format. The API supports providing the number of speakers, minimum and maximum number of speakers, and adjusting the segmentation onset threshold.\\'}', metadata={})]", "category": "generic"}
{"question_id": 814, "text": " Our company wants to predict the potential negative impact on the environment based on certain factors.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentiment Analysis\\', \\'api_name\\': \\'michellejieli/NSFW_text_classifier\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'michellejieli/NSFW_text_classification\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'classifier(I see you\u2019ve set aside this special time to humiliate yourself in public.)\\', \\'performance\\': {\\'dataset\\': \\'Reddit posts\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'DistilBERT is a transformer model that performs sentiment analysis. I fine-tuned the model on Reddit posts with the purpose of classifying not safe for work (NSFW) content, specifically text that is considered inappropriate and unprofessional. The model predicts 2 classes, which are NSFW or safe for work (SFW). The model is a fine-tuned version of DistilBERT. It was fine-tuned on 14317 Reddit posts pulled from the Reddit API.\\'}', metadata={})]", "category": "generic"}
{"question_id": 815, "text": " I have a dataset with CO2 emissions in a CSV file, and I want to classify which sources have high or low emissions.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Joblib\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'pachi107/autotrain-in-class-test-1780161764\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data.csv\\'], \\'python_environment_requirements\\': [\\'json\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(data.csv)\\\\ndata = data[features]\\\\ndata.columns = [feat_ + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'Validation Metrics\\', \\'accuracy\\': 0.974}, \\'description\\': \\'A binary classification model for predicting CO2 emissions based on tabular data. Trained using AutoTrain with a model ID of 1780161764.\\'}', metadata={})]", "category": "generic"}
{"question_id": 816, "text": " I am building a recommendation engine to recommend linkbuilding strategy to clients. I have data in csv files that needs to be classified.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Joblib\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'pachi107/autotrain-in-class-test-1780161764\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data.csv\\'], \\'python_environment_requirements\\': [\\'json\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(data.csv)\\\\ndata = data[features]\\\\ndata.columns = [feat_ + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'Validation Metrics\\', \\'accuracy\\': 0.974}, \\'description\\': \\'A binary classification model for predicting CO2 emissions based on tabular data. Trained using AutoTrain with a model ID of 1780161764.\\'}', metadata={})]", "category": "generic"}
{"question_id": 817, "text": " Create a model to estimate the carbon emissions of a specific device by using features like idle power, standby power, and active power.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'45473113800\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'samvelkoch/autotrain-data-prknsn-2\\', \\'accuracy\\': {\\'Loss\\': 5.079, \\'R2\\': 0.109, \\'MSE\\': 25.795, \\'MAE\\': 3.78, \\'RMSLE\\': 0.849}}, \\'description\\': \\'A tabular regression model trained with AutoTrain for predicting carbon emissions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 818, "text": " We have a virtual reality investment in soccer simulations. We need an intelligent learning-based soccer player to make it more engaging and challenging.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Reinforcement Learning\\', \\'framework\\': \\'Unity ML-Agents Library\\', \\'functionality\\': \\'Train and play SoccerTwos\\', \\'api_name\\': \\'poca-SoccerTwosv2\\', \\'api_call\\': \"mlagents-load-from-hf --repo-id=\\'Raiden-1001/poca-SoccerTwosv2\\' --local-dir=\\'./downloads\\'\", \\'api_arguments\\': [\\'your_configuration_file_path.yaml\\', \\'run_id\\'], \\'python_environment_requirements\\': [\\'ml-agents\\'], \\'example_code\\': \\'mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume\\', \\'performance\\': {\\'dataset\\': \\'SoccerTwos\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'A trained model of a poca agent playing SoccerTwos using the Unity ML-Agents Library.\\'}', metadata={})]", "category": "generic"}
{"question_id": 819, "text": " I want to build an AI that identifies the best marketing strategies for my website by trying different combinations of headlines and images.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'abhishek/autotrain-dog-vs-food\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'abhishek/autotrain-dog-vs-food\\')\", \\'api_arguments\\': \\'image_path\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'sasha/dog-food\\', \\'accuracy\\': 0.998}, \\'description\\': \"A pre-trained model for classifying images as either dog or food using Hugging Face\\'s AutoTrain framework.\"}', metadata={})]", "category": "generic"}
{"question_id": 820, "text": " A team is working on a video game where the player needs to land the spaceship on the lunar surface without crashing. They want to implement an AI module that can play the game and test it.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Reinforcement Learning\\', \\'framework\\': \\'Stable-Baselines3\\', \\'functionality\\': \\'LunarLander-v2\\', \\'api_name\\': \\'araffin/dqn-LunarLander-v2\\', \\'api_call\\': \"DQN.load(load_from_hub(\\'araffin/dqn-LunarLander-v2\\', \\'dqn-LunarLander-v2.zip\\'), **kwargs)\", \\'api_arguments\\': {\\'checkpoint\\': \\'araffin/dqn-LunarLander-v2\\', \\'kwargs\\': {\\'target_update_interval\\': 30}}, \\'python_environment_requirements\\': [\\'huggingface_sb3\\', \\'stable_baselines3\\'], \\'example_code\\': {\\'load_model\\': \\'from huggingface_sb3 import load_from_hub\\\\nfrom stable_baselines3 import DQN\\\\nfrom stable_baselines3.common.env_util import make_vec_env\\\\nfrom stable_baselines3.common.evaluation import evaluate_policy\\\\n\\\\ncheckpoint = load_from_hub(araffin/dqn-LunarLander-v2, dqn-LunarLander-v2.zip)\\\\n\\\\nkwargs = dict(target_update_interval=30)\\\\n\\\\nmodel = DQN.load(checkpoint, **kwargs)\\\\nenv = make_vec_env(LunarLander-v2, n_envs=1)\\', \\'evaluate\\': \\'mean_reward, std_reward = evaluate_policy(\\\\n model,\\\\n env,\\\\n n_eval_episodes=20,\\\\n deterministic=True,\\\\n)\\\\nprint(fMean reward = {mean_reward:.2f} +/- {std_reward:.2f})\\'}, \\'performance\\': {\\'dataset\\': \\'LunarLander-v2\\', \\'accuracy\\': \\'280.22 +/- 13.03\\'}, \\'description\\': \\'This is a trained model of a DQN agent playing LunarLander-v2 using the stable-baselines3 library.\\'}', metadata={})]", "category": "generic"}
{"question_id": 821, "text": " Develop an AI character that can play the SoccerTwos game with advanced strategies.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Reinforcement Learning\\', \\'framework\\': \\'Unity ML-Agents Library\\', \\'functionality\\': \\'Train and play SoccerTwos\\', \\'api_name\\': \\'poca-SoccerTwosv2\\', \\'api_call\\': \"mlagents-load-from-hf --repo-id=\\'Raiden-1001/poca-SoccerTwosv2\\' --local-dir=\\'./downloads\\'\", \\'api_arguments\\': [\\'your_configuration_file_path.yaml\\', \\'run_id\\'], \\'python_environment_requirements\\': [\\'ml-agents\\'], \\'example_code\\': \\'mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume\\', \\'performance\\': {\\'dataset\\': \\'SoccerTwos\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'A trained model of a poca agent playing SoccerTwos using the Unity ML-Agents Library.\\'}', metadata={})]", "category": "generic"}
{"question_id": 822, "text": " Our gaming company is looking for a reinforcement learning solution to implement an artificial agent that can play SoccerTwos proficiently.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Reinforcement Learning\\', \\'framework\\': \\'Unity ML-Agents Library\\', \\'functionality\\': \\'Train and play SoccerTwos\\', \\'api_name\\': \\'poca-SoccerTwosv2\\', \\'api_call\\': \"mlagents-load-from-hf --repo-id=\\'Raiden-1001/poca-SoccerTwosv2\\' --local-dir=\\'./downloads\\'\", \\'api_arguments\\': [\\'your_configuration_file_path.yaml\\', \\'run_id\\'], \\'python_environment_requirements\\': [\\'ml-agents\\'], \\'example_code\\': \\'mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume\\', \\'performance\\': {\\'dataset\\': \\'SoccerTwos\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'A trained model of a poca agent playing SoccerTwos using the Unity ML-Agents Library.\\'}', metadata={})]", "category": "generic"}
{"question_id": 823, "text": " Develop a language understanding feature for a chatbot that can help tourists with information on visiting Indonesia.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'hyunwoongko/blenderbot-9B\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'hyunwoongko/blenderbot-9B\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'Input a message to start chatting with hyunwoongko/blenderbot-9B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high-performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi-turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 824, "text": " I need to analyze images in real-time feeds from different locations for object recognition.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Object Detection\\', \\'api_name\\': \\'hustvl/yolos-tiny\\', \\'api_call\\': \"YolosForObjectDetection.from_pretrained(\\'hustvl/yolos-tiny\\')\", \\'api_arguments\\': {\\'images\\': \\'image\\', \\'return_tensors\\': \\'pt\\'}, \\'python_environment_requirements\\': [\\'transformers\\', \\'PIL\\', \\'requests\\'], \\'example_code\\': \"from transformers import YolosFeatureExtractor, YolosForObjectDetection\\\\nfrom PIL import Image\\\\nimport requests\\\\nurl = \\'http://images.cocodataset.org/val2017/000000039769.jpg\\'\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nfeature_extractor = YolosFeatureExtractor.from_pretrained(\\'hustvl/yolos-tiny\\')\\\\nmodel = YolosForObjectDetection.from_pretrained(\\'hustvl/yolos-tiny\\')\\\\ninputs = feature_extractor(images=image, return_tensors=pt)\\\\noutputs = model(**inputs)\\\\nlogits = outputs.logits\\\\nbboxes = outputs.pred_boxes\", \\'performance\\': {\\'dataset\\': \\'COCO 2017 validation\\', \\'accuracy\\': \\'28.7 AP\\'}, \\'description\\': \\'YOLOS is a Vision Transformer (ViT) trained using the DETR loss. Despite its simplicity, a base-sized YOLOS model is able to achieve 42 AP on COCO validation 2017 (similar to DETR and more complex frameworks such as Faster R-CNN). The model is trained using a bipartite matching loss: one compares the predicted classes + bounding boxes of each of the N = 100 object queries to the ground truth annotations, padded up to the same length N (so if an image only contains 4 objects, 96 annotations will just have a no object as class and no bounding box as bounding box). The Hungarian matching algorithm is used to create an optimal one-to-one mapping between each of the N queries and each of the N annotations. Next, standard cross-entropy (for the classes) and a linear combination of the L1 and generalized IoU loss (for the bounding boxes) are used to optimize the parameters of the model.\\'}', metadata={})]", "category": "generic"}
{"question_id": 825, "text": " Our travel agency needs to build a chatbot that helps tourists find the best attractions in various destinations. The chatbot should answer questions related to tourist attractions.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'hyunwoongko/blenderbot-9B\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'hyunwoongko/blenderbot-9B\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'Input a message to start chatting with hyunwoongko/blenderbot-9B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high-performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi-turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 826, "text": " We want to develop a chatbot that can engage with multilingual users. Please help us create a model to encode sentences and understand user input in various languages.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'hyunwoongko/blenderbot-9B\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'hyunwoongko/blenderbot-9B\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'Input a message to start chatting with hyunwoongko/blenderbot-9B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high-performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi-turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 827, "text": " We need to create a demo for a 3D printing company by generating images of some examples that the business will print.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'gsdf/Counterfeit-V2.5\\', \\'api_call\\': \"pipeline(\\'text-to-image\\', model=\\'gsdf/Counterfeit-V2.5\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'((masterpiece,best quality)),1girl, solo, animal ears, rabbit, barefoot, knees up, dress, sitting, rabbit ears, short sleeves, looking at viewer, grass, short hair, smile, white hair, puffy sleeves, outdoors, puffy short sleeves, bangs, on ground, full body, animal, white dress, sunlight, brown eyes, dappled sunlight, day, depth of field\\', \\'performance\\': {\\'dataset\\': \\'EasyNegative\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Counterfeit-V2.5 is a text-to-image model that generates anime-style images based on text prompts. It has been updated for ease of use and can be used with negative prompts to create high-quality images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 828, "text": " I work at an art school and our professor wants to create an AI chatbot that can study an image of a painting and answer questions about it.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'hyunwoongko/blenderbot-9B\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'hyunwoongko/blenderbot-9B\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'Input a message to start chatting with hyunwoongko/blenderbot-9B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high-performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi-turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 829, "text": " We are working on a project to develop an AI-based method to provide answers to questions about charts and diagrams. Please provide the necessary code for using the Pix2Struct model.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Visual Question Answering\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'google/pix2struct-chartqa-base\\', \\'api_call\\': \"Pix2StructForConditionalGeneration.from_pretrained(\\'google/pix2struct-chartqa-base\\')\", \\'api_arguments\\': [\\'t5x_checkpoint_path\\', \\'pytorch_dump_path\\', \\'use-large\\'], \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'python convert_pix2struct_checkpoint_to_pytorch.py --t5x_checkpoint_path PATH_TO_T5X_CHECKPOINTS --pytorch_dump_path PATH_TO_SAVE\\', \\'performance\\': {\\'dataset\\': \\'ChartQA\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Pix2Struct is an image encoder - text decoder model that is trained on image-text pairs for various tasks, including image captionning and visual question answering. The model is pretrained by learning to parse masked screenshots of web pages into simplified HTML. It can achieve state-of-the-art results in six out of nine tasks across four domains: documents, illustrations, user interfaces, and natural images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 830, "text": " Create a personalized online ad campaign for our new clothing line. The ad should generate a video showing a man wearing a stylish suit while walking in the city.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'chavinlo/TempoFunk\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'chavinlo/TempoFunk\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Text-to-Video model using Hugging Face Transformers library. Model is capable of generating video content based on the input text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 831, "text": " We need to generate a short video showing Spider-Man water skiing in redshift style based on a textual prompt.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video Generation\\', \\'api_name\\': \\'redshift-man-skiing\\', \\'api_call\\': \"TuneAVideoPipeline.from_pretrained(\\'nitrosocke/redshift-diffusion\\', unet=UNet3DConditionModel.from_pretrained(\\'Tune-A-Video-library/redshift-man-skiing\\', subfolder=\\'unet\\', torch_dtype=torch.float16), torch_dtype=torch.float16)\", \\'api_arguments\\': {\\'prompt\\': \\'string\\', \\'video_length\\': \\'int\\', \\'height\\': \\'int\\', \\'width\\': \\'int\\', \\'num_inference_steps\\': \\'int\\', \\'guidance_scale\\': \\'float\\'}, \\'python_environment_requirements\\': [\\'torch\\', \\'tuneavideo\\'], \\'example_code\\': \"from tuneavideo.pipelines.pipeline_tuneavideo import TuneAVideoPipeline\\\\nfrom tuneavideo.models.unet import UNet3DConditionModel\\\\nfrom tuneavideo.util import save_videos_grid\\\\nimport torch\\\\npretrained_model_path = nitrosocke/redshift-diffusion\\\\nunet_model_path = Tune-A-Video-library/redshift-man-skiing\\\\nunet = UNet3DConditionModel.from_pretrained(unet_model_path, subfolder=\\'unet\\', torch_dtype=torch.float16).to(\\'cuda\\')\\\\npipe = TuneAVideoPipeline.from_pretrained(pretrained_model_path, unet=unet, torch_dtype=torch.float16).to(cuda)\\\\npipe.enable_xformers_memory_efficient_attention()\\\\nprompt = (redshift style) spider man is skiing\\\\nvideo = pipe(prompt, video_length=8, height=512, width=512, num_inference_steps=50, guidance_scale=7.5).videos\\\\nsave_videos_grid(video, f./{prompt}.gif)\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \"Tune-A-Video - Redshift is a text-to-video generation model based on the nitrosocke/redshift-diffusion model. It generates videos based on textual prompts, such as \\'a man is skiing\\' or \\'(redshift style) spider man is skiing\\'.\"}', metadata={})]", "category": "generic"}
{"question_id": 832, "text": " Our company develops travel guides in Polish language. We are now planning to use artificial intelligence for quick captioning of Polish images.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Visual Question Answering\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'azwierzc/vilt-b32-finetuned-vqa-pl\\', \\'api_call\\': \"pipeline(\\'visual-question-answering\\', model=\\'azwierzc/vilt-b32-finetuned-vqa-pl\\')\", \\'api_arguments\\': {\\'image\\': \\'path_to_image\\', \\'question\\': \\'question_text\\'}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Visual Question Answering model fine-tuned on the Polish language.\\'}', metadata={})]", "category": "generic"}
{"question_id": 833, "text": " A user has difficulty in visualizing images and needs help answering questions about an image. We need a model to answer questions based on the image provided.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 834, "text": " A real estate company wants to extract information from a scan of a property listing. They need a solution to recognize and retrieve critical information from the OCR of the scan.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Image-to-Text\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'kha-white/manga-ocr-base\\', \\'api_call\\': \"pipeline(\\'ocr\\', model=\\'kha-white/manga-ocr-base\\')\", \\'api_arguments\\': \\'image\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'manga109s\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Optical character recognition for Japanese text, with the main focus being Japanese manga. It uses Vision Encoder Decoder framework. Manga OCR can be used as a general purpose printed Japanese OCR, but its main goal was to provide a high quality text recognition, robust against various scenarios specific to manga: both vertical and horizontal text, text with furigana, text overlaid on images, wide variety of fonts and font styles, and low quality images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 835, "text": " We want to utilize the machine learning model for predicting molecular properties in the drug discovery domain.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Carbon Emissions Prediction\\', \\'api_name\\': \\'bibekbehera/autotrain-numeric_prediction-40376105019\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'bibekbehera/autotrain-data-numeric_prediction\\', \\'accuracy\\': {\\'Loss\\': 0.152, \\'R2\\': 0.659, \\'MSE\\': 0.023, \\'MAE\\': 0.062, \\'RMSLE\\': 0.105}}, \\'description\\': \\'A tabular regression model trained with AutoTrain to predict carbon emissions based on input features.\\'}', metadata={})]", "category": "generic"}
{"question_id": 836, "text": " We need to estimate the depth of a scene in an image using a pretrained model. Can you please suggest a way?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Depth Estimation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Depth Estimation\\', \\'api_name\\': \\'glpn-kitti\\', \\'api_call\\': \"GLPNForDepthEstimation.from_pretrained(\\'vinvino02/glpn-kitti\\')\", \\'api_arguments\\': \\'images, return_tensors\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'from transformers import GLPNFeatureExtractor, GLPNForDepthEstimation\\\\nimport torch\\\\nimport numpy as np\\\\nfrom PIL import Image\\\\nimport requests\\\\nurl = http://images.cocodataset.org/val2017/000000039769.jpg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nfeature_extractor = GLPNFeatureExtractor.from_pretrained(vinvino02/glpn-kitti)\\\\nmodel = GLPNForDepthEstimation.from_pretrained(vinvino02/glpn-kitti)\\\\ninputs = feature_extractor(images=image, return_tensors=pt)\\\\nwith torch.no_grad():\\\\n outputs = model(**inputs)\\\\n predicted_depth = outputs.predicted_depth\\\\nprediction = torch.nn.functional.interpolate(\\\\n predicted_depth.unsqueeze(1),\\\\n size=image.size[::-1],\\\\n mode=bicubic,\\\\n align_corners=False,\\\\n)\\\\noutput = prediction.squeeze().cpu().numpy()\\\\nformatted = (output * 255 / np.max(output)).astype(uint8)\\\\ndepth = Image.fromarray(formatted)\\', \\'performance\\': {\\'dataset\\': \\'KITTI\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Global-Local Path Networks (GLPN) model trained on KITTI for monocular depth estimation. It was introduced in the paper Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth by Kim et al. and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 837, "text": " Our company is working on an autonomous robot and needs a solution to estimate the depth of objects in its environment.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Depth Estimation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Depth Estimation\\', \\'api_name\\': \\'glpn-kitti\\', \\'api_call\\': \"GLPNForDepthEstimation.from_pretrained(\\'vinvino02/glpn-kitti\\')\", \\'api_arguments\\': \\'images, return_tensors\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'from transformers import GLPNFeatureExtractor, GLPNForDepthEstimation\\\\nimport torch\\\\nimport numpy as np\\\\nfrom PIL import Image\\\\nimport requests\\\\nurl = http://images.cocodataset.org/val2017/000000039769.jpg\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\nfeature_extractor = GLPNFeatureExtractor.from_pretrained(vinvino02/glpn-kitti)\\\\nmodel = GLPNForDepthEstimation.from_pretrained(vinvino02/glpn-kitti)\\\\ninputs = feature_extractor(images=image, return_tensors=pt)\\\\nwith torch.no_grad():\\\\n outputs = model(**inputs)\\\\n predicted_depth = outputs.predicted_depth\\\\nprediction = torch.nn.functional.interpolate(\\\\n predicted_depth.unsqueeze(1),\\\\n size=image.size[::-1],\\\\n mode=bicubic,\\\\n align_corners=False,\\\\n)\\\\noutput = prediction.squeeze().cpu().numpy()\\\\nformatted = (output * 255 / np.max(output)).astype(uint8)\\\\ndepth = Image.fromarray(formatted)\\', \\'performance\\': {\\'dataset\\': \\'KITTI\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Global-Local Path Networks (GLPN) model trained on KITTI for monocular depth estimation. It was introduced in the paper Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth by Kim et al. and first released in this repository.\\'}', metadata={})]", "category": "generic"}
{"question_id": 838, "text": " I want to create a video streaming app for plants, in which plants are categorized by species. I need to find the species of plants in an image.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Image-to-Image\\', \\'api_name\\': \\'GreeneryScenery/SheepsControlV5\\', \\'api_call\\': \"pipeline(\\'image-to-image\\', model=\\'GreeneryScenery/SheepsControlV5\\')\", \\'api_arguments\\': {\\'input_image\\': \\'path/to/image/file\\'}, \\'python_environment_requirements\\': {\\'huggingface_hub\\': \\'>=0.0.17\\', \\'transformers\\': \\'>=4.13.0\\', \\'torch\\': \\'>=1.10.0\\'}, \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'poloclub/diffusiondb\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'SheepsControlV5 is an image-to-image model trained on the poloclub/diffusiondb dataset. It is designed for transforming input images into a different style or representation.\\'}', metadata={})]", "category": "generic"}
{"question_id": 839, "text": " Develop an image classifier for distinguishing images of cats and dogs.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'abhishek/autotrain-dog-vs-food\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'abhishek/autotrain-dog-vs-food\\')\", \\'api_arguments\\': \\'image_path\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'sasha/dog-food\\', \\'accuracy\\': 0.998}, \\'description\\': \"A pre-trained model for classifying images as either dog or food using Hugging Face\\'s AutoTrain framework.\"}', metadata={})]", "category": "generic"}
{"question_id": 840, "text": " We are building AI glasses that should tell us about things that we are seeing with classifications. We want to use a visual transformer architecture.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'martinezomg/vit-base-patch16-224-diabetic-retinopathy\\', \\'api_call\\': \"pipeline(\\'image-classification\\', \\'martinezomg/vit-base-patch16-224-diabetic-retinopathy\\')\", \\'api_arguments\\': {\\'model_name\\': \\'martinezomg/vit-base-patch16-224-diabetic-retinopathy\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'4.28.1\\', \\'pytorch\\': \\'2.0.0+cu118\\', \\'datasets\\': \\'2.11.0\\', \\'tokenizers\\': \\'0.13.3\\'}, \\'example_code\\': \"from transformers import pipeline\\\\nimage_classifier = pipeline(\\'image-classification\\', \\'martinezomg/vit-base-patch16-224-diabetic-retinopathy\\')\\\\nresult = image_classifier(\\'path/to/image.jpg\\')\", \\'performance\\': {\\'dataset\\': \\'None\\', \\'accuracy\\': 0.7744}, \\'description\\': \\'This model is a fine-tuned version of google/vit-base-patch16-224 on the None dataset. It is designed for image classification tasks, specifically for diabetic retinopathy detection.\\'}', metadata={})]", "category": "generic"}
{"question_id": 841, "text": " A marketing firm has asked us to build an application to classify social media images into various categories.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'abhishek/autotrain-dog-vs-food\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'abhishek/autotrain-dog-vs-food\\')\", \\'api_arguments\\': \\'image_path\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'sasha/dog-food\\', \\'accuracy\\': 0.998}, \\'description\\': \"A pre-trained model for classifying images as either dog or food using Hugging Face\\'s AutoTrain framework.\"}', metadata={})]", "category": "generic"}
{"question_id": 842, "text": " John is a car enthusiast and he wants to build an application that can recognize the car brand in a picture he takes.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Joblib\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'tejas23/autotrain-amx2-1702259725\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': {\\'accuracy\\': 0.827}}, \\'description\\': \\'Multi-class Classification Model for Carbon Emissions\\'}', metadata={})]", "category": "generic"}
{"question_id": 843, "text": " We are building an app to help Counter-Strike: Global Offensive players improve their gameplay. We need to detect players in a live game of CS:GO.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Object Detection\\', \\'api_name\\': \\'keremberke/yolov8n-csgo-player-detection\\', \\'api_call\\': \"YOLO(\\'keremberke/yolov8n-csgo-player-detection\\')\", \\'api_arguments\\': {\\'image\\': \\'URL or local path to image\\'}, \\'python_environment_requirements\\': \\'pip install ultralyticsplus==0.0.23 ultralytics==8.0.21\\', \\'example_code\\': \"from ultralyticsplus import YOLO, render_result\\\\n\\\\nmodel = YOLO(\\'keremberke/yolov8n-csgo-player-detection\\')\\\\n\\\\nmodel.overrides[\\'conf\\'] = 0.25\\\\nmodel.overrides[\\'iou\\'] = 0.45\\\\nmodel.overrides[\\'agnostic_nms\\'] = False\\\\nmodel.overrides[\\'max_det\\'] = 1000\\\\n\\\\nimage = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\\\\n\\\\nresults = model.predict(image)\\\\n\\\\nprint(results[0].boxes)\\\\nrender = render_result(model=model, image=image, result=results[0])\\\\nrender.show()\", \\'performance\\': {\\'dataset\\': \\'csgo-object-detection\\', \\'accuracy\\': 0.844}, \\'description\\': \"A YOLOv8 model for detecting Counter-Strike: Global Offensive (CS:GO) players with supported labels: [\\'ct\\', \\'cthead\\', \\'t\\', \\'thead\\'].\"}', metadata={})]", "category": "generic"}
{"question_id": 844, "text": " The security department needs assistance to detect suspicious objects and people using a zero-shot text-conditioned object detection system.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Detect Bordered and Borderless tables in documents\\', \\'api_name\\': \\'TahaDouaji/detr-doc-table-detection\\', \\'api_call\\': \"DetrForObjectDetection.from_pretrained(\\'TahaDouaji/detr-doc-table-detection\\')\", \\'api_arguments\\': [\\'images\\', \\'return_tensors\\', \\'threshold\\'], \\'python_environment_requirements\\': [\\'transformers\\', \\'torch\\', \\'PIL\\', \\'requests\\'], \\'example_code\\': \\'from transformers import DetrImageProcessor, DetrForObjectDetection\\\\nimport torch\\\\nfrom PIL import Image\\\\nimport requests\\\\nimage = Image.open(IMAGE_PATH)\\\\nprocessor = DetrImageProcessor.from_pretrained(TahaDouaji/detr-doc-table-detection)\\\\nmodel = DetrForObjectDetection.from_pretrained(TahaDouaji/detr-doc-table-detection)\\\\ninputs = processor(images=image, return_tensors=pt)\\\\noutputs = model(**inputs)\\\\ntarget_sizes = torch.tensor([image.size[::-1]])\\\\nresults = processor.post_process_object_detection(outputs, target_sizes=target_sizes, threshold=0.9)[0]\\\\nfor score, label, box in zip(results[scores], results[labels], results[boxes]):\\\\n box = [round(i, 2) for i in box.tolist()]\\\\n print(\\\\n fDetected {model.config.id2label[label.item()]} with confidence \\\\n f{round(score.item(), 3)} at location {box}\\\\n )\\', \\'performance\\': {\\'dataset\\': \\'ICDAR2019 Table Dataset\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'detr-doc-table-detection is a model trained to detect both Bordered and Borderless tables in documents, based on facebook/detr-resnet-50.\\'}', metadata={})]", "category": "generic"}
{"question_id": 845, "text": " As a specialist in computer vision, we need to use the OwlViT model to identify objects in an image described by specific text phrases like \\\"a photo of a cat\\\" and \\\"a photo of a dog.\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'zero-shot-object-detection\\', \\'api_name\\': \\'google/owlvit-large-patch14\\', \\'api_call\\': \"OwlViTForObjectDetection.from_pretrained(\\'google/owlvit-large-patch14\\')\", \\'api_arguments\\': {\\'model_name\\': \\'google/owlvit-large-patch14\\'}, \\'python_environment_requirements\\': [\\'torch\\', \\'transformers\\', \\'PIL\\', \\'requests\\'], \\'example_code\\': [\\'import requests\\', \\'from PIL import Image\\', \\'import torch\\', \\'from transformers import OwlViTProcessor, OwlViTForObjectDetection\\', \\'processor = OwlViTProcessor.from_pretrained(google/owlvit-large-patch14)\\', \\'model = OwlViTForObjectDetection.from_pretrained(google/owlvit-large-patch14)\\', \\'url = http://images.cocodataset.org/val2017/000000039769.jpg\\', \\'image = Image.open(requests.get(url, stream=True).raw)\\', \\'texts = [[a photo of a cat, a photo of a dog]\\', \\'inputs = processor(text=texts, images=image, return_tensors=pt)\\', \\'outputs = model(**inputs)\\', \\'target_sizes = torch.Tensor([image.size[::-1]])\\', \\'results = processor.post_process(outputs=outputs, target_sizes=target_sizes)\\', \\'i = 0\\', \\'text = texts[i]\\', \\'boxes, scores, labels = results[i][boxes], results[i][scores], results[i][labels]\\', \\'score_threshold = 0.1\\', \\'for box, score, label in zip(boxes, scores, labels):\\', \\' box = [round(i, 2) for i in box.tolist()]\\', \\' if score >= score_threshold:\\', \\' print(fDetected {text[label]} with confidence {round(score.item(), 3)} at location {box})\\'], \\'performance\\': {\\'dataset\\': \\'COCO\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'OWL-ViT is a zero-shot text-conditioned object detection model that can be used to query an image with one or multiple text queries. It uses CLIP as its multi-modal backbone, with a ViT-like Transformer to get visual features and a causal language model to get the text features. OWL-ViT is trained on publicly available image-caption data and fine-tuned on publicly available object detection datasets such as COCO and OpenImages.\\'}', metadata={})]", "category": "generic"}
{"question_id": 846, "text": " A medical research team requests an automated procedure for detecting blood cells in microscopic images of blood samples. Develop a solution to address this need.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Blood Cell Detection\\', \\'api_name\\': \\'keremberke/yolov8n-blood-cell-detection\\', \\'api_call\\': \"YOLO(\\'keremberke/yolov8n-blood-cell-detection\\')\", \\'api_arguments\\': {\\'conf\\': 0.25, \\'iou\\': 0.45, \\'agnostic_nms\\': False, \\'max_det\\': 1000}, \\'python_environment_requirements\\': \\'ultralyticsplus==0.0.23 ultralytics==8.0.21\\', \\'example_code\\': \"from ultralyticsplus import YOLO, render_result\\\\nmodel = YOLO(\\'keremberke/yolov8n-blood-cell-detection\\')\\\\nmodel.overrides[\\'conf\\'] = 0.25\\\\nmodel.overrides[\\'iou\\'] = 0.45\\\\nmodel.overrides[\\'agnostic_nms\\'] = False\\\\nmodel.overrides[\\'max_det\\'] = 1000\\\\nimage = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\\\\nresults = model.predict(image)\\\\nprint(results[0].boxes)\\\\nrender = render_result(model=model, image=image, result=results[0])\\\\nrender.show()\", \\'performance\\': {\\'dataset\\': \\'blood-cell-object-detection\\', \\'accuracy\\': 0.893}, \\'description\\': \\'This model detects blood cells in images, specifically Platelets, RBC, and WBC. It is based on the YOLOv8 architecture and trained on the blood-cell-object-detection dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 847, "text": " Our client has a traffic camera system and wants to detect vehicles in the images captured. Implement a solution.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'License Plate Detection\\', \\'api_name\\': \\'keremberke/yolov5s-license-plate\\', \\'api_call\\': \"yolov5.load(\\'keremberke/yolov5s-license-plate\\')\", \\'api_arguments\\': {\\'img\\': \\'image url or path\\', \\'size\\': \\'image resize dimensions\\', \\'augment\\': \\'optional, test time augmentation\\'}, \\'python_environment_requirements\\': \\'pip install -U yolov5\\', \\'example_code\\': [\\'import yolov5\\', \"model = yolov5.load(\\'keremberke/yolov5s-license-plate\\')\", \\'model.conf = 0.25\\', \\'model.iou = 0.45\\', \\'model.agnostic = False\\', \\'model.multi_label = False\\', \\'model.max_det = 1000\\', \"img = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\", \\'results = model(img, size=640)\\', \\'results = model(img, augment=True)\\', \\'predictions = results.pred[0]\\', \\'boxes = predictions[:, :4]\\', \\'scores = predictions[:, 4]\\', \\'categories = predictions[:, 5]\\', \\'results.show()\\', \"results.save(save_dir=\\'results/\\')\"], \\'performance\\': {\\'dataset\\': \\'keremberke/license-plate-object-detection\\', \\'accuracy\\': 0.985}, \\'description\\': \\'A YOLOv5 based license plate detection model trained on a custom dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 848, "text": " Detect the location of players in an image from a Counter-Strike: Global Offensive (CS:GO) game.\\n###Input: image_path = \\\"path/to/your/csgo_image.jpg\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Object Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Object Detection\\', \\'api_name\\': \\'keremberke/yolov8n-csgo-player-detection\\', \\'api_call\\': \"YOLO(\\'keremberke/yolov8n-csgo-player-detection\\')\", \\'api_arguments\\': {\\'image\\': \\'URL or local path to image\\'}, \\'python_environment_requirements\\': \\'pip install ultralyticsplus==0.0.23 ultralytics==8.0.21\\', \\'example_code\\': \"from ultralyticsplus import YOLO, render_result\\\\n\\\\nmodel = YOLO(\\'keremberke/yolov8n-csgo-player-detection\\')\\\\n\\\\nmodel.overrides[\\'conf\\'] = 0.25\\\\nmodel.overrides[\\'iou\\'] = 0.45\\\\nmodel.overrides[\\'agnostic_nms\\'] = False\\\\nmodel.overrides[\\'max_det\\'] = 1000\\\\n\\\\nimage = \\'https://github.com/ultralytics/yolov5/raw/master/data/images/zidane.jpg\\'\\\\n\\\\nresults = model.predict(image)\\\\n\\\\nprint(results[0].boxes)\\\\nrender = render_result(model=model, image=image, result=results[0])\\\\nrender.show()\", \\'performance\\': {\\'dataset\\': \\'csgo-object-detection\\', \\'accuracy\\': 0.844}, \\'description\\': \"A YOLOv8 model for detecting Counter-Strike: Global Offensive (CS:GO) players with supported labels: [\\'ct\\', \\'cthead\\', \\'t\\', \\'thead\\'].\"}', metadata={})]", "category": "generic"}
{"question_id": 849, "text": " Develop a fashion app that segments and identifies clothing items in an uploaded image.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Segmentation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Segmentation\\', \\'api_name\\': \\'mattmdjaga/segformer_b2_clothes\\', \\'api_call\\': \"SegformerForSemanticSegmentation.from_pretrained(\\'mattmdjaga/segformer_b2_clothes\\')\", \\'api_arguments\\': [\\'image\\'], \\'python_environment_requirements\\': [\\'transformers\\', \\'PIL\\', \\'requests\\', \\'matplotlib\\', \\'torch\\'], \\'example_code\\': \"from transformers import AutoFeatureExtractor, SegformerForSemanticSegmentation\\\\nfrom PIL import Image\\\\nimport requests\\\\nimport matplotlib.pyplot as plt\\\\nimport torch.nn as nn\\\\nextractor = AutoFeatureExtractor.from_pretrained(\\'mattmdjaga/segformer_b2_clothes\\')\\\\nmodel = SegformerForSemanticSegmentation.from_pretrained(\\'mattmdjaga/segformer_b2_clothes\\')\\\\nurl = \\'https://plus.unsplash.com/premium_photo-1673210886161-bfcc40f54d1f?ixlib=rb-4.0.3&amp;ixid=MnwxMjA3fDB8MHxzZWFyY2h8MXx8cGVyc29uJTIwc3RhbmRpbmd8ZW58MHx8MHx8&amp;w=1000&amp;q=80\\'\\\\nimage = Image.open(requests.get(url, stream=True).raw)\\\\ninputs = extractor(images=image, return_tensors=\\'pt\\')\\\\noutputs = model(**inputs)\\\\nlogits = outputs.logits.cpu()\\\\nupsampled_logits = nn.functional.interpolate(logits, size=image.size[::-1], mode=\\'bilinear\\', align_corners=False)\\\\npred_seg = upsampled_logits.argmax(dim=1)[0]\\\\nplt.imshow(pred_seg)\", \\'performance\\': {\\'dataset\\': \\'mattmdjaga/human_parsing_dataset\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'SegFormer model fine-tuned on ATR dataset for clothes segmentation.\\'}', metadata={})]", "category": "generic"}
{"question_id": 850, "text": " Our team is working on a project to develop autonomous vehicles. We need a model to identify different segments of an image captured from the vehicle's camera.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Image Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Image Classification\\', \\'api_name\\': \\'abhishek/autotrain-dog-vs-food\\', \\'api_call\\': \"pipeline(\\'image-classification\\', model=\\'abhishek/autotrain-dog-vs-food\\')\", \\'api_arguments\\': \\'image_path\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'sasha/dog-food\\', \\'accuracy\\': 0.998}, \\'description\\': \"A pre-trained model for classifying images as either dog or food using Hugging Face\\'s AutoTrain framework.\"}', metadata={})]", "category": "generic"}
{"question_id": 851, "text": " A real estate agency needs an application that can transform the floor plan images into simple straight line drawings, simplifying the visualization of the properties.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'Lykon/DreamShaper\\', \\'api_call\\': \"pipeline(\\'text-to-image\\', model=Lykon/DreamShaper)\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers, torch\\', \\'example_code\\': \\'https://huggingface.co/spaces/Lykon/DreamShaper-webui\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Dream Shaper is a text-to-image model that generates artistic images based on the given input text. Read more about this model here: https://civitai.com/models/4384/dreamshaper\\'}', metadata={})]", "category": "generic"}
{"question_id": 852, "text": " I'm working on creating images of various scenes based on their textual descriptions. The models should also consider the actual positions and poses of the objects in the scenes.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Image\\', \\'api_name\\': \\'gsdf/Counterfeit-V2.5\\', \\'api_call\\': \"pipeline(\\'text-to-image\\', model=\\'gsdf/Counterfeit-V2.5\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'((masterpiece,best quality)),1girl, solo, animal ears, rabbit, barefoot, knees up, dress, sitting, rabbit ears, short sleeves, looking at viewer, grass, short hair, smile, white hair, puffy sleeves, outdoors, puffy short sleeves, bangs, on ground, full body, animal, white dress, sunlight, brown eyes, dappled sunlight, day, depth of field\\', \\'performance\\': {\\'dataset\\': \\'EasyNegative\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Counterfeit-V2.5 is a text-to-image model that generates anime-style images based on text prompts. It has been updated for ease of use and can be used with negative prompts to create high-quality images.\\'}', metadata={})]", "category": "generic"}
{"question_id": 853, "text": " Develop a new product that uses machine learning to enhance the quality of low-resolution images by upscaling them to twice their size.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Image\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Image Upscaling\\', \\'api_name\\': \\'stabilityai/sd-x2-latent-upscaler\\', \\'api_call\\': \\'StableDiffusionLatentUpscalePipeline.from_pretrained(stabilityai/sd-x2-latent-upscaler, torch_dtype=torch.float16)\\', \\'api_arguments\\': {\\'prompt\\': \\'text prompt\\', \\'image\\': \\'low resolution latents\\', \\'num_inference_steps\\': 20, \\'guidance_scale\\': 0, \\'generator\\': \\'torch generator\\'}, \\'python_environment_requirements\\': [\\'git+https://github.com/huggingface/diffusers.git\\', \\'transformers\\', \\'accelerate\\', \\'scipy\\', \\'safetensors\\'], \\'example_code\\': \\'from diffusers import StableDiffusionLatentUpscalePipeline, StableDiffusionPipeline\\\\nimport torch\\\\npipeline = StableDiffusionPipeline.from_pretrained(CompVis/stable-diffusion-v1-4, torch_dtype=torch.float16)\\\\npipeline.to(cuda)\\\\nupscaler = StableDiffusionLatentUpscalePipeline.from_pretrained(stabilityai/sd-x2-latent-upscaler, torch_dtype=torch.float16)\\\\nupscaler.to(cuda)\\\\nprompt = a photo of an astronaut high resolution, unreal engine, ultra realistic\\\\ngenerator = torch.manual_seed(33)\\\\nlow_res_latents = pipeline(prompt, generator=generator, output_type=latent).images\\\\nupscaled_image = upscaler(prompt=prompt, image=low_res_latents, num_inference_steps=20, guidance_scale=0, generator=generator).images[0]\\\\nupscaled_image.save(astronaut_1024.png)\\', \\'performance\\': {\\'dataset\\': \\'LAION-2B\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \"Stable Diffusion x2 latent upscaler is a diffusion-based upscaler model developed by Katherine Crowson in collaboration with Stability AI. It is designed to upscale Stable Diffusion\\'s latent denoised image embeddings, allowing for fast text-to-image and upscaling pipelines. The model was trained on a high-resolution subset of the LAION-2B dataset and works with all Stable Diffusion checkpoints.\"}', metadata={})]", "category": "generic"}
{"question_id": 854, "text": " An animal shelter needs an original cat image for a fundraising event poster. Generate the image.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Computer Vision Unconditional Image Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Unconditional Image Generation\\', \\'api_name\\': \\'google/ddpm-cat-256\\', \\'api_call\\': \"DDPMPipeline.from_pretrained(\\'google/ddpm-cat-256\\')\", \\'api_arguments\\': [\\'model_id\\'], \\'python_environment_requirements\\': [\\'diffusers\\'], \\'example_code\\': \\'!pip install diffusers\\\\nfrom diffusers import DDPMPipeline, DDIMPipeline, PNDMPipeline\\\\nmodel_id = google/ddpm-cat-256\\\\nddpm = DDPMPipeline.from_pretrained(model_id)\\\\nimage = ddpm().images[0]\\\\nimage.save(ddpm_generated_image.png)\\', \\'performance\\': {\\'dataset\\': \\'CIFAR10\\', \\'accuracy\\': {\\'Inception_score\\': 9.46, \\'FID_score\\': 3.17}}, \\'description\\': \\'Denoising Diffusion Probabilistic Models (DDPM) is a class of latent variable models inspired by considerations from nonequilibrium thermodynamics. It can generate high-quality images using discrete noise schedulers such as scheduling_ddpm, scheduling_ddim, and scheduling_pndm. The model is trained on the unconditional CIFAR10 dataset and 256x256 LSUN, obtaining an Inception score of 9.46 and a state-of-the-art FID score of 3.17.\\'}', metadata={})]", "category": "generic"}
{"question_id": 855, "text": " Develop a video content recommendation engine that can understand and generate multiple categories, such as sports, comedy, and news, based on the videos.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'ImRma/Brucelee\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'ImRma/Brucelee\\')\", \\'api_arguments\\': [\\'your_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Hugging Face model for converting Persian and English text into video.\\'}', metadata={})]", "category": "generic"}
{"question_id": 856, "text": " Our customer is a fitness platform. We need to analyze workout videos for offering customized workout plans.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'chavinlo/TempoFunk\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'chavinlo/TempoFunk\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Text-to-Video model using Hugging Face Transformers library. Model is capable of generating video content based on the input text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 857, "text": " We are a security company and we need a video classification model to analyze CCTV footage for suspicious activities.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Text-to-Video\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Text-to-Video\\', \\'api_name\\': \\'chavinlo/TempoFunk\\', \\'api_call\\': \"pipeline(\\'text-to-video\\', model=\\'chavinlo/TempoFunk\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Text-to-Video model using Hugging Face Transformers library. Model is capable of generating video content based on the input text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 858, "text": " Design a model to classify the following image: a city park with a playground and a lake, surrounded by trees and skyscrapers.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Keras\\', \\'functionality\\': \\'Binary Classification\\', \\'api_name\\': \\'TF_Decision_Trees\\', \\'api_call\\': \\'TF_Decision_Trees(input_features, target)\\', \\'api_arguments\\': [\\'input_features\\', \\'target\\'], \\'python_environment_requirements\\': [\\'tensorflow >= 7.0\\'], \\'example_code\\': \\'https://github.com/tdubon/TF-GB-Forest/blob/c0cf4c7e3e29d819b996cfe4eecc1f2728115e52/TFDecisionTrees_Final.ipynb\\', \\'performance\\': {\\'dataset\\': \\'Census-Income Data Set\\', \\'accuracy\\': 96.57}, \\'description\\': \"Use TensorFlow\\'s Gradient Boosted Trees model in binary classification of structured data. Build a decision forests model by specifying the input feature usage. Implement a custom Binary Target encoder as a Keras Preprocessing layer to encode the categorical features with respect to their target value co-occurrences, and then use the encoded features to build a decision forests model. The model is trained on the US Census Income Dataset containing approximately 300k instances with 41 numerical and categorical variables. The task is to determine whether a person makes over 50k a year.\"}', metadata={})]", "category": "generic"}
{"question_id": 859, "text": " We are integrating a chatbot into our system. We want the chatbot to first detect the language of user input before providing a response.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Generation\\', \\'api_name\\': \\'mywateriswet/ShuanBot\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'mywateriswet/ShuanBot\\')\", \\'api_arguments\\': \\'message\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"response = chatbot(\\'What is your name?\\')\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'ShuanBot is a conversational chatbot model based on the GPT-2 architecture. It can be used for generating human-like responses in a chat context.\\'}', metadata={})]", "category": "generic"}
{"question_id": 860, "text": " We have a dataset with customer reviews of our financial service app, and we'd like to analyze their sentiment.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'sentiment_analysis_generic_dataset\\', \\'api_call\\': \"pipeline(\\'text-classification\\', model=\\'Seethal/sentiment_analysis_generic_dataset\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"sentiment_analysis(\\'I love this product!\\')\", \\'performance\\': {\\'dataset\\': \\'generic_dataset\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'This is a fine-tuned downstream version of the bert-base-uncased model for sentiment analysis, this model is not intended for further downstream fine-tuning for any other tasks. This model is trained on a classified dataset for text classification.\\'}', metadata={})]", "category": "generic"}
{"question_id": 861, "text": " Our organization sells movies. We need to collect reviews from various platforms to understand the popularity of a movie.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Joblib\\', \\'functionality\\': \\'Multi-class Classification\\', \\'api_name\\': \\'Alexei1/imdb\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'json\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'IMDB\\', \\'accuracy\\': 0.487}, \\'description\\': \\'A tabular classification model trained using AutoTrain for sentiment analysis on the IMDB dataset. The model has a CO2 emission of 0.0186 grams and an accuracy of 0.487.\\'}', metadata={})]", "category": "generic"}
{"question_id": 862, "text": " We want to enhance our search function by improving the ranking of search results.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Table Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Table Question Answering\\', \\'api_name\\': \\'google/tapas-small-finetuned-sqa\\', \\'api_call\\': \"pipeline(\\'table-question-answering\\', model=\\'google/tapas-small-finetuned-sqa\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'msr_sqa\\', \\'accuracy\\': 0.6155}, \\'description\\': \\'TAPAS small model fine-tuned on Sequential Question Answering (SQA). It uses relative position embeddings (i.e. resetting the position index at every cell of the table).\\'}', metadata={})]", "category": "generic"}
{"question_id": 863, "text": " We are developing an AI chatbot to interact with users. We need the bot to recognize the user's emotions based on their text input.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'hyunwoongko/blenderbot-9B\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'hyunwoongko/blenderbot-9B\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'Input a message to start chatting with hyunwoongko/blenderbot-9B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high-performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi-turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 864, "text": " Please help me extract the names of people, organizations, and locations mentioned in the given text.\\n###Input: Hello, my name is John Doe, and I work at Microsoft. Tomorrow, I'll be going to a conference in San Francisco.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Text Classification\\', \\'api_name\\': \\'joeddav/distilbert-base-uncased-go-emotions-student\\', \\'api_call\\': \"pipeline(\\'text-classification\\', model=\\'joeddav/distilbert-base-uncased-go-emotions-student\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': [\\'transformers\\', \\'torch\\', \\'tensorflow\\'], \\'example_code\\': \"from transformers import pipeline\\\\nnlp = pipeline(\\'text-classification\\', model=\\'joeddav/distilbert-base-uncased-go-emotions-student\\')\\\\nresult = nlp(\\'I am so happy today!\\')\", \\'performance\\': {\\'dataset\\': \\'go_emotions\\'}, \\'description\\': \\'This model is distilled from the zero-shot classification pipeline on the unlabeled GoEmotions dataset. It is primarily intended as a demo of how an expensive NLI-based zero-shot model can be distilled to a more efficient student, allowing a classifier to be trained with only unlabeled data.\\'}', metadata={})]", "category": "generic"}
{"question_id": 865, "text": " The school wants a tool to teach foreign students Chinese grammar. They want you to develop a part-of-speech tagging system to detect the words' grammatical roles.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Token Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Part-of-speech tagging\\', \\'api_name\\': \\'ckiplab/bert-base-chinese-pos\\', \\'api_call\\': \"AutoModel.from_pretrained(\\'ckiplab/bert-base-chinese-pos\\')\", \\'api_arguments\\': {\\'tokenizer\\': \"BertTokenizerFast.from_pretrained(\\'bert-base-chinese\\')\"}, \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import (\\\\n  BertTokenizerFast,\\\\n  AutoModel,\\\\n)\\\\ntokenizer = BertTokenizerFast.from_pretrained(\\'bert-base-chinese\\')\\\\nmodel = AutoModel.from_pretrained(\\'ckiplab/bert-base-chinese-pos\\')\", \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'This project provides traditional Chinese transformers models (including ALBERT, BERT, GPT2) and NLP tools (including word segmentation, part-of-speech tagging, named entity recognition).\\'}', metadata={})]", "category": "generic"}
{"question_id": 866, "text": " Our company collects data on the salesperson performance in different regions for each month. We want to use the most suitable API from our subscribed_huggingface.co to analyze that in specific table format and then based on provided question, answer accordingly.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Table Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Table Question Answering\\', \\'api_name\\': \\'google/tapas-small-finetuned-sqa\\', \\'api_call\\': \"pipeline(\\'table-question-answering\\', model=\\'google/tapas-small-finetuned-sqa\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'msr_sqa\\', \\'accuracy\\': 0.6155}, \\'description\\': \\'TAPAS small model fine-tuned on Sequential Question Answering (SQA). It uses relative position embeddings (i.e. resetting the position index at every cell of the table).\\'}', metadata={})]", "category": "generic"}
{"question_id": 867, "text": " Assist me in finding the accurate information in a table related to the Korean stock market.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Table Question Answering\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'dsba-lab/koreapas-finetuned-korwikitq\\', \\'api_call\\': \"pipeline(\\'table-question-answering\\', model=\\'dsba-lab/koreapas-finetuned-korwikitq\\')\", \\'api_arguments\\': {}, \\'python_environment_requirements\\': {\\'transformers\\': \\'>=4.0.0\\'}, \\'example_code\\': \"from transformers import pipeline; table_qa = pipeline(\\'table-question-answering\\', model=\\'dsba-lab/koreapas-finetuned-korwikitq\\'); table_qa(table=table, query=\\'\uc9c8\ubb38\\')\", \\'performance\\': {\\'dataset\\': \\'korwikitq\\', \\'accuracy\\': None}, \\'description\\': \\'A Korean Table Question Answering model finetuned on the korwikitq dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 868, "text": " My company has a large data table of employees, containing their names, titles, departments, and hire dates. We need a tool that can find all employees with the title of \\\"Software Engineer\\\" hired in 2020.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Table Question Answering\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Table Question Answering\\', \\'api_name\\': \\'google/tapas-small-finetuned-sqa\\', \\'api_call\\': \"pipeline(\\'table-question-answering\\', model=\\'google/tapas-small-finetuned-sqa\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'msr_sqa\\', \\'accuracy\\': 0.6155}, \\'description\\': \\'TAPAS small model fine-tuned on Sequential Question Answering (SQA). It uses relative position embeddings (i.e. resetting the position index at every cell of the table).\\'}', metadata={})]", "category": "generic"}
{"question_id": 869, "text": " We received a business document in French. We need to extract some specific information from it. \\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'moussaKam/barthez-orangesum-abstract\\', \\'api_call\\': \"BarthezModel.from_pretrained(\\'moussaKam/barthez-orangesum-abstract\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'orangeSum\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'Barthez model finetuned on orangeSum for abstract generation in French language\\'}', metadata={})]", "category": "generic"}
{"question_id": 870, "text": " We have a coffee shop with different types of coffee on the menu. Determine the caffeine levels in each cup of coffee.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Joblib\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'pachi107/autotrain-in-class-test-1780161764\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data.csv\\'], \\'python_environment_requirements\\': [\\'json\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(data.csv)\\\\ndata = data[features]\\\\ndata.columns = [feat_ + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'Validation Metrics\\', \\'accuracy\\': 0.974}, \\'description\\': \\'A binary classification model for predicting CO2 emissions based on tabular data. Trained using AutoTrain with a model ID of 1780161764.\\'}', metadata={})]", "category": "generic"}
{"question_id": 871, "text": " You are the head of the QA department and you want to create an application that extracts answers from large product manuals. Explain how you can use this API for the project.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 872, "text": " I want to have a personal assistant app that can answer questions from a given text.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Text2Text Generation\\', \\'api_name\\': \\'castorini/doc2query-t5-base-msmarco\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'castorini/doc2query-t5-base-msmarco\\')\", \\'api_arguments\\': \\'text, max_length\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'MS MARCO\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'A T5 model trained on the MS MARCO dataset for generating queries from documents.\\'}', metadata={})]", "category": "generic"}
{"question_id": 873, "text": " I want to build a tool that helps me answer questions about specific information in a Korean newspaper article.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Table Question Answering\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'dsba-lab/koreapas-finetuned-korwikitq\\', \\'api_call\\': \"pipeline(\\'table-question-answering\\', model=\\'dsba-lab/koreapas-finetuned-korwikitq\\')\", \\'api_arguments\\': {}, \\'python_environment_requirements\\': {\\'transformers\\': \\'>=4.0.0\\'}, \\'example_code\\': \"from transformers import pipeline; table_qa = pipeline(\\'table-question-answering\\', model=\\'dsba-lab/koreapas-finetuned-korwikitq\\'); table_qa(table=table, query=\\'\uc9c8\ubb38\\')\", \\'performance\\': {\\'dataset\\': \\'korwikitq\\', \\'accuracy\\': None}, \\'description\\': \\'A Korean Table Question Answering model finetuned on the korwikitq dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 874, "text": " Develop a tool to help our team members find answers to essential questions from a long document.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Multimodal Document Question Answer\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Document Question Answering\\', \\'api_name\\': \\'CQI_Visual_Question_Awnser_PT_v0\\', \\'api_call\\': \"pipeline(\\'question-answering\\', model=LayoutLMForQuestionAnswering.from_pretrained(\\'microsoft/layoutlm-base-uncased\\'))\", \\'api_arguments\\': [\\'url\\', \\'question\\'], \\'python_environment_requirements\\': [\\'PIL\\', \\'pytesseract\\', \\'PyTorch\\', \\'transformers\\'], \\'example_code\\': [\"nlp(\\'https://templates.invoicehome.com/invoice-template-us-neat-750px.png\\', \\'What is the invoice number?\\')\", \"nlp(\\'https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg\\', \\'What is the purchase amount?\\')\", \"nlp(\\'https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png\\', \\'What are the 2020 net sales?\\')\"], \\'performance\\': {\\'dataset\\': [{\\'accuracy\\': 0.9943977}, {\\'accuracy\\': 0.9912159}, {\\'accuracy\\': 0.59147286}]}, \\'description\\': \\'A model for visual question answering in Portuguese and English, capable of processing PDFs and images to extract information and answer questions.\\'}', metadata={})]", "category": "generic"}
{"question_id": 875, "text": " There is a news article stating, \\\"Angela Merkel ist eine Politikerin in Deutschland und Vorsitzende der CDU.\\\" We need to determine which category this article should fall under.\\n###Input: Angela Merkel ist eine Politikerin in Deutschland und Vorsitzende der CDU.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Zero-Shot Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Zero-Shot Classification\\', \\'api_name\\': \\'MoritzLaurer/mDeBERTa-v3-base-xnli-multilingual-nli-2mil7\\', \\'api_call\\': \"AutoModelForSequenceClassification.from_pretrained(\\'MoritzLaurer/DeBERTa-v3-xsmall-mnli-fever-anli-ling-binary\\')\", \\'api_arguments\\': {\\'sequence_to_classify\\': \\'Angela Merkel ist eine Politikerin in Deutschland und Vorsitzende der CDU\\', \\'candidate_labels\\': [\\'politics\\', \\'economy\\', \\'entertainment\\', \\'environment\\'], \\'multi_label\\': False}, \\'python_environment_requirements\\': [\\'transformers==4.13\\'], \\'example_code\\': \\'from transformers import pipeline\\\\nclassifier = pipeline(zero-shot-classification, model=MoritzLaurer/mDeBERTa-v3-base-mnli-xnli)\\\\nsequence_to_classify = Angela Merkel ist eine Politikerin in Deutschland und Vorsitzende der CDU\\\\ncandidate_labels = [politics, economy, entertainment, environment]\\\\noutput = classifier(sequence_to_classify, candidate_labels, multi_label=False)\\\\nprint(output)\\', \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'MultiNLI-matched\\', \\'accuracy\\': 0.857}, {\\'name\\': \\'MultiNLI-mismatched\\', \\'accuracy\\': 0.856}, {\\'name\\': \\'ANLI-all\\', \\'accuracy\\': 0.537}, {\\'name\\': \\'ANLI-r3\\', \\'accuracy\\': 0.497}, {\\'name\\': \\'WANLI\\', \\'accuracy\\': 0.732}, {\\'name\\': \\'LingNLI\\', \\'accuracy\\': 0.788}, {\\'name\\': \\'fever-nli\\', \\'accuracy\\': 0.761}]}, \\'description\\': \\'This multilingual model can perform natural language inference (NLI) on 100 languages and is therefore also suitable for multilingual zero-shot classification. The underlying mDeBERTa-v3-base model was pre-trained by Microsoft on the CC100 multilingual dataset with 100 languages. The model was then fine-tuned on the XNLI dataset and on the multilingual-NLI-26lang-2mil7 dataset. Both datasets contain more than 2.7 million hypothesis-premise pairs in 27 languages spoken by more than 4 billion people.\\'}', metadata={})]", "category": "generic"}
{"question_id": 876, "text": " We have generated user reviews for movies. We need to check user opinions about the movie 'Inception' based on their reviews.\\n###Input: The movie 'Inception' is an exceptional piece of cinematic art. The storyline is thought-provoking and keeps the audience engaged till the end. The special effects are breathtaking and complement the plot perfectly.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Classification\\', \\'api_name\\': \\'lvwerra/distilbert-imdb\\', \\'api_call\\': \"pipeline(\\'sentiment-analysis\\', model=\\'lvwerra/distilbert-imdb\\')\", \\'api_arguments\\': [], \\'python_environment_requirements\\': [\\'transformers\\', \\'pytorch\\'], \\'example_code\\': \"classifier(\\'I love this movie!\\')\", \\'performance\\': {\\'dataset\\': \\'imdb\\', \\'accuracy\\': 0.928}, \\'description\\': \\'This model is a fine-tuned version of distilbert-base-uncased on the imdb dataset. It is used for sentiment analysis on movie reviews and achieves an accuracy of 0.928 on the evaluation set.\\'}', metadata={})]", "category": "generic"}
{"question_id": 877, "text": " In our new app, we are building a feature that recommends books in different languages. To do this, first, we need to translate the book title and details from English to French. Help us to decide the best translation model to use here.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Audio-to-Audio\\', \\'framework\\': \\'Fairseq\\', \\'functionality\\': \\'speech-to-speech-translation\\', \\'api_name\\': \\'facebook/textless_sm_en_fr\\', \\'api_call\\': \"load_model_ensemble_and_task_from_hf_hub(\\'facebook/textless_sm_en_fr\\')\", \\'api_arguments\\': [\\'input_file\\'], \\'python_environment_requirements\\': [\\'huggingface_hub\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'This model is a speech-to-speech translation model trained by Facebook. It is designed for translating English speech to French speech.\\'}', metadata={})]", "category": "generic"}
{"question_id": 878, "text": " Our multinational company is dealing with a French client. Please help us communicate in French by translating an English sentence into French.\\n###Input: \\\"Hello, how are you?\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Translation\\', \\'api_name\\': \\'Helsinki-NLP/opus-mt-en-fr\\', \\'api_call\\': \"translate(\\'input_text\\', model=\\'Helsinki-NLP/opus-mt-en-fr\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'opus\\', \\'accuracy\\': {\\'BLEU\\': {\\'newsdiscussdev2015-enfr.en.fr\\': 33.8, \\'newsdiscusstest2015-enfr.en.fr\\': 40.0, \\'newssyscomb2009.en.fr\\': 29.8, \\'news-test2008.en.fr\\': 27.5, \\'newstest2009.en.fr\\': 29.4, \\'newstest2010.en.fr\\': 32.7, \\'newstest2011.en.fr\\': 34.3, \\'newstest2012.en.fr\\': 31.8, \\'newstest2013.en.fr\\': 33.2, \\'Tatoeba.en.fr\\': 50.5}}}, \\'description\\': \\'Helsinki-NLP/opus-mt-en-fr is a translation model that translates English text to French using the Hugging Face Transformers library. It is based on the OPUS dataset and uses a transformer-align architecture with normalization and SentencePiece pre-processing.\\'}', metadata={})]", "category": "generic"}
{"question_id": 879, "text": " We want to communicate product information to online customers. Translate the information from English to French.\\n###Input: \\\"Introducing the new eco-friendly water bottle made of high-quality stainless steel with double-wall insulation to keep your drinks cool for 24 hours or hot for 12 hours.\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Translation\\', \\'api_name\\': \\'Helsinki-NLP/opus-mt-en-fr\\', \\'api_call\\': \"translate(\\'input_text\\', model=\\'Helsinki-NLP/opus-mt-en-fr\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'opus\\', \\'accuracy\\': {\\'BLEU\\': {\\'newsdiscussdev2015-enfr.en.fr\\': 33.8, \\'newsdiscusstest2015-enfr.en.fr\\': 40.0, \\'newssyscomb2009.en.fr\\': 29.8, \\'news-test2008.en.fr\\': 27.5, \\'newstest2009.en.fr\\': 29.4, \\'newstest2010.en.fr\\': 32.7, \\'newstest2011.en.fr\\': 34.3, \\'newstest2012.en.fr\\': 31.8, \\'newstest2013.en.fr\\': 33.2, \\'Tatoeba.en.fr\\': 50.5}}}, \\'description\\': \\'Helsinki-NLP/opus-mt-en-fr is a translation model that translates English text to French using the Hugging Face Transformers library. It is based on the OPUS dataset and uses a transformer-align architecture with normalization and SentencePiece pre-processing.\\'}', metadata={})]", "category": "generic"}
{"question_id": 880, "text": " Our Customer is a Swedish travel magazine with an English version. Translate this Swedish text to English for them: \\\"Stockholm \\u00e4r Sveriges huvudstad och st\\u00f6rsta stad. Den har en rik historia och erbjuder m\\u00e5nga kulturella och historiska sev\\u00e4rdheter.\\\"\\n###Input: \\\"Stockholm \\u00e4r Sveriges huvudstad och st\\u00f6rsta stad. Den har en rik historia och erbjuder m\\u00e5nga kulturella och historiska sev\\u00e4rdheter.\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Translation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Translation\\', \\'api_name\\': \\'Helsinki-NLP/opus-mt-es-en\\', \\'api_call\\': \"pipeline(\\'translation_es_to_en\\', model=\\'Helsinki-NLP/opus-mt-es-en\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"from transformers import pipeline\\\\ntranslation = pipeline(\\'translation_es_to_en\\', model=\\'Helsinki-NLP/opus-mt-es-en\\')(\\'Hola, \u00bfc\u00f3mo est\u00e1s?\\')\", \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'newssyscomb2009-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 30.6, \\'chr-F\\': 0.57}}, {\\'name\\': \\'news-test2008-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 27.9, \\'chr-F\\': 0.553}}, {\\'name\\': \\'newstest2009-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 30.4, \\'chr-F\\': 0.572}}, {\\'name\\': \\'newstest2010-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 36.1, \\'chr-F\\': 0.614}}, {\\'name\\': \\'newstest2011-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 34.2, \\'chr-F\\': 0.599}}, {\\'name\\': \\'newstest2012-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 37.9, \\'chr-F\\': 0.624}}, {\\'name\\': \\'newstest2013-spaeng.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 35.3, \\'chr-F\\': 0.609}}, {\\'name\\': \\'Tatoeba-test.spa.eng\\', \\'accuracy\\': {\\'BLEU\\': 59.6, \\'chr-F\\': 0.739}}]}, \\'description\\': \\'Helsinki-NLP/opus-mt-es-en is a machine translation model trained to translate from Spanish to English using the Hugging Face Transformers library. The model is based on the Marian framework and was trained on the OPUS dataset.\\'}', metadata={})]", "category": "generic"}
{"question_id": 881, "text": " I have written a small story in English about the adventures of a superhero who saves the day from evil villains. I would like to translate it into French so that my friends in France can read it too.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Abstractive Text Summarization\\', \\'api_name\\': \\'plguillou/t5-base-fr-sum-cnndm\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'plguillou/t5-base-fr-sum-cnndm\\')\", \\'api_arguments\\': {\\'input_text\\': \\'summarize: ARTICLE\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'from transformers import T5Tokenizer, T5ForConditionalGeneration\\'}, \\'example_code\\': \\'tokenizer = T5Tokenizer.from_pretrained(plguillou/t5-base-fr-sum-cnndm)\\\\nmodel = T5ForConditionalGeneration.from_pretrained(plguillou/t5-base-fr-sum-cnndm)\\', \\'performance\\': {\\'dataset\\': \\'cnn_dailymail\\', \\'ROUGE-1\\': 44.5252, \\'ROUGE-2\\': 22.652, \\'ROUGE-L\\': 29.8866}, \\'description\\': \\'This model is a T5 Transformers model (JDBN/t5-base-fr-qg-fquad) that was fine-tuned in French for abstractive text summarization.\\'}', metadata={})]", "category": "generic"}
{"question_id": 882, "text": " Create an application that reads news articles and provides a brief summary of the article contents.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Summarization\\', \\'api_name\\': \\'it5-base-news-summarization\\', \\'api_call\\': \"pipeline(\\'summarization\\', model=\\'it5/it5-base-news-summarization\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'newsum(Dal 31 maggio \u00e8 infine partita la piattaforma ITsART, a pi\u00f9 di un anno da quando \u2013 durante il primo lockdown \u2013 il ministro della Cultura Dario Franceschini ne aveva parlato come di \u00abuna sorta di Netflix della cultura\u00bb, pensata per \u00aboffrire a tutto il mondo la cultura italiana a pagamento\u00bb. \u00c8 presto per dare giudizi definitivi sulla piattaforma, e di certo sar\u00e0 difficile farlo anche pi\u00f9 avanti senza numeri precisi. Al momento, l\u2019unica cosa che si pu\u00f2 fare \u00e8 guardare com\u2019\u00e8 fatto il sito, contare quanti contenuti ci sono (circa 700 \u201ctitoli\u201d, tra film, documentari, spettacoli teatrali e musicali e altri eventi) e provare a dare un giudizio sul loro valore e sulla loro variet\u00e0. Intanto, una cosa notata da pi\u00f9 parti \u00e8 che diversi contenuti di ITsART sono a pagamento sulla piattaforma sebbene altrove, per esempio su RaiPlay, siano invece disponibili gratuitamente.)\\', \\'performance\\': {\\'dataset\\': \\'NewsSum-IT\\', \\'accuracy\\': {\\'Rouge1\\': 0.339, \\'Rouge2\\': 0.16, \\'RougeL\\': 0.263}}, \\'description\\': \\'IT5 Base model fine-tuned on news summarization on the Fanpage and Il Post corpora for Italian Language Understanding and Generation.\\'}', metadata={})]", "category": "generic"}
{"question_id": 883, "text": " We are building a platform for developers and want to provide automatic code documentation generation for Python functions.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Summarization\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Code Documentation Generation\\', \\'api_name\\': \\'code_trans_t5_base_code_documentation_generation_python\\', \\'api_call\\': \"AutoModelWithLMHead.from_pretrained(\\'SEBIS/code_trans_t5_base_code_documentation_generation_python\\')\", \\'api_arguments\\': [\\'tokenized_code\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'from transformers import AutoTokenizer, AutoModelWithLMHead, SummarizationPipeline\\\\npipeline = SummarizationPipeline(\\\\n model=AutoModelWithLMHead.from_pretrained(SEBIS/code_trans_t5_base_code_documentation_generation_python),\\\\n tokenizer=AutoTokenizer.from_pretrained(SEBIS/code_trans_t5_base_code_documentation_generation_python, skip_special_tokens=True),\\\\n device=0\\\\n)\\\\ntokenized_code = def e ( message , exit_code = None ) : print_log ( message , YELLOW , BOLD ) if exit_code is not None : sys . exit ( exit_code )\\\\npipeline([tokenized_code])\\', \\'performance\\': {\\'dataset\\': \\'CodeSearchNet Corpus python dataset\\', \\'accuracy\\': \\'20.26 BLEU score\\'}, \\'description\\': \\'This CodeTrans model is based on the t5-base model and is trained on tokenized python code functions. It can be used to generate descriptions for python functions or be fine-tuned on other python code tasks. The model works best with tokenized python functions but can also be used on unparsed and untokenized python code.\\'}', metadata={})]", "category": "generic"}
{"question_id": 884, "text": " In order to engage our users and keep them interested in our platform, we require a conversational chatbot that discusses a wide range of topics.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'hyunwoongko/blenderbot-9B\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'hyunwoongko/blenderbot-9B\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'Input a message to start chatting with hyunwoongko/blenderbot-9B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high-performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi-turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 885, "text": " Use this API to get a suggestion on how to respond to a customer's complaint about the late delivery of their package.\\n###Input: {\\\"instruction\\\": \\\"How can I respond to a customer complaint about late delivery?\\\", \\\"knowledge\\\": \\\"The courier had external delays due to bad winter weather.\\\", \\\"dialog\\\": [\\\"Customer: My package is late. What's going on?\\\", \\\"Support: I apologize for the inconvenience. I'll check what's happening with the package and get back to you.\\\"]}\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Conversational\\', \\'api_name\\': \\'Pi3141/DialoGPT-medium-elon-3\\', \\'api_call\\': \"pipeline(\\'text-generation\\', model=\\'Pi3141/DialoGPT-medium-elon-3\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'Input a message to start chatting with Pi3141/DialoGPT-medium-elon-3.\\', \\'performance\\': {\\'dataset\\': \\'Twitter tweets by Elon Musk\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'DialoGPT model that talks like Elon Musk, trained on Twitter tweets by Elon Musk. This model will spew meaningless shit about 40% of the time. Trained on 8 epochs. But with a larger dataset this time. The AI can now use more emojis, I think.\\'}', metadata={})]", "category": "generic"}
{"question_id": 886, "text": " A game studio is now creating a story for their new action game, they need a hint for creating the setting of the game.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Text2Text Generation\\', \\'api_name\\': \\'castorini/doc2query-t5-base-msmarco\\', \\'api_call\\': \"T5ForConditionalGeneration.from_pretrained(\\'castorini/doc2query-t5-base-msmarco\\')\", \\'api_arguments\\': \\'text, max_length\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'MS MARCO\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'A T5 model trained on the MS MARCO dataset for generating queries from documents.\\'}', metadata={})]", "category": "generic"}
{"question_id": 887, "text": " We are working on an AI chatbot for customer support and we need our chatbot to generate human-like responses to customers' questions.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'hyunwoongko/blenderbot-9B\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'hyunwoongko/blenderbot-9B\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'Input a message to start chatting with hyunwoongko/blenderbot-9B.\\', \\'performance\\': {\\'dataset\\': \\'blended_skill_talk\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high-performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi-turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.\\'}', metadata={})]", "category": "generic"}
{"question_id": 888, "text": " Translate the following English sentence to German: \\\"I have a doctor's appointment tomorrow morning.\\\"\\n###Input: \\\"I have a doctor's appointment tomorrow morning.\\\"\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentence Correction\\', \\'api_name\\': \\'flexudy/t5-base-multi-sentence-doctor\\', \\'api_call\\': \"AutoModelWithLMHead.from_pretrained(\\'flexudy/t5-base-multi-sentence-doctor\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'from transformers import AutoTokenizer, AutoModelWithLMHead\\\\ntokenizer = AutoTokenizer.from_pretrained(flexudy/t5-base-multi-sentence-doctor)\\\\nmodel = AutoModelWithLMHead.from_pretrained(flexudy/t5-base-multi-sentence-doctor)\\\\ninput_text = repair_sentence: m a medical doct context: {That is my job I a}{or I save lives} </s>\\\\ninput_ids = tokenizer.encode(input_text, return_tensors=pt)\\\\noutputs = model.generate(input_ids, max_length=32, num_beams=1)\\\\nsentence = tokenizer.decode(outputs[0], skip_special_tokens=True, clean_up_tokenization_spaces=True)\\\\nassert sentence == I am a medical doctor.\\', \\'performance\\': {\\'dataset\\': \\'tatoeba\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'Sentence doctor is a T5 model that attempts to correct the errors or mistakes found in sentences. Model works on English, German and French text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 889, "text": " I am struggling with grammar while writing. Thus, I want to create a grammar correction tool for myself.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Grammar Correction\\', \\'api_name\\': \\'vennify/t5-base-grammar-correction\\', \\'api_call\\': \"HappyTextToText(\\'T5\\', \\'vennify/t5-base-grammar-correction\\')\", \\'api_arguments\\': {\\'num_beams\\': 5, \\'min_length\\': 1}, \\'python_environment_requirements\\': {\\'package\\': \\'happytransformer\\', \\'installation\\': \\'pip install happytransformer\\'}, \\'example_code\\': \\'from happytransformer import HappyTextToText, TTSettings\\\\nhappy_tt = HappyTextToText(T5, vennify/t5-base-grammar-correction)\\\\nargs = TTSettings(num_beams=5, min_length=1)\\\\nresult = happy_tt.generate_text(grammar: This sentences has has bads grammar., args=args)\\\\nprint(result.text)\\', \\'performance\\': {\\'dataset\\': \\'jfleg\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'This model generates a revised version of inputted text with the goal of containing fewer grammatical errors. It was trained with Happy Transformer using a dataset called JFLEG.\\'}', metadata={})]", "category": "generic"}
{"question_id": 890, "text": " Our company is cooperating with a German partner. We have some materials in English, but need to translate them into German.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transformers\\', \\'api_name\\': \\'sshleifer/tiny-marian-en-de\\', \\'api_call\\': \"pipeline(\\'translation_en_to_de\\', model=\\'sshleifer/tiny-marian-en-de\\')\", \\'api_arguments\\': \\'\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A tiny English to German translation model using the Marian framework in Hugging Face Transformers.\\'}', metadata={})]", "category": "generic"}
{"question_id": 891, "text": " The publisher has sent us a draft of an article, but some of the words have been masked. We need to identify the masked words.\\n###Input: \\\"<mask> are large, slow-moving reptiles native to the southeastern United States. They are well-adapted to life in <mask>, and they are a common sight in swamps, rivers, and lakes.\\\" \\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Fill-Mask\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Masked Language Modeling and Next Sentence Prediction\\', \\'api_name\\': \\'bert-large-uncased\\', \\'api_call\\': \"pipeline(\\'fill-mask\\', model=\\'bert-large-uncased\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"from transformers import pipeline\\\\nunmasker = pipeline(\\'fill-mask\\', model=\\'bert-large-uncased\\')\\\\nunmasker(Hello I\\'m a [MASK] model.)\", \\'performance\\': {\\'dataset\\': {\\'SQUAD 1.1 F1/EM\\': \\'91.0/84.3\\', \\'Multi NLI Accuracy\\': \\'86.05\\'}}, \\'description\\': \\'BERT large model (uncased) is a transformer model pretrained on a large corpus of English data using a masked language modeling (MLM) objective. It has 24 layers, 1024 hidden dimensions, 16 attention heads, and 336M parameters. The model is intended to be fine-tuned on a downstream task, such as sequence classification, token classification, or question answering.\\'}', metadata={})]", "category": "generic"}
{"question_id": 892, "text": " Our company is exploring the Chinese market and we need to communicate effectively with our clients. Help me create a generic Chinese response.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Conversational\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Text Generation\\', \\'api_name\\': \\'mywateriswet/ShuanBot\\', \\'api_call\\': \"pipeline(\\'conversational\\', model=\\'mywateriswet/ShuanBot\\')\", \\'api_arguments\\': \\'message\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"response = chatbot(\\'What is your name?\\')\", \\'performance\\': {\\'dataset\\': \\'N/A\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'ShuanBot is a conversational chatbot model based on the GPT-2 architecture. It can be used for generating human-like responses in a chat context.\\'}', metadata={})]", "category": "generic"}
{"question_id": 893, "text": " We are an educational platform improving student writing skills. We need a program that can fill in the blanks in sentences.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text2Text Generation\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Sentence Correction\\', \\'api_name\\': \\'flexudy/t5-base-multi-sentence-doctor\\', \\'api_call\\': \"AutoModelWithLMHead.from_pretrained(\\'flexudy/t5-base-multi-sentence-doctor\\')\", \\'api_arguments\\': [\\'input_text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'from transformers import AutoTokenizer, AutoModelWithLMHead\\\\ntokenizer = AutoTokenizer.from_pretrained(flexudy/t5-base-multi-sentence-doctor)\\\\nmodel = AutoModelWithLMHead.from_pretrained(flexudy/t5-base-multi-sentence-doctor)\\\\ninput_text = repair_sentence: m a medical doct context: {That is my job I a}{or I save lives} </s>\\\\ninput_ids = tokenizer.encode(input_text, return_tensors=pt)\\\\noutputs = model.generate(input_ids, max_length=32, num_beams=1)\\\\nsentence = tokenizer.decode(outputs[0], skip_special_tokens=True, clean_up_tokenization_spaces=True)\\\\nassert sentence == I am a medical doctor.\\', \\'performance\\': {\\'dataset\\': \\'tatoeba\\', \\'accuracy\\': \\'Not specified\\'}, \\'description\\': \\'Sentence doctor is a T5 model that attempts to correct the errors or mistakes found in sentences. Model works on English, German and French text.\\'}', metadata={})]", "category": "generic"}
{"question_id": 894, "text": " I am building a plagiarism-detection tool. I need to evaluate the similarity between two sentences.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Sentence Similarity\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Sentence Transformers\\', \\'api_name\\': \\'nikcheerla/nooks-amd-detection-realtime\\', \\'api_call\\': \"SentenceTransformer(\\'nikcheerla/nooks-amd-detection-realtime\\')\", \\'api_arguments\\': [\\'sentences\\'], \\'python_environment_requirements\\': \\'pip install -U sentence-transformers\\', \\'example_code\\': \"from sentence_transformers import SentenceTransformer\\\\nsentences = [This is an example sentence, Each sentence is converted]\\\\nmodel = SentenceTransformer(\\'{MODEL_NAME}\\')\\\\nembeddings = model.encode(sentences)\\\\nprint(embeddings)\", \\'performance\\': {\\'dataset\\': \\'https://seb.sbert.net\\', \\'accuracy\\': \\'Automated evaluation\\'}, \\'description\\': \\'This is a sentence-transformers model that maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search.\\'}', metadata={})]", "category": "generic"}
{"question_id": 895, "text": " We are working on creating an audiobook. Convert this text: \\\"The sun was shining brightly, and the birds were singing sweetly\\\" into speech.\\n###Input: The sun was shining brightly, and the birds were singing sweetly.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'lakahaga/novel_reading_tts\\', \\'api_call\\': \"AutoModelForTTS.from_pretrained(\\'lakahaga/novel_reading_tts\\')\", \\'api_arguments\\': \\'text\\', \\'python_environment_requirements\\': \\'transformers\\', \\'example_code\\': \"inputs = processor(text, return_tensors=\\'pt\\'); generated_audio = model.generate(**inputs);\", \\'performance\\': {\\'dataset\\': \\'novelspeech\\', \\'accuracy\\': None}, \\'description\\': \\'This model was trained by lakahaga using novelspeech recipe in espnet. It is designed for Korean text-to-speech tasks.\\'}', metadata={})]", "category": "generic"}
{"question_id": 896, "text": " A researcher needs information about how to use the Whisper ASR model to transcribe and analyze the sentiment of an audio file.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Transcription and Translation\\', \\'api_name\\': \\'openai/whisper-medium\\', \\'api_call\\': \"WhisperForConditionalGeneration.from_pretrained(\\'openai/whisper-medium\\')\", \\'api_arguments\\': [\\'sample\\', \\'sampling_rate\\', \\'language\\', \\'task\\', \\'skip_special_tokens\\'], \\'python_environment_requirements\\': [\\'transformers\\', \\'datasets\\'], \\'example_code\\': \\'from transformers import WhisperProcessor, WhisperForConditionalGeneration\\\\nfrom datasets import load_dataset\\\\n\\\\nprocessor = WhisperProcessor.from_pretrained(openai/whisper-medium)\\\\nmodel = WhisperForConditionalGeneration.from_pretrained(openai/whisper-medium)\\\\n\\\\nmodel.config.forced_decoder_ids = None\\\\n\\\\nds = load_dataset(hf-internal-testing/librispeech_asr_dummy, clean, split=validation)\\\\nsample = ds[0][audio]\\\\ninput_features = processor(sample[array], sampling_rate=sample[sampling_rate], return_tensors=pt).input_features\\\\n\\\\npredicted_ids = model.generate(input_features)\\\\ntranscription = processor.batch_decode(predicted_ids, skip_special_tokens=True)\\', \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'LibriSpeech (clean)\\', \\'accuracy\\': 2.9}, {\\'name\\': \\'LibriSpeech (other)\\', \\'accuracy\\': 5.9}, {\\'name\\': \\'Common Voice 11.0\\', \\'accuracy\\': 53.87}]}, \\'description\\': \\'Whisper is a pre-trained model for automatic speech recognition (ASR) and speech translation. Trained on 680k hours of labelled data, Whisper models demonstrate a strong ability to generalise to many datasets and domains without the need for fine-tuning. It is a Transformer-based encoder-decoder model and was trained on either English-only data or multilingual data.\\'}', metadata={})]", "category": "generic"}
{"question_id": 897, "text": " I recently interviewed a person in Japanese. I need to transcribe the interview in order to find relevant quotes for my article.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Fill-Mask\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Fill-Mask\\', \\'api_name\\': \\'cl-tohoku/bert-base-japanese\\', \\'api_call\\': \"AutoModelForMaskedLM.from_pretrained(\\'cl-tohoku/bert-base-japanese\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"fill_mask(\\'[MASK]\\')\", \\'performance\\': {\\'dataset\\': \\'wikipedia\\', \\'accuracy\\': \\'N/A\\'}, \\'description\\': \\'This is a BERT model pretrained on texts in the Japanese language. This version of the model processes input texts with word-level tokenization based on the IPA dictionary, followed by the WordPiece subword tokenization.\\'}', metadata={})]", "category": "generic"}
{"question_id": 898, "text": " Our startup produces a podcast editing software. We need to convert speech to text automatically for transcription purposes.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'ESPnet\\', \\'functionality\\': \\'Text-to-Speech\\', \\'api_name\\': \\'SYSPIN/Telugu_Male_TTS\\', \\'api_call\\': \"pipeline(\\'text-to-speech\\', model=\\'SYSPIN/Telugu_Male_TTS\\')\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \\'\\', \\'performance\\': {\\'dataset\\': \\'\\', \\'accuracy\\': \\'\\'}, \\'description\\': \\'A Telugu Male Text-to-Speech model using the ESPnet framework, provided by Hugging Face.\\'}', metadata={})]", "category": "generic"}
{"question_id": 899, "text": " Lately, the quality of our audio has degraded due to background noise, and our robot needs to process speech to understand voice commands. Please help.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'CTranslate2\\', \\'functionality\\': \\'Automatic Speech Recognition\\', \\'api_name\\': \\'guillaumekln/faster-whisper-large-v2\\', \\'api_call\\': \"WhisperModel(\\'large-v2\\')\", \\'api_arguments\\': [\\'audio.mp3\\'], \\'python_environment_requirements\\': [\\'faster_whisper\\'], \\'example_code\\': \\'from faster_whisper import WhisperModel\\\\nmodel = WhisperModel(large-v2)\\\\nsegments, info = model.transcribe(audio.mp3)\\\\nfor segment in segments:\\\\n print([%.2fs -&gt; %.2fs] %s % (segment.start, segment.end, segment.text))\\', \\'performance\\': {\\'dataset\\': \\'99 languages\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Whisper large-v2 model for CTranslate2. This model can be used in CTranslate2 or projets based on CTranslate2 such as faster-whisper.\\'}', metadata={})]", "category": "generic"}
{"question_id": 900, "text": " Create a recommendation engine for a podcast platform that enhances the audio quality of low-quality recordings before recommending it to users seeking high-quality content.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Voice Activity Detection, Speech-to-Noise Ratio, and C50 Room Acoustics Estimation\\', \\'api_name\\': \\'pyannote/brouhaha\\', \\'api_call\\': \"Model.from_pretrained(\\'pyannote/brouhaha\\', use_auth_token=\\'ACCESS_TOKEN_GOES_HERE\\')\", \\'api_arguments\\': [\\'audio.wav\\'], \\'python_environment_requirements\\': [\\'pyannote-audio\\', \\'brouhaha-vad\\'], \\'example_code\\': [\\'from pyannote.audio import Model\\', \\'model = Model.from_pretrained(pyannote/brouhaha, use_auth_token=ACCESS_TOKEN_GOES_HERE)\\', \\'from pyannote.audio import Inference\\', \\'inference = Inference(model)\\', \\'output = inference(audio.wav)\\', \\'for frame, (vad, snr, c50) in output:\\', \\'  t = frame.middle\\', \\'  print(f{t:8.3f} vad={100*vad:.0f}% snr={snr:.0f} c50={c50:.0f})\\'], \\'performance\\': {\\'dataset\\': \\'LibriSpeech, AudioSet, EchoThief, MIT-Acoustical-Reverberation-Scene\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Brouhaha is a joint voice activity detection, speech-to-noise ratio, and C50 room acoustics estimation model. It is based on the PyTorch framework and uses the pyannote.audio library.\\'}', metadata={})]", "category": "generic"}
{"question_id": 901, "text": " Create a system that translates and synthesizes speech from one language to another using the given model.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Text-to-Speech\\', \\'framework\\': \\'Fairseq\\', \\'functionality\\': \\'Speech-to-speech translation\\', \\'api_name\\': \\'facebook/unit_hifigan_mhubert_vp_en_es_fr_it3_400k_layer11_km1000_lj_dur\\', \\'api_call\\': \"load_model_ensemble_and_task_from_hf_hub(\\'facebook/unit_hifigan_mhubert_vp_en_es_fr_it3_400k_layer11_km1000_lj_dur\\')\", \\'api_arguments\\': {\\'audio\\': \\'16000Hz mono channel audio\\'}, \\'python_environment_requirements\\': [\\'fairseq\\', \\'hub_utils\\', \\'huggingface_hub\\', \\'IPython.display\\', \\'torchaudio\\'], \\'example_code\\': [\\'import json\\', \\'import os\\', \\'from pathlib import Path\\', \\'import IPython.display as ipd\\', \\'from fairseq import hub_utils\\', \\'from fairseq.checkpoint_utils import load_model_ensemble_and_task_from_hf_hub\\', \\'from fairseq.models.speech_to_text.hub_interface import S2THubInterface\\', \\'from fairseq.models.text_to_speech import CodeHiFiGANVocoder\\', \\'from fairseq.models.text_to_speech.hub_interface import VocoderHubInterface\\', \\'from huggingface_hub import snapshot_download\\', \\'import torchaudio\\', \\'cache_dir = os.getenv(HUGGINGFACE_HUB_CACHE)\\', \\'models, cfg, task = load_model_ensemble_and_task_from_hf_hub(\\', \\'facebook/xm_transformer_s2ut_800m-es-en-st-asr-bt_h1_2022,\\', \\'arg_overrides={config_yaml: config.yaml, task: speech_to_text},\\', \\'cache_dir=cache_dir,\\', \\')\\', \\'model = models[0].cpu()\\', \\'cfg[task].cpu = True\\', \\'generator = task.build_generator([model], cfg)\\', \\'# requires 16000Hz mono channel audio\\', \\'audio, _ = torchaudio.load(/Users/lpw/git/api-inference-community/docker_images/fairseq/tests/samples/sample2.flac)\\', \\'sample = S2THubInterface.get_model_input(task, audio)\\', \\'unit = S2THubInterface.get_prediction(task, model, generator, sample)\\', \\'library_name = fairseq\\', \\'cache_dir = (\\', \\' cache_dir or (Path.home() / .cache / library_name).as_posix()\\', \\')\\', \\'cache_dir = snapshot_download(\\', \\' ffacebook/unit_hifigan_mhubert_vp_en_es_fr_it3_400k_layer11_km1000_lj_dur, cache_dir=cache_dir, library_name=library_name\\', \\')\\', \\'x = hub_utils.from_pretrained(\\', \\' cache_dir,\\', \\' model.pt,\\', \\' .,\\', \\' archive_map=CodeHiFiGANVocoder.hub_models(),\\', \\' config_yaml=config.json,\\', \\' fp16=False,\\', \\' is_vocoder=True,\\', \\')\\', \"with open(f{x[\\'args\\'][\\'data\\']}/config.json) as f:\", \\' vocoder_cfg = json.load(f)\\', \\'assert (\\', \\' len(x[args][model_path]) == 1\\', \\'), Too many vocoder models in the input\\', \\'vocoder = CodeHiFiGANVocoder(x[args][model_path][0], vocoder_cfg)\\', \\'tts_model = VocoderHubInterface(vocoder_cfg, vocoder)\\', \\'tts_sample = tts_model.get_model_input(unit)\\', \\'wav, sr = tts_model.get_prediction(tts_sample)\\', \\'ipd.Audio(wav, rate=sr)\\'], \\'performance\\': {\\'dataset\\': \\'covost2\\', \\'accuracy\\': None}, \\'description\\': \\'Speech-to-speech translation model from fairseq S2UT (paper/code) for Spanish-English. Trained on mTEDx, CoVoST 2, Europarl-ST, and VoxPopuli.\\'}', metadata={})]", "category": "generic"}
{"question_id": 902, "text": " I have a single channel audio recording containing the voices of two speakers very close, and I would like to separate the voices in this audio recording.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'pyannote.audio\\', \\'functionality\\': \\'Speaker diarization\\', \\'api_name\\': \\'johnislarry/cloned-pyannote-speaker-diarization-endpoint\\', \\'api_call\\': \"Pipeline.from_pretrained(\\'pyannote/speaker-diarization@2.1\\',use_auth_token=\\'ACCESS_TOKEN_GOES_HERE\\')\", \\'api_arguments\\': [\\'num_speakers\\', \\'min_speakers\\', \\'max_speakers\\', \\'segmentation_onset\\'], \\'python_environment_requirements\\': \\'pyannote.audio 2.0\\', \\'example_code\\': {\\'load_pipeline\\': \\'from pyannote.audio import Pipeline\\\\npipeline = Pipeline.from_pretrained(pyannote/speaker-diarization@2022.07)\\', \\'apply_pipeline\\': \\'diarization = pipeline(audio.wav)\\', \\'save_output\\': \\'with open(audio.rttm, w) as rttm:\\\\n  diarization.write_rttm(rttm)\\'}, \\'performance\\': {\\'dataset\\': [{\\'name\\': \\'AISHELL-4\\', \\'accuracy\\': {\\'DER%\\': 14.61, \\'FA%\\': 3.31, \\'Miss%\\': 4.35, \\'Conf%\\': 6.95}}, {\\'name\\': \\'AMI Mix-Headset only_words\\', \\'accuracy\\': {\\'DER%\\': 18.21, \\'FA%\\': 3.28, \\'Miss%\\': 11.07, \\'Conf%\\': 3.87}}, {\\'name\\': \\'AMI Array1-01 only_words\\', \\'accuracy\\': {\\'DER%\\': 29.0, \\'FA%\\': 2.71, \\'Miss%\\': 21.61, \\'Conf%\\': 4.68}}, {\\'name\\': \\'CALLHOME Part2\\', \\'accuracy\\': {\\'DER%\\': 30.24, \\'FA%\\': 3.71, \\'Miss%\\': 16.86, \\'Conf%\\': 9.66}}, {\\'name\\': \\'DIHARD 3 Full\\', \\'accuracy\\': {\\'DER%\\': 20.99, \\'FA%\\': 4.25, \\'Miss%\\': 10.74, \\'Conf%\\': 6.0}}, {\\'name\\': \\'REPERE Phase 2\\', \\'accuracy\\': {\\'DER%\\': 12.62, \\'FA%\\': 1.55, \\'Miss%\\': 3.3, \\'Conf%\\': 7.76}}, {\\'name\\': \\'VoxConverse v0.0.2\\', \\'accuracy\\': {\\'DER%\\': 12.76, \\'FA%\\': 3.45, \\'Miss%\\': 3.85, \\'Conf%\\': 5.46}}]}, \\'description\\': \\'This API provides speaker diarization functionality using the pyannote.audio framework. It is capable of processing audio files and outputting speaker diarization results in RTTM format. The API supports providing the number of speakers, minimum and maximum number of speakers, and adjusting the segmentation onset threshold.\\'}', metadata={})]", "category": "generic"}
{"question_id": 903, "text": " We are building an app to improve people's public speaking skills. We want to analyze their emotional speech as feedback.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'emotion\\', \\'api_name\\': \\'bhadresh-savani/distilbert-base-uncased-emotion\\', \\'api_call\\': \"pipeline(\\'text-classification\\', model=\\'bhadresh-savani/distilbert-base-uncased-emotion\\', return_all_scores=True)\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"prediction = classifier(\\'I love using transformers. The best part is wide range of support and its easy to use\\')\", \\'performance\\': {\\'dataset\\': \\'Twitter-Sentiment-Analysis\\', \\'accuracy\\': 0.938}, \\'description\\': \"Distilbert is created with knowledge distillation during the pre-training phase which reduces the size of a BERT model by 40%, while retaining 97% of its language understanding. It\\'s smaller, faster than Bert and any other Bert-based model. Distilbert-base-uncased finetuned on the emotion dataset using HuggingFace Trainer.\"}', metadata={})]", "category": "generic"}
{"question_id": 904, "text": " A psychology company is building a revolutionary means to detect emotions of its clients. Help them create a system to understand emotions from spoken words.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Transformers\\', \\'functionality\\': \\'Emotion Classification\\', \\'api_name\\': \\'j-hartmann/emotion-english-distilroberta-base\\', \\'api_call\\': \"pipeline(\\'text-classification\\', model=\\'j-hartmann/emotion-english-distilroberta-base\\', return_all_scores=True)\", \\'api_arguments\\': {\\'text\\': \\'string\\'}, \\'python_environment_requirements\\': {\\'transformers\\': \\'latest\\'}, \\'example_code\\': \\'from transformers import pipeline\\\\nclassifier = pipeline(text-classification, model=j-hartmann/emotion-english-distilroberta-base, return_all_scores=True)\\\\nclassifier(I love this!)\\', \\'performance\\': {\\'dataset\\': \\'Balanced subset from 6 diverse datasets\\', \\'accuracy\\': \\'66%\\'}, \\'description\\': \"This model classifies emotions in English text data. It predicts Ekman\\'s 6 basic emotions, plus a neutral class: anger, disgust, fear, joy, neutral, sadness, and surprise. The model is a fine-tuned checkpoint of DistilRoBERTa-base.\"}', metadata={})]", "category": "generic"}
{"question_id": 905, "text": " We are a company manufacturing AI-based toothbrushes for children. We want to analyze the emotion of children while they brush their teeth.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Natural Language Processing Text Classification\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'emotion\\', \\'api_name\\': \\'bhadresh-savani/distilbert-base-uncased-emotion\\', \\'api_call\\': \"pipeline(\\'text-classification\\', model=\\'bhadresh-savani/distilbert-base-uncased-emotion\\', return_all_scores=True)\", \\'api_arguments\\': [\\'text\\'], \\'python_environment_requirements\\': [\\'transformers\\'], \\'example_code\\': \"prediction = classifier(\\'I love using transformers. The best part is wide range of support and its easy to use\\')\", \\'performance\\': {\\'dataset\\': \\'Twitter-Sentiment-Analysis\\', \\'accuracy\\': 0.938}, \\'description\\': \"Distilbert is created with knowledge distillation during the pre-training phase which reduces the size of a BERT model by 40%, while retaining 97% of its language understanding. It\\'s smaller, faster than Bert and any other Bert-based model. Distilbert-base-uncased finetuned on the emotion dataset using HuggingFace Trainer.\"}', metadata={})]", "category": "generic"}
{"question_id": 906, "text": " I am a hearing impaired individual who relies on technology to lead a qualitative life. Develop an application that measures noise levels in the environment to help me decide if it's suitable for communication.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Voice Activity Detection\\', \\'framework\\': \\'Hugging Face Transformers\\', \\'functionality\\': \\'Voice Activity Detection, Speech-to-Noise Ratio, and C50 Room Acoustics Estimation\\', \\'api_name\\': \\'pyannote/brouhaha\\', \\'api_call\\': \"Model.from_pretrained(\\'pyannote/brouhaha\\', use_auth_token=\\'ACCESS_TOKEN_GOES_HERE\\')\", \\'api_arguments\\': [\\'audio.wav\\'], \\'python_environment_requirements\\': [\\'pyannote-audio\\', \\'brouhaha-vad\\'], \\'example_code\\': [\\'from pyannote.audio import Model\\', \\'model = Model.from_pretrained(pyannote/brouhaha, use_auth_token=ACCESS_TOKEN_GOES_HERE)\\', \\'from pyannote.audio import Inference\\', \\'inference = Inference(model)\\', \\'output = inference(audio.wav)\\', \\'for frame, (vad, snr, c50) in output:\\', \\'  t = frame.middle\\', \\'  print(f{t:8.3f} vad={100*vad:.0f}% snr={snr:.0f} c50={c50:.0f})\\'], \\'performance\\': {\\'dataset\\': \\'LibriSpeech, AudioSet, EchoThief, MIT-Acoustical-Reverberation-Scene\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Brouhaha is a joint voice activity detection, speech-to-noise ratio, and C50 room acoustics estimation model. It is based on the PyTorch framework and uses the pyannote.audio library.\\'}', metadata={})]", "category": "generic"}
{"question_id": 907, "text": " Can you classify the audio clip to determine whether it is silent or contains speech?\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'CTranslate2\\', \\'functionality\\': \\'Automatic Speech Recognition\\', \\'api_name\\': \\'guillaumekln/faster-whisper-large-v2\\', \\'api_call\\': \"WhisperModel(\\'large-v2\\')\", \\'api_arguments\\': [\\'audio.mp3\\'], \\'python_environment_requirements\\': [\\'faster_whisper\\'], \\'example_code\\': \\'from faster_whisper import WhisperModel\\\\nmodel = WhisperModel(large-v2)\\\\nsegments, info = model.transcribe(audio.mp3)\\\\nfor segment in segments:\\\\n print([%.2fs -&gt; %.2fs] %s % (segment.start, segment.end, segment.text))\\', \\'performance\\': {\\'dataset\\': \\'99 languages\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Whisper large-v2 model for CTranslate2. This model can be used in CTranslate2 or projets based on CTranslate2 such as faster-whisper.\\'}', metadata={})]", "category": "generic"}
{"question_id": 908, "text": " I need to predict digit categories based on some tabular data inputs. I need a pretrained model to use for classification.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Regression\\', \\'framework\\': \\'Hugging Face\\', \\'functionality\\': \\'Carbon Emissions Prediction\\', \\'api_name\\': \\'bibekbehera/autotrain-numeric_prediction-40376105019\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'bibekbehera/autotrain-data-numeric_prediction\\', \\'accuracy\\': {\\'Loss\\': 0.152, \\'R2\\': 0.659, \\'MSE\\': 0.023, \\'MAE\\': 0.062, \\'RMSLE\\': 0.105}}, \\'description\\': \\'A tabular regression model trained with AutoTrain to predict carbon emissions based on input features.\\'}', metadata={})]", "category": "generic"}
{"question_id": 909, "text": " In surveillance operations, transcripts are used to turn the audio feed into chunks such that after processing transcripts there are no speakers talking over each other in the output.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Audio Automatic Speech Recognition\\', \\'framework\\': \\'CTranslate2\\', \\'functionality\\': \\'Automatic Speech Recognition\\', \\'api_name\\': \\'guillaumekln/faster-whisper-large-v2\\', \\'api_call\\': \"WhisperModel(\\'large-v2\\')\", \\'api_arguments\\': [\\'audio.mp3\\'], \\'python_environment_requirements\\': [\\'faster_whisper\\'], \\'example_code\\': \\'from faster_whisper import WhisperModel\\\\nmodel = WhisperModel(large-v2)\\\\nsegments, info = model.transcribe(audio.mp3)\\\\nfor segment in segments:\\\\n print([%.2fs -&gt; %.2fs] %s % (segment.start, segment.end, segment.text))\\', \\'performance\\': {\\'dataset\\': \\'99 languages\\', \\'accuracy\\': \\'Not provided\\'}, \\'description\\': \\'Whisper large-v2 model for CTranslate2. This model can be used in CTranslate2 or projets based on CTranslate2 such as faster-whisper.\\'}', metadata={})]", "category": "generic"}
{"question_id": 910, "text": " I work at GreenTech, a startup that provides eco-friendly solutions, and need to determine if a given set of input data will result in high carbon emissions or not.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Joblib\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'tejas23/autotrain-amx2-1702259728\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data\\'], \\'python_environment_requirements\\': [\\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(\\'data.csv\\')\\\\ndata = data[features]\\\\ndata.columns = [\\'feat_\\' + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'Validation Metrics\\', \\'accuracy\\': 0.831}, \\'description\\': \\'A multi-class classification model trained using AutoTrain to predict CO2 emissions based on tabular data.\\'}', metadata={})]", "category": "generic"}
{"question_id": 911, "text": " Can you figure a way to predict electricity consumption? We need to better allocate resources.\\n \n Use this API documentation for reference: [Document(page_content='{\\'domain\\': \\'Tabular Tabular Classification\\', \\'framework\\': \\'Joblib\\', \\'functionality\\': \\'Carbon Emissions\\', \\'api_name\\': \\'pachi107/autotrain-in-class-test-1780161764\\', \\'api_call\\': \"joblib.load(\\'model.joblib\\')\", \\'api_arguments\\': [\\'data.csv\\'], \\'python_environment_requirements\\': [\\'json\\', \\'joblib\\', \\'pandas\\'], \\'example_code\\': \"import json\\\\nimport joblib\\\\nimport pandas as pd\\\\nmodel = joblib.load(\\'model.joblib\\')\\\\nconfig = json.load(open(\\'config.json\\'))\\\\nfeatures = config[\\'features\\']\\\\ndata = pd.read_csv(data.csv)\\\\ndata = data[features]\\\\ndata.columns = [feat_ + str(col) for col in data.columns]\\\\npredictions = model.predict(data)\", \\'performance\\': {\\'dataset\\': \\'Validation Metrics\\', \\'accuracy\\': 0.974}, \\'description\\': \\'A binary classification model for predicting CO2 emissions based on tabular data. Trained using AutoTrain with a model ID of 1780161764.\\'}', metadata={})]", "category": "generic"}