Datasets:

---

ftopal

/

huggingface-datasets

like 0

[ "region:us" ]

[]

[]

[]

[ "TAGS\n#region-us \n" ]

[ "arxiv:1611.09268", "region:us" ]

[ "1611.09268" ]

[]

[]

[ "TAGS\n#arxiv-1611.09268 #region-us \n" ]

[ "region:us" ]

[]

[]

[]

[ "TAGS\n#region-us \n" ]

[ "multilinguality:monolingual", "multilinguality:multilingual", "language:de", "language:en", "language:fr", "language:ru", "language:zh", "license:cc-by-sa-4.0", "arxiv:2104.06893", "arxiv:2010.02573", "arxiv:2003.04807", "arxiv:2204.08582", "arxiv:2008.09335", "arxiv:2104.07081", "region:us" ]

[ "2104.06893", "2010.02573", "2003.04807", "2204.08582", "2008.09335", "2104.07081" ]

[ "de", "en", "fr", "ru", "zh" ]

[ "# Dataset Card for MTEB Benchmark", "## Dataset Description\n\n- Homepage: URL\n- Repository: URL\n- Paper: soon\n- Leaderboard: URL\n- Point of Contact: nouamane@URL", "### Dataset Summary\n\nMTEB is a heterogeneous benchmark that has been built from diverse tasks:\n* BitextMining: BUCC, Tatoeba\n* Classification: AmazonCounterfactualClassification, AmazonPolarityClassification, AmazonReviewsClassification, Banking77Classification, EmotionClassification, ImdbClassification, MassiveIntentClassification, MassiveScenarioClassification, MTOPDomainClassification, MTOPIntentClassification, ToxicConversationsClassification, TweetSentimentExtractionClassification\n* Clustering: ArxivClusteringP2P, ArxivClusteringS2S, BiorxivClusteringP2P, BiorxivClusteringS2S, MedrxivClusteringP2P, MedrxivClusteringS2S, RedditClustering, RedditClusteringP2P, StackExchangeClustering, StackExchangeClusteringP2P, TwentyNewsgroupsClustering\n* Pair Classification: SprintDuplicateQuestions, TwitterSemEval2015, TwitterURLCorpus\n* Reranking: AskUbuntuDupQuestions, MindSmallReranking, SciDocs, StackOverflowDupQuestions\n* Retrieval: ArguAna, ClimateFEVER, CQADupstackRetrieval, DBPedia, FEVER, FiQA2018, HotpotQA, MSMARCO, MSMARCOv2, NFCorpus, NQ, QuoraRetrieval, SCIDOCS, SciFact, Touche2020, TRECCOVID\n* STS: BIOSSES, SICK-R, STS12, STS13, STS14, STS15, STS16, STS17, STS22, STSBenchmark\n* Summarization: SummEval\n\nAll these datasets have been preprocessed and can be used for your experiments." ]

[ "TAGS\n#multilinguality-monolingual #multilinguality-multilingual #language-German #language-English #language-French #language-Russian #language-Chinese #license-cc-by-sa-4.0 #arxiv-2104.06893 #arxiv-2010.02573 #arxiv-2003.04807 #arxiv-2204.08582 #arxiv-2008.09335 #arxiv-2104.07081 #region-us \n", "# Dataset Card for MTEB Benchmark", "## Dataset Description\n\n- Homepage: URL\n- Repository: URL\n- Paper: soon\n- Leaderboard: URL\n- Point of Contact: nouamane@URL", "### Dataset Summary\n\nMTEB is a heterogeneous benchmark that has been built from diverse tasks:\n* BitextMining: BUCC, Tatoeba\n* Classification: AmazonCounterfactualClassification, AmazonPolarityClassification, AmazonReviewsClassification, Banking77Classification, EmotionClassification, ImdbClassification, MassiveIntentClassification, MassiveScenarioClassification, MTOPDomainClassification, MTOPIntentClassification, ToxicConversationsClassification, TweetSentimentExtractionClassification\n* Clustering: ArxivClusteringP2P, ArxivClusteringS2S, BiorxivClusteringP2P, BiorxivClusteringS2S, MedrxivClusteringP2P, MedrxivClusteringS2S, RedditClustering, RedditClusteringP2P, StackExchangeClustering, StackExchangeClusteringP2P, TwentyNewsgroupsClustering\n* Pair Classification: SprintDuplicateQuestions, TwitterSemEval2015, TwitterURLCorpus\n* Reranking: AskUbuntuDupQuestions, MindSmallReranking, SciDocs, StackOverflowDupQuestions\n* Retrieval: ArguAna, ClimateFEVER, CQADupstackRetrieval, DBPedia, FEVER, FiQA2018, HotpotQA, MSMARCO, MSMARCOv2, NFCorpus, NQ, QuoraRetrieval, SCIDOCS, SciFact, Touche2020, TRECCOVID\n* STS: BIOSSES, SICK-R, STS12, STS13, STS14, STS15, STS16, STS17, STS22, STSBenchmark\n* Summarization: SummEval\n\nAll these datasets have been preprocessed and can be used for your experiments." ]

[ "language:en", "region:us" ]

[]

[ "en" ]

[ "### Dataset Summary\n\n\nCoQA: A Conversational Question Answering Challenge", "### Supported Tasks and Leaderboards", "### Languages\n\n\nDataset Structure\n-----------------", "### Data Instances", "#### default\n\n\n* Size of downloaded dataset files: 55.40 MB\n* Size of the generated dataset: 18.35 MB\n* Total amount of disk used: 73.75 MB\n\n\nAn example of 'train' looks as follows.", "### Data Fields\n\n\nThe data fields are the same among all splits.", "#### default\n\n\n* 'source': a 'string' feature.\n* 'story': a 'string' feature.\n* 'questions': a 'list' of 'string' features.\n* 'answers': a dictionary feature containing:\n\t+ 'input\\_text': a 'string' feature.\n\t+ 'answer\\_start': a 'int32' feature.\n\t+ 'answer\\_end': a 'int32' feature.", "### Data Splits\n\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations\n\n\nAdditional Information\n----------------------", "### Dataset Curators", "### Licensing Information", "### Contributions\n\n\nThanks to @patrickvonplaten, @lewtun, @thomwolf, @mariamabarham, @ojasaar, @lhoestq for adding this dataset." ]

[ "TAGS\n#language-English #region-us \n", "### Dataset Summary\n\n\nCoQA: A Conversational Question Answering Challenge", "### Supported Tasks and Leaderboards", "### Languages\n\n\nDataset Structure\n-----------------", "### Data Instances", "#### default\n\n\n* Size of downloaded dataset files: 55.40 MB\n* Size of the generated dataset: 18.35 MB\n* Total amount of disk used: 73.75 MB\n\n\nAn example of 'train' looks as follows.", "### Data Fields\n\n\nThe data fields are the same among all splits.", "#### default\n\n\n* 'source': a 'string' feature.\n* 'story': a 'string' feature.\n* 'questions': a 'list' of 'string' features.\n* 'answers': a dictionary feature containing:\n\t+ 'input\\_text': a 'string' feature.\n\t+ 'answer\\_start': a 'int32' feature.\n\t+ 'answer\\_end': a 'int32' feature.", "### Data Splits\n\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations\n\n\nAdditional Information\n----------------------", "### Dataset Curators", "### Licensing Information", "### Contributions\n\n\nThanks to @patrickvonplaten, @lewtun, @thomwolf, @mariamabarham, @ojasaar, @lhoestq for adding this dataset." ]

[ "language:ro", "region:us" ]

[]

[ "ro" ]

[ "# Multi-microworld conversational agent dataset (RASA)\n\nIncluded microworlds (domains of knowledge):\n- generic\n- memory assistance\n- university guidance" ]

[ "TAGS\n#language-Romanian #region-us \n", "# Multi-microworld conversational agent dataset (RASA)\n\nIncluded microworlds (domains of knowledge):\n- generic\n- memory assistance\n- university guidance" ]

[ "task_categories:text-classification", "task_ids:sentiment-analysis", "task_ids:hate-speech-detection", "task_ids:language-identification", "multilinguality:multilingual", "language:si", "language:en", "license:mit", "region:us" ]

[]

[ "si", "en" ]

[ "# Sinhala-English-Code-Mixed-Code-Switched-Dataset\n\nThis dataset contains 10,000 comments that have been annotated at the sentence level for sentiment analysis, humor detection, hate speech detection, aspect identification, and language identification.\n\nThe following is the tag scheme.\n* Sentiment - Positive, Negative, Neutral, Conflict\n* Humor - Humorous, Non humorous\n* Hate Speech - Hate-Inducing, Abusive, Not offensive\n* Aspect - Network, Billing or Price, Package, Customer Service, Data, Service or product, None\n* Language ID - Sinhala, English, Sin-Eng, Eng-Sin, Mixed, Named-Entity, Symbol" ]

[ "TAGS\n#task_categories-text-classification #task_ids-sentiment-analysis #task_ids-hate-speech-detection #task_ids-language-identification #multilinguality-multilingual #language-Sinhala #language-English #license-mit #region-us \n", "# Sinhala-English-Code-Mixed-Code-Switched-Dataset\n\nThis dataset contains 10,000 comments that have been annotated at the sentence level for sentiment analysis, humor detection, hate speech detection, aspect identification, and language identification.\n\nThe following is the tag scheme.\n* Sentiment - Positive, Negative, Neutral, Conflict\n* Humor - Humorous, Non humorous\n* Hate Speech - Hate-Inducing, Abusive, Not offensive\n* Aspect - Network, Billing or Price, Package, Customer Service, Data, Service or product, None\n* Language ID - Sinhala, English, Sin-Eng, Eng-Sin, Mixed, Named-Entity, Symbol" ]

[ "task_categories:text-classification", "task_ids:fact-checking", "annotations_creators:crowdsourced", "language_creators:found", "multilinguality:monolingual", "size_categories:1K<n<10K", "source_datasets:original", "language:ar", "license:apache-2.0", "stance-detection", "arxiv:2005.10410", "region:us" ]

[ "2005.10410" ]

[ "ar" ]

[ "### Dataset Summary\n\n\nThe dataset is a collection of news titles in arabic along with paraphrased and corrupted titles. The stance prediction version is a 3-class classification task. Data contains three columns: s1, s2, stance.", "### Languages\n\n\nArabic\n\n\nDataset Structure\n-----------------", "### Data Instances\n\n\nAn example of 'train' looks as follows:", "### Data Fields\n\n\n* 'id': a 'string' feature.\n* 's1': a 'string' expressing a claim/topic.\n* 's2': a 'string' to be classified for its stance to the source.\n* 'stance': a class label representing the stance the article expresses towards the claim. Full tagset with indices:", "### Data Splits\n\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations\n\n\nAdditional Information\n----------------------", "### Dataset Curators\n\n\nThe dataset is curated by the paper's authors", "### Licensing Information\n\n\nThe authors distribute this data under the Apache License, Version 2.0", "### Contributions\n\n\nThanks to mkonxd for adding this dataset." ]

[ "TAGS\n#task_categories-text-classification #task_ids-fact-checking #annotations_creators-crowdsourced #language_creators-found #multilinguality-monolingual #size_categories-1K<n<10K #source_datasets-original #language-Arabic #license-apache-2.0 #stance-detection #arxiv-2005.10410 #region-us \n", "### Dataset Summary\n\n\nThe dataset is a collection of news titles in arabic along with paraphrased and corrupted titles. The stance prediction version is a 3-class classification task. Data contains three columns: s1, s2, stance.", "### Languages\n\n\nArabic\n\n\nDataset Structure\n-----------------", "### Data Instances\n\n\nAn example of 'train' looks as follows:", "### Data Fields\n\n\n* 'id': a 'string' feature.\n* 's1': a 'string' expressing a claim/topic.\n* 's2': a 'string' to be classified for its stance to the source.\n* 'stance': a class label representing the stance the article expresses towards the claim. Full tagset with indices:", "### Data Splits\n\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations\n\n\nAdditional Information\n----------------------", "### Dataset Curators\n\n\nThe dataset is curated by the paper's authors", "### Licensing Information\n\n\nThe authors distribute this data under the Apache License, Version 2.0", "### Contributions\n\n\nThanks to mkonxd for adding this dataset." ]

[ "region:us" ]

[]

[]

[]

[ "TAGS\n#region-us \n" ]

[ "task_categories:image-classification", "task_ids:multi-class-image-classification", "annotations_creators:crowdsourced", "language_creators:crowdsourced", "multilinguality:monolingual", "size_categories:10K<n<100K", "source_datasets:extended|imagenet-1k", "language:en", "license:unknown", "arxiv:1905.13549", "region:us" ]

[ "1905.13549" ]

[ "en" ]

[ "### Dataset Summary\n\n\nImageNet-Sketch data set consists of 50000 images, 50 images for each of the 1000 ImageNet classes. We construct the data set with Google Image queries \"sketch of \\_\\_\", where \\_\\_ is the standard class name. We only search within the \"black and white\" color scheme. We initially query 100 images for every class, and then manually clean the pulled images by deleting the irrelevant images and images that are for similar but different classes. For some classes, there are less than 50 images after manually cleaning, and then we augment the data set by flipping and rotating the images.\n\n\nThe scripts used to conduct queries and clean images can be found in the GitHub repository.", "### Supported Tasks and Leaderboards\n\n\n* 'image\\_classification': The goal of this task is to classify a given image into one of 1000 ImageNet classes. The leaderboard is available here.\n\n\nThe goal of the leaderboard is to evaluate the out-of-domain classification performance of vision models trained on ImageNet. The evaluation metrics used in the leaderboard are top-1 accuracy and top-5 accuracy.", "### Languages\n\n\nThe class labels in the dataset are in English.\n\n\nDataset Structure\n-----------------", "### Data Instances\n\n\nA sample from the training set is provided below:", "### Data Fields\n\n\nThe data instances have the following fields:\n\n\n* 'image': A 'PIL.Image.Image' object containing the image. Note that when accessing the image column: 'dataset[0][\"image\"]' the image file is automatically decoded. Decoding of a large number of image files might take a significant amount of time. Thus it is important to first query the sample index before the '\"image\"' column, *i.e.* 'dataset[0][\"image\"]' should always be preferred over 'dataset[\"image\"][0]'.\n* 'label': an 'int' classification label.\n\n\nThe labels are indexed based on a sorted list of synset ids such as 'n07565083' which we automatically map to original class names. The original dataset is divided into folders based on these synset ids. To get a mapping from original synset names, use the file LOC\\_synset\\_mapping.txt available on Kaggle challenge page. You can also use 'dataset\\_instance.features[\"label\"].int2str' function to get the class for a particular label index.\n\n\n\n\n Click here to see the full list of ImageNet class label mapping:", "### Data Splits\n\n\n\nDataset Creation\n----------------", "### Curation Rationale\n\n\nFrom the paper:\n\n\n\n> \n> Inspired by the Sketch data of (Li et al., 2017a) with seven classes, and several other Sketch datasets,\n> such as the Sketchy dataset (Sangkloy et al., 2016) with 125 classes and the Quick Draw! dataset\n> (QuickDraw, 2018) with 345 classes, and motivated by absence of a large-scale sketch dataset fitting\n> the shape and size of popular image classification benchmarks, we construct the ImageNet-Sketch\n> data set for evaluating the out-of-domain classification performance of vision models trained on\n> ImageNet.\n> \n> \n>", "### Source Data", "#### Initial Data Collection and Normalization\n\n\nThe initial data collection and normalization is inherited from ImageNet. More information on it can be found here.\n\n\nAdditional preprocessing from the paper:\n\n\n\n> \n> We construct the data set with Google Image queries “sketch of ”, where is the\n> standard class name. We only search within the “black and white” color scheme. We initially query\n> 100 images for every class, and then manually clean the pulled images by deleting the irrelevant\n> images and images that are for similar but different classes. For some classes, there are less than 50\n> images after manually cleaning, and then we augment the data set by flipping and rotating the images.\n> \n> \n>", "#### Who are the source language producers?\n\n\nThe source language is inherited from ImageNet. More information on the source language produces can be found here.", "### Annotations", "#### Annotation process\n\n\nThe annotations are inherited from ImageNet. More information about the process can be found here.", "#### Who are the annotators?\n\n\nThe same as in ImageNet.", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases\n\n\nThe biases are inherited from ImageNet. More information about the process can be found here.", "### Other Known Limitations\n\n\n1. Since most of the images were collected from internet, keep in mind that some images in ImageNet-Sketch might be subject to copyrights.\n\n\nAdditional Information\n----------------------", "### Dataset Curators\n\n\nAuthors of Learning Robust Global Representations by Penalizing Local Predictive Power:\n\n\n* Haohan Wang\n* Songwei Ge\n* Eric P. Xing\n* Zachary C. Lipton\n\n\nThe dataset was curated using the scripts found in the GitHub repository.", "### Licensing Information", "### Contributions\n\n\nThanks to @nateraw for adding this dataset." ]

[ "TAGS\n#task_categories-image-classification #task_ids-multi-class-image-classification #annotations_creators-crowdsourced #language_creators-crowdsourced #multilinguality-monolingual #size_categories-10K<n<100K #source_datasets-extended|imagenet-1k #language-English #license-unknown #arxiv-1905.13549 #region-us \n", "### Dataset Summary\n\n\nImageNet-Sketch data set consists of 50000 images, 50 images for each of the 1000 ImageNet classes. We construct the data set with Google Image queries \"sketch of \\_\\_\", where \\_\\_ is the standard class name. We only search within the \"black and white\" color scheme. We initially query 100 images for every class, and then manually clean the pulled images by deleting the irrelevant images and images that are for similar but different classes. For some classes, there are less than 50 images after manually cleaning, and then we augment the data set by flipping and rotating the images.\n\n\nThe scripts used to conduct queries and clean images can be found in the GitHub repository.", "### Supported Tasks and Leaderboards\n\n\n* 'image\\_classification': The goal of this task is to classify a given image into one of 1000 ImageNet classes. The leaderboard is available here.\n\n\nThe goal of the leaderboard is to evaluate the out-of-domain classification performance of vision models trained on ImageNet. The evaluation metrics used in the leaderboard are top-1 accuracy and top-5 accuracy.", "### Languages\n\n\nThe class labels in the dataset are in English.\n\n\nDataset Structure\n-----------------", "### Data Instances\n\n\nA sample from the training set is provided below:", "### Data Fields\n\n\nThe data instances have the following fields:\n\n\n* 'image': A 'PIL.Image.Image' object containing the image. Note that when accessing the image column: 'dataset[0][\"image\"]' the image file is automatically decoded. Decoding of a large number of image files might take a significant amount of time. Thus it is important to first query the sample index before the '\"image\"' column, *i.e.* 'dataset[0][\"image\"]' should always be preferred over 'dataset[\"image\"][0]'.\n* 'label': an 'int' classification label.\n\n\nThe labels are indexed based on a sorted list of synset ids such as 'n07565083' which we automatically map to original class names. The original dataset is divided into folders based on these synset ids. To get a mapping from original synset names, use the file LOC\\_synset\\_mapping.txt available on Kaggle challenge page. You can also use 'dataset\\_instance.features[\"label\"].int2str' function to get the class for a particular label index.\n\n\n\n\n Click here to see the full list of ImageNet class label mapping:", "### Data Splits\n\n\n\nDataset Creation\n----------------", "### Curation Rationale\n\n\nFrom the paper:\n\n\n\n> \n> Inspired by the Sketch data of (Li et al., 2017a) with seven classes, and several other Sketch datasets,\n> such as the Sketchy dataset (Sangkloy et al., 2016) with 125 classes and the Quick Draw! dataset\n> (QuickDraw, 2018) with 345 classes, and motivated by absence of a large-scale sketch dataset fitting\n> the shape and size of popular image classification benchmarks, we construct the ImageNet-Sketch\n> data set for evaluating the out-of-domain classification performance of vision models trained on\n> ImageNet.\n> \n> \n>", "### Source Data", "#### Initial Data Collection and Normalization\n\n\nThe initial data collection and normalization is inherited from ImageNet. More information on it can be found here.\n\n\nAdditional preprocessing from the paper:\n\n\n\n> \n> We construct the data set with Google Image queries “sketch of ”, where is the\n> standard class name. We only search within the “black and white” color scheme. We initially query\n> 100 images for every class, and then manually clean the pulled images by deleting the irrelevant\n> images and images that are for similar but different classes. For some classes, there are less than 50\n> images after manually cleaning, and then we augment the data set by flipping and rotating the images.\n> \n> \n>", "#### Who are the source language producers?\n\n\nThe source language is inherited from ImageNet. More information on the source language produces can be found here.", "### Annotations", "#### Annotation process\n\n\nThe annotations are inherited from ImageNet. More information about the process can be found here.", "#### Who are the annotators?\n\n\nThe same as in ImageNet.", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases\n\n\nThe biases are inherited from ImageNet. More information about the process can be found here.", "### Other Known Limitations\n\n\n1. Since most of the images were collected from internet, keep in mind that some images in ImageNet-Sketch might be subject to copyrights.\n\n\nAdditional Information\n----------------------", "### Dataset Curators\n\n\nAuthors of Learning Robust Global Representations by Penalizing Local Predictive Power:\n\n\n* Haohan Wang\n* Songwei Ge\n* Eric P. Xing\n* Zachary C. Lipton\n\n\nThe dataset was curated using the scripts found in the GitHub repository.", "### Licensing Information", "### Contributions\n\n\nThanks to @nateraw for adding this dataset." ]

[ "region:us" ]

[]

[]

[ "## Generation procedure\n\nThe dataset was constructed using documents from the Pile scored using using Perspective API toxicity scores.\n\nThe procedure was the following:\n1. A chunk of the Pile (2.2m documents) was scored using the Perspective API (on May 18-20 2022) giving 'tomekkorbak/pile-chunk-toxicity-scored-3'.\n1. The first half of this dataset is 100k *most* toxic documents from 'pile-chunk-toxicity-scored-3'\n2. The first half of this dataset is 100k documents sampled randomly from of 'pile-chunk-toxicity-scored-3'\n3. Then, the dataset was shuffled and a 9:1 train-test split was done", "## Basic stats\n\nThe average document-level scores of the bad and random halves are 0.34 and 0.05, respectively. The average token-level score of the whole dataset is 0.2025. The average document-level score is 0.1983.", "## Score histogram\n\n![](URL" ]

[ "TAGS\n#region-us \n", "## Generation procedure\n\nThe dataset was constructed using documents from the Pile scored using using Perspective API toxicity scores.\n\nThe procedure was the following:\n1. A chunk of the Pile (2.2m documents) was scored using the Perspective API (on May 18-20 2022) giving 'tomekkorbak/pile-chunk-toxicity-scored-3'.\n1. The first half of this dataset is 100k *most* toxic documents from 'pile-chunk-toxicity-scored-3'\n2. The first half of this dataset is 100k documents sampled randomly from of 'pile-chunk-toxicity-scored-3'\n3. Then, the dataset was shuffled and a 9:1 train-test split was done", "## Basic stats\n\nThe average document-level scores of the bad and random halves are 0.34 and 0.05, respectively. The average token-level score of the whole dataset is 0.2025. The average document-level score is 0.1983.", "## Score histogram\n\n![](URL" ]

[ "region:us" ]

[]

[]

[ "# Kinyarwanda dataset for text to speech model\n\nKinyarwanda dataset for text to speech model holds data for ai modelling of Kinyarwanda chatbots or other use cases." ]

[ "TAGS\n#region-us \n", "# Kinyarwanda dataset for text to speech model\n\nKinyarwanda dataset for text to speech model holds data for ai modelling of Kinyarwanda chatbots or other use cases." ]

[ "annotations_creators:crowd-sourced", "language_creators:Digital Umuganda", "multilinguality:monolingual", "size_categories:1M<n<3M", "source_datasets:original", "language:rw", "license:cc-by-4.0", "region:us" ]

[]

[ "rw" ]

[ "# Dataset Card for DigitalUmuganda/common-voice-kinyarwanda-text-dataset" ]

[ "TAGS\n#annotations_creators-crowd-sourced #language_creators-Digital Umuganda #multilinguality-monolingual #size_categories-1M<n<3M #source_datasets-original #language-Kinyarwanda #license-cc-by-4.0 #region-us \n", "# Dataset Card for DigitalUmuganda/common-voice-kinyarwanda-text-dataset" ]

[ "arxiv:2201.06499", "region:us" ]

[ "2201.06499" ]

[]

[ "# Dataset Card for ru-med-ner", "## Table of Contents\n- Dataset Description\n - Dataset Summary\n - Supported Tasks\n - Languages\n- Dataset Structure\n - Data Instances\n - Data Fields\n- Additional Information\n - Citation Information", "## Dataset Description\n\n- Homepage: URL\n- Repository: URL\n- Paper: URL\n- Leaderboard: URL\n- Point of Contact: Blinov.P.D@URL", "### Dataset Summary\n\nNER dataset for Russian language, extracted from medical records\\\\\nSee URL for details", "### Supported Tasks and Leaderboards", "### Languages\n\n- ru-RU", "## Dataset Structure", "### Data Instances", "### Data Fields\n\n- idx: example id\n- tokens: list of words from example\n- ner_tags: ner tags" ]

[ "TAGS\n#arxiv-2201.06499 #region-us \n", "# Dataset Card for ru-med-ner", "## Table of Contents\n- Dataset Description\n - Dataset Summary\n - Supported Tasks\n - Languages\n- Dataset Structure\n - Data Instances\n - Data Fields\n- Additional Information\n - Citation Information", "## Dataset Description\n\n- Homepage: URL\n- Repository: URL\n- Paper: URL\n- Leaderboard: URL\n- Point of Contact: Blinov.P.D@URL", "### Dataset Summary\n\nNER dataset for Russian language, extracted from medical records\\\\\nSee URL for details", "### Supported Tasks and Leaderboards", "### Languages\n\n- ru-RU", "## Dataset Structure", "### Data Instances", "### Data Fields\n\n- idx: example id\n- tokens: list of words from example\n- ner_tags: ner tags" ]

[ "region:us" ]

[]

[]

[]

[ "TAGS\n#region-us \n" ]

[ "license:other", "region:us" ]

[]

[]

[]

[ "TAGS\n#license-other #region-us \n" ]

[ "region:us" ]

[]

[]

[]

[ "TAGS\n#region-us \n" ]

[ "task_categories:text-classification", "task_ids:sentiment-classification", "annotations_creators:expert-generated", "language_creators:expert-generated", "multilinguality:monolingual", "size_categories:10K<n<100K", "source_datasets:extended|imdb", "language:en", "license:mit", "region:us" ]

[]

[ "en" ]

[ "### Dataset Summary\n\n\nThis dataset was crafted to be used in our tutorial [Link to the tutorial when ready]. It consists on a large Movie Review Dataset mixed with some reviews from a Hotel Review Dataset. The training/validation set are purely obtained from the Movie Review Dataset while the production set is mixed. Some other features have been added ('age', 'gender', 'context') as well as a made up timestamp 'prediction\\_ts' of when the inference took place.", "### Supported Tasks and Leaderboards\n\n\n'text-classification', 'sentiment-classification': The dataset is mainly used for text classification: given the text, predict the sentiment (positive or negative).", "### Languages\n\n\nText is mainly written in english.\n\n\nDataset Structure\n-----------------", "### Data Instances", "#### default\n\n\nAn example of 'training' looks as follows:", "### Data Fields", "#### default\n\n\nThe data fields are the same among all splits. An example of 'training' looks as follows:\n\n\n* 'prediction\\_ts': a 'float' feature.\n* 'age': an 'int' feature.\n* 'gender': a 'string' feature.\n* 'context': a 'string' feature.\n* 'text': a 'string' feature.\n* 'label': a 'ClassLabel' feature, with possible values including negative(0) and positive(1).", "### Data Splits\n\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Contributions\n\n\nThanks to @fjcasti1 for adding this dataset." ]

[ "TAGS\n#task_categories-text-classification #task_ids-sentiment-classification #annotations_creators-expert-generated #language_creators-expert-generated #multilinguality-monolingual #size_categories-10K<n<100K #source_datasets-extended|imdb #language-English #license-mit #region-us \n", "### Dataset Summary\n\n\nThis dataset was crafted to be used in our tutorial [Link to the tutorial when ready]. It consists on a large Movie Review Dataset mixed with some reviews from a Hotel Review Dataset. The training/validation set are purely obtained from the Movie Review Dataset while the production set is mixed. Some other features have been added ('age', 'gender', 'context') as well as a made up timestamp 'prediction\\_ts' of when the inference took place.", "### Supported Tasks and Leaderboards\n\n\n'text-classification', 'sentiment-classification': The dataset is mainly used for text classification: given the text, predict the sentiment (positive or negative).", "### Languages\n\n\nText is mainly written in english.\n\n\nDataset Structure\n-----------------", "### Data Instances", "#### default\n\n\nAn example of 'training' looks as follows:", "### Data Fields", "#### default\n\n\nThe data fields are the same among all splits. An example of 'training' looks as follows:\n\n\n* 'prediction\\_ts': a 'float' feature.\n* 'age': an 'int' feature.\n* 'gender': a 'string' feature.\n* 'context': a 'string' feature.\n* 'text': a 'string' feature.\n* 'label': a 'ClassLabel' feature, with possible values including negative(0) and positive(1).", "### Data Splits\n\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Contributions\n\n\nThanks to @fjcasti1 for adding this dataset." ]

[ "region:us" ]

[]

[]

[]

[ "TAGS\n#region-us \n" ]

[ "task_categories:summarization", "task_categories:text2text-generation", "multilinguality:monolingual", "size_categories:100K<n<1M", "language:en", "conditional-text-generation", "region:us" ]

[]

[ "en" ]

[ "### Data Fields\n\n\n* 'id': paper id\n* 'document': a string/list containing the body of a set of documents\n* 'summary': a string containing the abstract of the set", "### Data Splits\n\n\nThis dataset has 3 splits: *train*, *validation*, and *test*. \\\n\n\n\nCite original article\n=====================" ]

[ "TAGS\n#task_categories-summarization #task_categories-text2text-generation #multilinguality-monolingual #size_categories-100K<n<1M #language-English #conditional-text-generation #region-us \n", "### Data Fields\n\n\n* 'id': paper id\n* 'document': a string/list containing the body of a set of documents\n* 'summary': a string containing the abstract of the set", "### Data Splits\n\n\nThis dataset has 3 splits: *train*, *validation*, and *test*. \\\n\n\n\nCite original article\n=====================" ]

[ "annotations_creators:no-annotation", "language_creators:expert-generated", "multilinguality:monolingual", "size_categories:1M<n<10M", "source_datasets:original", "language:en", "license:cc0-1.0", "region:us" ]

[]

[ "en" ]

[ "# Dataset Card for Million Headlines", "## Table of Contents\n- Dataset Description\n - Dataset Summary\n - Supported Tasks and Leaderboards\n - Languages\n- Dataset Structure\n - Data Instances\n - Data Fields\n - Data Splits\n- Dataset Creation\n - Curation Rationale\n - Source Data\n - Annotations\n - Personal and Sensitive Information\n- Considerations for Using the Data\n - Social Impact of Dataset\n - Discussion of Biases\n - Other Known Limitations\n- Additional Information\n - Dataset Curators\n - Licensing Information\n - Citation Information\n - Contributions", "## Dataset Description\n\n- Homepage: Kaggle dataset\n- Point of Contact: Rohit Kulkarni)", "### Dataset Summary\n\nThis contains data of news headlines published over a period of eighteen years. Sourced from the reputable Australian news source ABC (Australian Broadcasting Corporation)", "## Dataset Structure", "### Data Instances\n\nFor each instance, there is a integer for the data, a string for news headline.", "### Data Fields\n\n- 'publish date': a integer that represents the data\n- 'headline': a string for the news headline", "### Personal and Sensitive Information\n\nThe dataset does not contain any personal information about the authors or the crowdworkers, but may contain descriptions of the people that were in the headlines.", "## Considerations for Using the Data", "### Social Impact of Dataset\n\nThis dataset represents one news service in Australia and should not be considered representative of all news or headlines.", "### Discussion of Biases\n\nNews headlines may contain biases and should not be considered neutral.", "### Licensing Information\n\nCC0: Public Domain." ]

[ "TAGS\n#annotations_creators-no-annotation #language_creators-expert-generated #multilinguality-monolingual #size_categories-1M<n<10M #source_datasets-original #language-English #license-cc0-1.0 #region-us \n", "# Dataset Card for Million Headlines", "## Table of Contents\n- Dataset Description\n - Dataset Summary\n - Supported Tasks and Leaderboards\n - Languages\n- Dataset Structure\n - Data Instances\n - Data Fields\n - Data Splits\n- Dataset Creation\n - Curation Rationale\n - Source Data\n - Annotations\n - Personal and Sensitive Information\n- Considerations for Using the Data\n - Social Impact of Dataset\n - Discussion of Biases\n - Other Known Limitations\n- Additional Information\n - Dataset Curators\n - Licensing Information\n - Citation Information\n - Contributions", "## Dataset Description\n\n- Homepage: Kaggle dataset\n- Point of Contact: Rohit Kulkarni)", "### Dataset Summary\n\nThis contains data of news headlines published over a period of eighteen years. Sourced from the reputable Australian news source ABC (Australian Broadcasting Corporation)", "## Dataset Structure", "### Data Instances\n\nFor each instance, there is a integer for the data, a string for news headline.", "### Data Fields\n\n- 'publish date': a integer that represents the data\n- 'headline': a string for the news headline", "### Personal and Sensitive Information\n\nThe dataset does not contain any personal information about the authors or the crowdworkers, but may contain descriptions of the people that were in the headlines.", "## Considerations for Using the Data", "### Social Impact of Dataset\n\nThis dataset represents one news service in Australia and should not be considered representative of all news or headlines.", "### Discussion of Biases\n\nNews headlines may contain biases and should not be considered neutral.", "### Licensing Information\n\nCC0: Public Domain." ]

[ "task_categories:text-generation", "task_ids:natural-language-inference", "annotations_creators:expert-generated", "language_creators:found", "language_creators:expert-generated", "multilinguality:monolingual", "size_categories:1K<n<10K", "source_datasets:original", "language:en", "license:mit", "region:us" ]

[]

[ "en" ]

[ "# Dataset Card for BBNLI", "## Table of Contents\n- Dataset Description\n - Dataset Summary\n - Supported Tasks\n - Languages\n- Dataset Structure\n - Data Instances\n - Data Fields\n - Data Splits\n- Dataset Creation\n - Curation Rationale\n - Source Data\n - Annotations\n - Personal and Sensitive Information\n- Considerations for Using the Data\n - Social Impact of Dataset\n - Discussion of Biases\n - Other Known Limitations\n- Additional Information\n - Dataset Curators\n - Licensing Information\n - Citation Information", "## Dataset Description\n\n- Homepage: \n- Repository: URL\n- Paper: URL\n- Leaderboard: \n- Point of Contact: Afra Feyza Akyurek", "### Dataset Summary\n\nBBNLI (Bias Benchmark for Natural Language Inference) is bias measurement benchmark for the tasks of both natural language inference and question answering. BBNLI consists of 16 subtopics each tailored to measure a specific stereotype that is negatively impacting certain classes. Each subtopic includes a set of 3 to 11 premises, 5 to 11 stereotypical hypotheses that are geared towards measuring biases and 3 to 5 test hypotheses.", "### Supported Tasks and Leaderboards\n\n- 'natural-language-inference', 'question-answering': The dataset can be used to test social biases presented in natural language inference or question answering systems.", "### Languages\n\nThe text is in English.", "## Dataset Structure", "### Data Instances\n\nA typical data point comprises of a premise or context and a hypothesis or a question depending on the task being considered. We also indicate if the hypothesis is designed to measure stereotypes or simple as a test hypothesis. For the distinction between the two cases please refer to the paper for more examples here.\n\nAn example from the BBNLI test set looks as follows:", "### Data Fields\n\n- Index: index\n- domain: domain among gender, religion or race\n- name: stereotype being tested\n- id: premise id\n- type: pro or anti stereotypical premise\n- unique_id: combination of name, type and id\n- premise: premise or context\n- hypothesis_type: test or stereotypical\n- hypothesis: hypothesis\n- question: question form of the hypothesis\n- true_label: correct label\n- bias_label: label is a stereotypical hypothesis/question\n- reference: source of the premise sentence", "### Data Splits\n\nThis dataset is configured only as a test set.", "## Dataset Creation", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information", "## Considerations for Using the Data", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations", "## Additional Information", "### Dataset Curators", "### Licensing Information" ]

[ "TAGS\n#task_categories-text-generation #task_ids-natural-language-inference #annotations_creators-expert-generated #language_creators-found #language_creators-expert-generated #multilinguality-monolingual #size_categories-1K<n<10K #source_datasets-original #language-English #license-mit #region-us \n", "# Dataset Card for BBNLI", "## Table of Contents\n- Dataset Description\n - Dataset Summary\n - Supported Tasks\n - Languages\n- Dataset Structure\n - Data Instances\n - Data Fields\n - Data Splits\n- Dataset Creation\n - Curation Rationale\n - Source Data\n - Annotations\n - Personal and Sensitive Information\n- Considerations for Using the Data\n - Social Impact of Dataset\n - Discussion of Biases\n - Other Known Limitations\n- Additional Information\n - Dataset Curators\n - Licensing Information\n - Citation Information", "## Dataset Description\n\n- Homepage: \n- Repository: URL\n- Paper: URL\n- Leaderboard: \n- Point of Contact: Afra Feyza Akyurek", "### Dataset Summary\n\nBBNLI (Bias Benchmark for Natural Language Inference) is bias measurement benchmark for the tasks of both natural language inference and question answering. BBNLI consists of 16 subtopics each tailored to measure a specific stereotype that is negatively impacting certain classes. Each subtopic includes a set of 3 to 11 premises, 5 to 11 stereotypical hypotheses that are geared towards measuring biases and 3 to 5 test hypotheses.", "### Supported Tasks and Leaderboards\n\n- 'natural-language-inference', 'question-answering': The dataset can be used to test social biases presented in natural language inference or question answering systems.", "### Languages\n\nThe text is in English.", "## Dataset Structure", "### Data Instances\n\nA typical data point comprises of a premise or context and a hypothesis or a question depending on the task being considered. We also indicate if the hypothesis is designed to measure stereotypes or simple as a test hypothesis. For the distinction between the two cases please refer to the paper for more examples here.\n\nAn example from the BBNLI test set looks as follows:", "### Data Fields\n\n- Index: index\n- domain: domain among gender, religion or race\n- name: stereotype being tested\n- id: premise id\n- type: pro or anti stereotypical premise\n- unique_id: combination of name, type and id\n- premise: premise or context\n- hypothesis_type: test or stereotypical\n- hypothesis: hypothesis\n- question: question form of the hypothesis\n- true_label: correct label\n- bias_label: label is a stereotypical hypothesis/question\n- reference: source of the premise sentence", "### Data Splits\n\nThis dataset is configured only as a test set.", "## Dataset Creation", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information", "## Considerations for Using the Data", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations", "## Additional Information", "### Dataset Curators", "### Licensing Information" ]

[ "region:us" ]

[]

[]

[]

[ "TAGS\n#region-us \n" ]

[ "region:us" ]

[]

[]

[]

[ "TAGS\n#region-us \n" ]

[ "task_categories:summarization", "annotations_creators:no-annotation", "language_creators:expert-generated", "multilinguality:monolingual", "size_categories:10K<n<100K", "source_datasets:original", "language:en", "license:cc-by-4.0", "region:us" ]

[]

[ "en" ]

[ "# Dataset Card for GovReport", "## Table of Contents\n- Table of Contents\n- Dataset Description\n - Dataset Summary\n - Versions\n - Supported Tasks and Leaderboards\n - Languages\n- Dataset Structure\n - Data Instances\n - Data Fields\n - Data Splits\n- Dataset Creation\n - Curation Rationale\n - Source Data\n - Annotations\n - Personal and Sensitive Information\n- Considerations for Using the Data\n - Social Impact of Dataset\n - Discussion of Biases\n - Other Known Limitations\n- Additional Information\n - Dataset Curators\n - Licensing Information\n - Citation Information\n - Contributions", "## Dataset Description\n\n- Homepage: URL\n- Repository: URL\n- Paper: URL\n- Leaderboard: \n- Point of Contact:", "### Dataset Summary\n\nGovernment report dataset consists of reports and associated summaries written by government research agencies including Congressional Research Service and U.S. Government Accountability Office.\n\nCompared with other long document summarization datasets, government report dataset has longer summaries and documents and requires reading in more context to cover salient words to be summarized.", "### Versions\n\n- '1.0.1' (default): remove extra whitespace.\n- '1.0.0': the dataset used in the original paper.\n\nTo use different versions, set the 'revision' argument of the 'load_dataset' function.", "### Supported Tasks and Leaderboards", "### Languages\n\nEnglish", "## Dataset Structure\n\nThree configs are available:\n- plain_text (default): the text-to-text summarization setting used as in the original paper.\n- plain_text_with_recommendations: the text-to-text summarization setting, with \"What GAO recommends\" included in the summary.\n- structure: data with the section structure.\n\nTo use different configs, set the 'name' argument of the 'load_dataset' function.", "### Data Instances", "#### plain_text & plain_text_with_recommendations\n\nAn example looks as follows.", "#### structure\n\nAn example looks as follows.", "### Data Fields", "#### plain_text & plain_text_with_recommendations\n\n- 'id': a 'string' feature.\n- 'document': a 'string' feature.\n- 'summary': a 'string' feature.", "#### structure\n\n- 'id': a 'string' feature.\n- 'document_sections': a dictionary feature containing lists of (each element corresponds to a section):\n - 'title': a 'string' feature.\n - 'paragraphs': a of 'string' feature, with '\\n' separating different paragraphs.\n - 'depth': a 'int32' feature.\n- 'summary_sections': a dictionary feature containing lists of (each element corresponds to a section):\n - 'title': a 'string' feature.\n - 'paragraphs': a 'string' feature, with '\\n' separating different paragraphs.", "### Data Splits\n\n- train: 17519\n- valid: 974\n- test: 973", "## Dataset Creation", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?\n\nEditors of the Congressional Research Service and U.S. Government Accountability Office.", "### Personal and Sensitive Information\n\nNone.", "## Considerations for Using the Data", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations", "## Additional Information", "### Dataset Curators", "### Licensing Information\n\nCC BY 4.0" ]

[ "TAGS\n#task_categories-summarization #annotations_creators-no-annotation #language_creators-expert-generated #multilinguality-monolingual #size_categories-10K<n<100K #source_datasets-original #language-English #license-cc-by-4.0 #region-us \n", "# Dataset Card for GovReport", "## Table of Contents\n- Table of Contents\n- Dataset Description\n - Dataset Summary\n - Versions\n - Supported Tasks and Leaderboards\n - Languages\n- Dataset Structure\n - Data Instances\n - Data Fields\n - Data Splits\n- Dataset Creation\n - Curation Rationale\n - Source Data\n - Annotations\n - Personal and Sensitive Information\n- Considerations for Using the Data\n - Social Impact of Dataset\n - Discussion of Biases\n - Other Known Limitations\n- Additional Information\n - Dataset Curators\n - Licensing Information\n - Citation Information\n - Contributions", "## Dataset Description\n\n- Homepage: URL\n- Repository: URL\n- Paper: URL\n- Leaderboard: \n- Point of Contact:", "### Dataset Summary\n\nGovernment report dataset consists of reports and associated summaries written by government research agencies including Congressional Research Service and U.S. Government Accountability Office.\n\nCompared with other long document summarization datasets, government report dataset has longer summaries and documents and requires reading in more context to cover salient words to be summarized.", "### Versions\n\n- '1.0.1' (default): remove extra whitespace.\n- '1.0.0': the dataset used in the original paper.\n\nTo use different versions, set the 'revision' argument of the 'load_dataset' function.", "### Supported Tasks and Leaderboards", "### Languages\n\nEnglish", "## Dataset Structure\n\nThree configs are available:\n- plain_text (default): the text-to-text summarization setting used as in the original paper.\n- plain_text_with_recommendations: the text-to-text summarization setting, with \"What GAO recommends\" included in the summary.\n- structure: data with the section structure.\n\nTo use different configs, set the 'name' argument of the 'load_dataset' function.", "### Data Instances", "#### plain_text & plain_text_with_recommendations\n\nAn example looks as follows.", "#### structure\n\nAn example looks as follows.", "### Data Fields", "#### plain_text & plain_text_with_recommendations\n\n- 'id': a 'string' feature.\n- 'document': a 'string' feature.\n- 'summary': a 'string' feature.", "#### structure\n\n- 'id': a 'string' feature.\n- 'document_sections': a dictionary feature containing lists of (each element corresponds to a section):\n - 'title': a 'string' feature.\n - 'paragraphs': a of 'string' feature, with '\\n' separating different paragraphs.\n - 'depth': a 'int32' feature.\n- 'summary_sections': a dictionary feature containing lists of (each element corresponds to a section):\n - 'title': a 'string' feature.\n - 'paragraphs': a 'string' feature, with '\\n' separating different paragraphs.", "### Data Splits\n\n- train: 17519\n- valid: 974\n- test: 973", "## Dataset Creation", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?\n\nEditors of the Congressional Research Service and U.S. Government Accountability Office.", "### Personal and Sensitive Information\n\nNone.", "## Considerations for Using the Data", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations", "## Additional Information", "### Dataset Curators", "### Licensing Information\n\nCC BY 4.0" ]

[ "task_categories:summarization", "annotations_creators:expert-generated", "language_creators:expert-generated", "multilinguality:monolingual", "size_categories:10K<n<100K", "source_datasets:launch/gov_report", "language:en", "license:cc-by-4.0", "region:us" ]

[]

[ "en" ]

[ "# Dataset Card for GovReport-QS", "## Table of Contents\n- Table of Contents\n- Dataset Description\n - Dataset Summary\n - Supported Tasks and Leaderboards\n - Languages\n- Dataset Structure\n - Data Instances\n - Data Fields\n - Data Splits\n- Dataset Creation\n - Curation Rationale\n - Source Data\n - Annotations\n - Personal and Sensitive Information\n- Considerations for Using the Data\n - Social Impact of Dataset\n - Discussion of Biases\n - Other Known Limitations\n- Additional Information\n - Dataset Curators\n - Licensing Information\n - Citation Information\n - Contributions", "## Dataset Description\n\n- Homepage: URL\n- Repository: URL\n- Paper: URL\n- Leaderboard: \n- Point of Contact:", "### Dataset Summary\n\nBased on the GovReport dataset, GovReport-QS additionally includes annotated question-summary hierarchies for government reports. This hierarchy proactively highlights the document structure, to further promote content engagement and comprehension.", "### Supported Tasks and Leaderboards", "### Languages\n\nEnglish", "## Dataset Structure\n\nTwo configs are available:\n- paragraph (default): paragraph-level annotated data\n- document: aggregated paragraph-level annotated data for the same document\n\nTo use different configs, set the 'name' argument of the 'load_dataset' function.", "### Data Instances", "#### paragraph\n\nAn example looks as follows.", "#### document\n\nAn example looks as follows.", "### Data Fields", "#### paragraph\n\nNote that document_sections in this config are the sections aligned with the annotated summary paragraph.\n\n- 'doc_id': a 'string' feature.\n- 'summary_paragraph_index': a 'int32' feature.\n- 'document_sections': a dictionary feature containing lists of (each element corresponds to a section):\n - 'title': a 'string' feature.\n - 'paragraphs': a of 'string' feature, with '\\n' separating different paragraphs.\n - 'depth': a 'int32' feature.\n- 'question_summary_pairs': a dictionary feature containing lists of (each element corresponds to a question-summary pair):\n - 'question': a 'string' feature.\n - 'summary': a 'string' feature.\n - 'parent_pair_index': a 'int32' feature indicating which question-summary pair is the parent of the current pair. '-1' indicates that the current pair does not have parent.", "#### document\n\nNote that document_sections in this config are the all sections in the document.\n\n- 'id': a 'string' feature.\n- 'document_sections': a dictionary feature containing lists of (each element corresponds to a section):\n - 'title': a 'string' feature.\n - 'paragraphs': a of 'string' feature, with '\\n' separating different paragraphs.\n - 'depth': a 'int32' feature.\n - 'alignment': a 'string' feature. Whether the 'full' section or the 'title' of the section should be included when aligned with each annotated hierarchy. For example, 'h0_full' indicates that the full section should be included for the hierarchy indexed '0'.\n- 'question_summary_pairs': a dictionary feature containing lists of:\n - 'question': a 'string' feature.\n - 'summary': a 'string' feature.\n - 'parent_pair_index': a 'int32' feature indicating which question-summary pair is the parent of the current pair. '-1' indicates that the current pair does not have parent. Note that the indices start from '0' for pairs with the same 'summary_paragraph_index'.\n - 'summary_paragraph_index': a 'int32' feature indicating which summary paragraph the question-summary pair is annotated for.", "### Data Splits", "#### paragraph\n\n- train: 17519\n- valid: 974\n- test: 973", "#### document\n\n- train: 1371\n- valid: 171\n- test: 172", "## Dataset Creation", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?\n\nEditors of the Congressional Research Service and U.S. Government Accountability Office.", "### Personal and Sensitive Information\n\nNone.", "## Considerations for Using the Data", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations", "## Additional Information", "### Dataset Curators", "### Licensing Information\n\nCC BY 4.0" ]

[ "TAGS\n#task_categories-summarization #annotations_creators-expert-generated #language_creators-expert-generated #multilinguality-monolingual #size_categories-10K<n<100K #source_datasets-launch/gov_report #language-English #license-cc-by-4.0 #region-us \n", "# Dataset Card for GovReport-QS", "## Table of Contents\n- Table of Contents\n- Dataset Description\n - Dataset Summary\n - Supported Tasks and Leaderboards\n - Languages\n- Dataset Structure\n - Data Instances\n - Data Fields\n - Data Splits\n- Dataset Creation\n - Curation Rationale\n - Source Data\n - Annotations\n - Personal and Sensitive Information\n- Considerations for Using the Data\n - Social Impact of Dataset\n - Discussion of Biases\n - Other Known Limitations\n- Additional Information\n - Dataset Curators\n - Licensing Information\n - Citation Information\n - Contributions", "## Dataset Description\n\n- Homepage: URL\n- Repository: URL\n- Paper: URL\n- Leaderboard: \n- Point of Contact:", "### Dataset Summary\n\nBased on the GovReport dataset, GovReport-QS additionally includes annotated question-summary hierarchies for government reports. This hierarchy proactively highlights the document structure, to further promote content engagement and comprehension.", "### Supported Tasks and Leaderboards", "### Languages\n\nEnglish", "## Dataset Structure\n\nTwo configs are available:\n- paragraph (default): paragraph-level annotated data\n- document: aggregated paragraph-level annotated data for the same document\n\nTo use different configs, set the 'name' argument of the 'load_dataset' function.", "### Data Instances", "#### paragraph\n\nAn example looks as follows.", "#### document\n\nAn example looks as follows.", "### Data Fields", "#### paragraph\n\nNote that document_sections in this config are the sections aligned with the annotated summary paragraph.\n\n- 'doc_id': a 'string' feature.\n- 'summary_paragraph_index': a 'int32' feature.\n- 'document_sections': a dictionary feature containing lists of (each element corresponds to a section):\n - 'title': a 'string' feature.\n - 'paragraphs': a of 'string' feature, with '\\n' separating different paragraphs.\n - 'depth': a 'int32' feature.\n- 'question_summary_pairs': a dictionary feature containing lists of (each element corresponds to a question-summary pair):\n - 'question': a 'string' feature.\n - 'summary': a 'string' feature.\n - 'parent_pair_index': a 'int32' feature indicating which question-summary pair is the parent of the current pair. '-1' indicates that the current pair does not have parent.", "#### document\n\nNote that document_sections in this config are the all sections in the document.\n\n- 'id': a 'string' feature.\n- 'document_sections': a dictionary feature containing lists of (each element corresponds to a section):\n - 'title': a 'string' feature.\n - 'paragraphs': a of 'string' feature, with '\\n' separating different paragraphs.\n - 'depth': a 'int32' feature.\n - 'alignment': a 'string' feature. Whether the 'full' section or the 'title' of the section should be included when aligned with each annotated hierarchy. For example, 'h0_full' indicates that the full section should be included for the hierarchy indexed '0'.\n- 'question_summary_pairs': a dictionary feature containing lists of:\n - 'question': a 'string' feature.\n - 'summary': a 'string' feature.\n - 'parent_pair_index': a 'int32' feature indicating which question-summary pair is the parent of the current pair. '-1' indicates that the current pair does not have parent. Note that the indices start from '0' for pairs with the same 'summary_paragraph_index'.\n - 'summary_paragraph_index': a 'int32' feature indicating which summary paragraph the question-summary pair is annotated for.", "### Data Splits", "#### paragraph\n\n- train: 17519\n- valid: 974\n- test: 973", "#### document\n\n- train: 1371\n- valid: 171\n- test: 172", "## Dataset Creation", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?\n\nEditors of the Congressional Research Service and U.S. Government Accountability Office.", "### Personal and Sensitive Information\n\nNone.", "## Considerations for Using the Data", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations", "## Additional Information", "### Dataset Curators", "### Licensing Information\n\nCC BY 4.0" ]

[ "region:us" ]

[]

[]

[]

[ "TAGS\n#region-us \n" ]

[ "task_categories:text2text-generation", "language:ms", "conditional-text-generation", "region:us" ]

[]

[ "ms" ]

[ "# rumi-jawi\n\nNotebooks to gather the dataset at URL" ]

[ "TAGS\n#task_categories-text2text-generation #language-Malay (macrolanguage) #conditional-text-generation #region-us \n", "# rumi-jawi\n\nNotebooks to gather the dataset at URL" ]

[ "task_categories:sentence-similarity", "language:si", "language:ta", "language:en", "region:us" ]

[]

[ "si", "ta", "en" ]

[ "### Dataset summary\n\n\nThis is a gold-standard benchmark dataset for document alignment, between Sinhala-English-Tamil languages.\nData had been crawled from the following news websites.\n\n\n\nThe aligned documents have been manually annotated.", "### Dataset\n\n\nThe folder structure for each news source is as follows.\n\n\nSinhala/English/Tamil - contain the crawled data for the respective news source\narmy\\_news\\_english\\_sinhala.txt - contains the annotated aligned documents between English and Sinhala languages.\narmynews\\_english\\_tamil.txt - contains the annotated aligned documents between English and Tamil languages.\narmynews\\_sinhala\\_tamil.txt - contains the annotated aligned documents between Sinhala and Tamil languages.\n\n\nCitation Information\n--------------------\n\n\n@article{fernando2022exploiting, \n\ntitle={Exploiting bilingual lexicons to improve multilingual embedding-based document and sentence alignment for low-resource languages}, \n\nauthor={Fernando, Aloka and Ranathunga, Surangika and Sachintha, Dilan and Piyarathna, Lakmali and Rajitha, Charith}, \n\njournal={Knowledge and Information Systems}, \n\npages={1--42}, \n\nyear={2022}, \n\npublisher={Springer} \n\n}" ]

[ "TAGS\n#task_categories-sentence-similarity #language-Sinhala #language-Tamil #language-English #region-us \n", "### Dataset summary\n\n\nThis is a gold-standard benchmark dataset for document alignment, between Sinhala-English-Tamil languages.\nData had been crawled from the following news websites.\n\n\n\nThe aligned documents have been manually annotated.", "### Dataset\n\n\nThe folder structure for each news source is as follows.\n\n\nSinhala/English/Tamil - contain the crawled data for the respective news source\narmy\\_news\\_english\\_sinhala.txt - contains the annotated aligned documents between English and Sinhala languages.\narmynews\\_english\\_tamil.txt - contains the annotated aligned documents between English and Tamil languages.\narmynews\\_sinhala\\_tamil.txt - contains the annotated aligned documents between Sinhala and Tamil languages.\n\n\nCitation Information\n--------------------\n\n\n@article{fernando2022exploiting, \n\ntitle={Exploiting bilingual lexicons to improve multilingual embedding-based document and sentence alignment for low-resource languages}, \n\nauthor={Fernando, Aloka and Ranathunga, Surangika and Sachintha, Dilan and Piyarathna, Lakmali and Rajitha, Charith}, \n\njournal={Knowledge and Information Systems}, \n\npages={1--42}, \n\nyear={2022}, \n\npublisher={Springer} \n\n}" ]

[ "license:apache-2.0", "region:us" ]

[]

[]

[]

[ "TAGS\n#license-apache-2.0 #region-us \n" ]

[ "task_categories:sentence-similarity", "task_categories:translation", "language:si", "language:ta", "language:en", "region:us" ]

[]

[ "si", "ta", "en" ]

[ "### Dataset summary\n\n\nThis is a gold-standard benchmark dataset for sentence alignment, between Sinhala-English-Tamil languages. Data had been crawled from the following news websites. The aligned documents annotated in the dataset NLPC-UOM/document\\_alignment\\_dataset-Sinhala-Tamil-English had been considered to annotate the aligned sentences.\n\n\n\nThe aligned sentences have been manually annotated.", "### Dataset\n\n\nThe folder structure for each news source is as follows.\n\n\nSinhala/English/Tamil - contain the aligned documents in the two languages with respect to the news source. (army/hiru/itn/newsfirst) Aligned documents contain the same ID. \n\nURL-en - golden aligned sentence alignment. Each sentence is referenced according to the languageprefix\\_fileid\\_sentenceId.", "### Citation Information\n\n\n@article{fernando2022exploiting, \n\ntitle={Exploiting bilingual lexicons to improve multilingual embedding-based document and sentence alignment for low-resource languages}, \n\nauthor={Fernando, Aloka and Ranathunga, Surangika and Sachintha, Dilan and Piyarathna, Lakmali and Rajitha, Charith}, \n\njournal={Knowledge and Information Systems}, \n\npages={1--42}, \n\nyear={2022}, \n\npublisher={Springer} \n\n}" ]

[ "TAGS\n#task_categories-sentence-similarity #task_categories-translation #language-Sinhala #language-Tamil #language-English #region-us \n", "### Dataset summary\n\n\nThis is a gold-standard benchmark dataset for sentence alignment, between Sinhala-English-Tamil languages. Data had been crawled from the following news websites. The aligned documents annotated in the dataset NLPC-UOM/document\\_alignment\\_dataset-Sinhala-Tamil-English had been considered to annotate the aligned sentences.\n\n\n\nThe aligned sentences have been manually annotated.", "### Dataset\n\n\nThe folder structure for each news source is as follows.\n\n\nSinhala/English/Tamil - contain the aligned documents in the two languages with respect to the news source. (army/hiru/itn/newsfirst) Aligned documents contain the same ID. \n\nURL-en - golden aligned sentence alignment. Each sentence is referenced according to the languageprefix\\_fileid\\_sentenceId.", "### Citation Information\n\n\n@article{fernando2022exploiting, \n\ntitle={Exploiting bilingual lexicons to improve multilingual embedding-based document and sentence alignment for low-resource languages}, \n\nauthor={Fernando, Aloka and Ranathunga, Surangika and Sachintha, Dilan and Piyarathna, Lakmali and Rajitha, Charith}, \n\njournal={Knowledge and Information Systems}, \n\npages={1--42}, \n\nyear={2022}, \n\npublisher={Springer} \n\n}" ]

[ "task_ids:masked-language-modeling", "multilinguality:monolingual", "size_categories:1M<n<10M", "source_datasets:TRex", "source_datasets:Lama", "language:en", "license:cc-by-sa-4.0", "license:cc-by-nc-4.0", "arxiv:2205.11482", "region:us" ]

[ "2205.11482" ]

[ "en" ]

[ "### Dataset Summary\n\n\n[PAPER]\nFTRACE is a zero-shot infromation retrieval benchmark deviced for tracing a language model’s predictions back to training examples. In the accompanying paper, we evaluate commonly studied influence methods, including gradient-based (TracIn) and embedding-based approaches. The dataset contains two parts. First, factual queries for that we trace the knowledge are extracted from existing LAMA queries (Petroni et al., 2019). Second, Wikidata sentences are extracted from TREx corpus (Elsahar et al., 2018). We annotate the extracted sentences with their stated facts, and these facts can be mathed with the facts in query set. In both parts, we provide (input, target) pairs as a masked language modeling task -- see examples in the below. However, one can use the same data in other formalities for example auto-regressive completion via a processing of 'input\\_pretokenized' and 'targets\\_pretokenized' field.", "### Supported Tasks and Leaderboards", "### Languages\n\n\nDataset Structure\n-----------------", "### Data Instances", "#### Abstracts\n\n\n* Size of downloaded dataset files: 112 MB\n* Size of the generated dataset: 884 MB\n* Total amount of disk used: 996 MB\n\n\nAn example of 'abstract' looks as follows.", "#### Queries\n\n\n* Size of downloaded dataset files: 1.7 MB\n* Size of the generated dataset: 8.9 MB\n* Total amount of disk used: 10.6 MB\n\n\nAn example of 'query' looks as follows.", "### Data Fields\n\n\nThe data fields are the same among all splits.", "#### Abstracts\n\n\n* 'inputs\\_pretokenized': a 'string' feature.\n* 'targets\\_pretokenized': a 'string' feature.\n* 'masked\\_uri': a 'string' feature.\n* 'masked\\_type': a 'string' feature.\n* 'facts': a 'string' feature.\n* 'id': a 'string' feature.\n* 'example\\_uris': a 'string' feature.\n* 'page\\_uri': a 'string' feature.", "#### Queries\n\n\n* 'inputs\\_pretokenized': a 'string' feature.\n* 'targets\\_pretokenized': a 'string' feature.\n* 'obj\\_surface': a 'string' feature.\n* 'sub\\_surface': a 'string' feature.\n* 'obj\\_uri': a 'string' feature.\n* 'sub\\_uri': a 'string' feature.\n* 'predicate\\_id': a 'string' feature.\n* 'uuid': a 'string' feature.", "### Data Splits\n\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Source Data\n\n\nLAMA: URL \n\nTRex: URL", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations\n\n\nAdditional Information\n----------------------", "### Dataset Curators", "### Licensing Information\n\n\nThe parts of this dataset are available under the Creative Commons Attribution-ShareAlike License (CC BY-SA 4.0) and The Creative Commons Attribution-Noncommercial 4.0 International License\n\n\nThe main paper should be cited as follow:\n\n\nPlease also cite Petroni et al., 2019 for the query set, and Elsahar et al., 2018 for the abstract set.", "### Contributions" ]

[ "TAGS\n#task_ids-masked-language-modeling #multilinguality-monolingual #size_categories-1M<n<10M #source_datasets-TRex #source_datasets-Lama #language-English #license-cc-by-sa-4.0 #license-cc-by-nc-4.0 #arxiv-2205.11482 #region-us \n", "### Dataset Summary\n\n\n[PAPER]\nFTRACE is a zero-shot infromation retrieval benchmark deviced for tracing a language model’s predictions back to training examples. In the accompanying paper, we evaluate commonly studied influence methods, including gradient-based (TracIn) and embedding-based approaches. The dataset contains two parts. First, factual queries for that we trace the knowledge are extracted from existing LAMA queries (Petroni et al., 2019). Second, Wikidata sentences are extracted from TREx corpus (Elsahar et al., 2018). We annotate the extracted sentences with their stated facts, and these facts can be mathed with the facts in query set. In both parts, we provide (input, target) pairs as a masked language modeling task -- see examples in the below. However, one can use the same data in other formalities for example auto-regressive completion via a processing of 'input\\_pretokenized' and 'targets\\_pretokenized' field.", "### Supported Tasks and Leaderboards", "### Languages\n\n\nDataset Structure\n-----------------", "### Data Instances", "#### Abstracts\n\n\n* Size of downloaded dataset files: 112 MB\n* Size of the generated dataset: 884 MB\n* Total amount of disk used: 996 MB\n\n\nAn example of 'abstract' looks as follows.", "#### Queries\n\n\n* Size of downloaded dataset files: 1.7 MB\n* Size of the generated dataset: 8.9 MB\n* Total amount of disk used: 10.6 MB\n\n\nAn example of 'query' looks as follows.", "### Data Fields\n\n\nThe data fields are the same among all splits.", "#### Abstracts\n\n\n* 'inputs\\_pretokenized': a 'string' feature.\n* 'targets\\_pretokenized': a 'string' feature.\n* 'masked\\_uri': a 'string' feature.\n* 'masked\\_type': a 'string' feature.\n* 'facts': a 'string' feature.\n* 'id': a 'string' feature.\n* 'example\\_uris': a 'string' feature.\n* 'page\\_uri': a 'string' feature.", "#### Queries\n\n\n* 'inputs\\_pretokenized': a 'string' feature.\n* 'targets\\_pretokenized': a 'string' feature.\n* 'obj\\_surface': a 'string' feature.\n* 'sub\\_surface': a 'string' feature.\n* 'obj\\_uri': a 'string' feature.\n* 'sub\\_uri': a 'string' feature.\n* 'predicate\\_id': a 'string' feature.\n* 'uuid': a 'string' feature.", "### Data Splits\n\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Source Data\n\n\nLAMA: URL \n\nTRex: URL", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations\n\n\nAdditional Information\n----------------------", "### Dataset Curators", "### Licensing Information\n\n\nThe parts of this dataset are available under the Creative Commons Attribution-ShareAlike License (CC BY-SA 4.0) and The Creative Commons Attribution-Noncommercial 4.0 International License\n\n\nThe main paper should be cited as follow:\n\n\nPlease also cite Petroni et al., 2019 for the query set, and Elsahar et al., 2018 for the abstract set.", "### Contributions" ]

[ "benchmark:gem", "evaluation", "benchmark", "region:us" ]

[]

[]

[ "# GEM Submission\n\nSubmission name: This is a test name" ]

[ "TAGS\n#benchmark-gem #evaluation #benchmark #region-us \n", "# GEM Submission\n\nSubmission name: This is a test name" ]

[ "benchmark:gem", "evaluation", "benchmark", "region:us" ]

[]

[]

[ "# GEM Submission\n\nSubmission name: This is a test name" ]

[ "TAGS\n#benchmark-gem #evaluation #benchmark #region-us \n", "# GEM Submission\n\nSubmission name: This is a test name" ]

[ "region:us" ]

[]

[]

[ "### Languages\n\n\nThe BCP-47 code for the dataset's language is unk.\n\n\nDataset Structure\n-----------------", "### Data Instances\n\n\nA sample from this dataset looks as follows:", "### Dataset Fields\n\n\nThe dataset has the following fields (also called \"features\"):", "### Dataset Splits\n\n\nThis dataset is split into a train and validation split. The split sizes are as follow:" ]

[ "TAGS\n#region-us \n", "### Languages\n\n\nThe BCP-47 code for the dataset's language is unk.\n\n\nDataset Structure\n-----------------", "### Data Instances\n\n\nA sample from this dataset looks as follows:", "### Dataset Fields\n\n\nThe dataset has the following fields (also called \"features\"):", "### Dataset Splits\n\n\nThis dataset is split into a train and validation split. The split sizes are as follow:" ]

[ "language:en", "region:us" ]

[]

[ "en" ]

[]

[ "TAGS\n#language-English #region-us \n" ]

[ "task_categories:text-classification", "task_ids:fact-checking", "annotations_creators:crowdsourced", "language_creators:found", "multilinguality:monolingual", "size_categories:1K<n<10K", "source_datasets:original", "language:ar", "license:cc-by-4.0", "stance-detection", "arxiv:2104.13559", "region:us" ]

[ "2104.13559" ]

[ "ar" ]

[ "### Dataset Summary\n\n\nThe AraStance dataset contains true and false claims, where each claim is paired with one or more documents. Each claim–article pair has a stance label: agree, disagree, discuss, or unrelated.", "### Languages\n\n\nArabic\n\n\nDataset Structure\n-----------------", "### Data Instances\n\n\nAn example of 'train' looks as follows:", "### Data Fields\n\n\n* 'id': a 'string' feature.\n* 'claim': a 'string' expressing a claim/topic.\n* 'article': a 'string' to be classified for its stance to the source.\n* 'stance': a class label representing the stance the article expresses towards the claim. Full tagset with indices:", "### Data Splits\n\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations\n\n\nAdditional Information\n----------------------", "### Dataset Curators\n\n\nThe dataset is curated by the paper's authors", "### Licensing Information\n\n\nThe authors distribute this data under Creative Commons attribution license, CC-BY 4.0", "### Contributions\n\n\nThanks to mkonxd for adding this dataset." ]

[ "TAGS\n#task_categories-text-classification #task_ids-fact-checking #annotations_creators-crowdsourced #language_creators-found #multilinguality-monolingual #size_categories-1K<n<10K #source_datasets-original #language-Arabic #license-cc-by-4.0 #stance-detection #arxiv-2104.13559 #region-us \n", "### Dataset Summary\n\n\nThe AraStance dataset contains true and false claims, where each claim is paired with one or more documents. Each claim–article pair has a stance label: agree, disagree, discuss, or unrelated.", "### Languages\n\n\nArabic\n\n\nDataset Structure\n-----------------", "### Data Instances\n\n\nAn example of 'train' looks as follows:", "### Data Fields\n\n\n* 'id': a 'string' feature.\n* 'claim': a 'string' expressing a claim/topic.\n* 'article': a 'string' to be classified for its stance to the source.\n* 'stance': a class label representing the stance the article expresses towards the claim. Full tagset with indices:", "### Data Splits\n\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations\n\n\nAdditional Information\n----------------------", "### Dataset Curators\n\n\nThe dataset is curated by the paper's authors", "### Licensing Information\n\n\nThe authors distribute this data under Creative Commons attribution license, CC-BY 4.0", "### Contributions\n\n\nThanks to mkonxd for adding this dataset." ]

[ "region:us" ]

[]

[]

[ "### SimRelUz: Similarity and Relatedness scores as a Semantic Evaluation dataset for Uzbek language\n\n\n\n We present a semantic model evaluation dataset: SimRelUz - a collection of similarity and relatedness scores of word pairs for Uzbek language. The dataset consists of more than a thousand pairs of words carefully selected based on their morphological features, occurrence frequency, semantic relation, as well as annotated by eleven native Uzbek speakers from different age groups and gender. \n Additionally, we also present a web-based tool to annotate similarity and relatedness scores. We also share the code to generate the scatter-plot to visualize word-pairs in a vector space.\n \n\n\n\nGitHub repo of the project\n\n\nAbout The Project\n-----------------\n\n\n\n\n\nThere are many language models that have been created that yield good quality semantic knowledge, yet their evaluation depends on gold standard datasets that have word/concept pairs scored by their semantic relations (such as synonymy, antonymy, meronymy, hypernymy, etc.), that come with cost due to their time-consuming context-generation process and high dependence on human annotators.\n\n\nCurrent project aims to present, to our knowledge, the first semantic similarity and relatedness dataset for Uzbek language. Furthermore, this repository includes a publicly-availabel code for the web-based tool created for semantic evaluation annotation.\n\n\nFeel free to use the dataset and the tool presented in this project, and if you find it useful, plese make sure to cite the paper here (coming soon...)\nDemo of the web-based annotation tool can be seen here.\n\n\n([back to top](#top))", "### Built With\n\n\nProgramming language used:\n\n\n* Python\n\n\nThese are the major libraries used inside Python:\n\n\n* scikit-learn : A set of python modules for machine learning\n* Matplotlib: Visualization with Python\n\n\n([back to top](#top))\n\n\nDataset\n-------\n\n\nThe visual representation of the dataset (word-pairs of database in the vector space) can be seen at the above diagram.\nThe dataset is composed of 1418 word pairs from different word types (nouns, adjectives and verbs), different word forms (root, inflectional, derivational), with different frequencies (high, mid, low frequencies, rare and OOV words), and with diverse pre-established semantic relations (synonym, antonym, meronym, hypernym, not related). All the pairs have two scores, one for semantic similarity, while the other is for semantic relatedness. No field in the dataset was left empty (as was requested from annotators in the guidelines, even for the OOV cases).\n\n\nMore detailed information can be seen in the table below:\n\n\n\n([back to top](#top))\n\n\nWeb tool\n--------\n\n\nThe user-interface of the presented web-based semantic evaluation tool designed for multiple-user annotation can be seen below in this picture:\n\n\n\n\n\n([back to top](#top))\n\n\nLicense\n-------\n\n\nDistributed under the GNU GENERAL PUBLIC LICENSE. See 'URL' for more information.\n\n\n([back to top](#top))\n\n\nAcknowledgments\n---------------\n\n\nWe would like to thank the NLP team of the Department of Information Technologies, Urgench State university for their huge help with the annotation.\n\n\nWe are grateful for these resources and tutorials for making this repository possible:\n\n\n* GitHub Readme template\n* Visual Studio Code\n\n\n([back to top](#top))" ]

[ "TAGS\n#region-us \n", "### SimRelUz: Similarity and Relatedness scores as a Semantic Evaluation dataset for Uzbek language\n\n\n\n We present a semantic model evaluation dataset: SimRelUz - a collection of similarity and relatedness scores of word pairs for Uzbek language. The dataset consists of more than a thousand pairs of words carefully selected based on their morphological features, occurrence frequency, semantic relation, as well as annotated by eleven native Uzbek speakers from different age groups and gender. \n Additionally, we also present a web-based tool to annotate similarity and relatedness scores. We also share the code to generate the scatter-plot to visualize word-pairs in a vector space.\n \n\n\n\nGitHub repo of the project\n\n\nAbout The Project\n-----------------\n\n\n\n\n\nThere are many language models that have been created that yield good quality semantic knowledge, yet their evaluation depends on gold standard datasets that have word/concept pairs scored by their semantic relations (such as synonymy, antonymy, meronymy, hypernymy, etc.), that come with cost due to their time-consuming context-generation process and high dependence on human annotators.\n\n\nCurrent project aims to present, to our knowledge, the first semantic similarity and relatedness dataset for Uzbek language. Furthermore, this repository includes a publicly-availabel code for the web-based tool created for semantic evaluation annotation.\n\n\nFeel free to use the dataset and the tool presented in this project, and if you find it useful, plese make sure to cite the paper here (coming soon...)\nDemo of the web-based annotation tool can be seen here.\n\n\n([back to top](#top))", "### Built With\n\n\nProgramming language used:\n\n\n* Python\n\n\nThese are the major libraries used inside Python:\n\n\n* scikit-learn : A set of python modules for machine learning\n* Matplotlib: Visualization with Python\n\n\n([back to top](#top))\n\n\nDataset\n-------\n\n\nThe visual representation of the dataset (word-pairs of database in the vector space) can be seen at the above diagram.\nThe dataset is composed of 1418 word pairs from different word types (nouns, adjectives and verbs), different word forms (root, inflectional, derivational), with different frequencies (high, mid, low frequencies, rare and OOV words), and with diverse pre-established semantic relations (synonym, antonym, meronym, hypernym, not related). All the pairs have two scores, one for semantic similarity, while the other is for semantic relatedness. No field in the dataset was left empty (as was requested from annotators in the guidelines, even for the OOV cases).\n\n\nMore detailed information can be seen in the table below:\n\n\n\n([back to top](#top))\n\n\nWeb tool\n--------\n\n\nThe user-interface of the presented web-based semantic evaluation tool designed for multiple-user annotation can be seen below in this picture:\n\n\n\n\n\n([back to top](#top))\n\n\nLicense\n-------\n\n\nDistributed under the GNU GENERAL PUBLIC LICENSE. See 'URL' for more information.\n\n\n([back to top](#top))\n\n\nAcknowledgments\n---------------\n\n\nWe would like to thank the NLP team of the Department of Information Technologies, Urgench State university for their huge help with the annotation.\n\n\nWe are grateful for these resources and tutorials for making this repository possible:\n\n\n* GitHub Readme template\n* Visual Studio Code\n\n\n([back to top](#top))" ]

[ "license:cc-by-4.0", "region:us" ]

[]

[]

[]

[ "TAGS\n#license-cc-by-4.0 #region-us \n" ]

[ "task_categories:text-classification", "language:en", "region:us" ]

[]

[ "en" ]

[ "### Languages\n\n\nThe BCP-47 code for the dataset's language is en.\n\n\nDataset Structure\n-----------------", "### Data Instances\n\n\nA sample from this dataset looks as follows:", "### Dataset Fields\n\n\nThe dataset has the following fields (also called \"features\"):", "### Dataset Splits\n\n\nThis dataset is split into a train and validation split. The split sizes are as follow:" ]

[ "TAGS\n#task_categories-text-classification #language-English #region-us \n", "### Languages\n\n\nThe BCP-47 code for the dataset's language is en.\n\n\nDataset Structure\n-----------------", "### Data Instances\n\n\nA sample from this dataset looks as follows:", "### Dataset Fields\n\n\nThe dataset has the following fields (also called \"features\"):", "### Dataset Splits\n\n\nThis dataset is split into a train and validation split. The split sizes are as follow:" ]

[ "language:en", "region:us" ]

[]

[ "en" ]

[ "# Dataset Card for \"amazon-shoe-reviews\"\n\nMore Information needed" ]

[ "TAGS\n#language-English #region-us \n", "# Dataset Card for \"amazon-shoe-reviews\"\n\nMore Information needed" ]

[ "region:us" ]

[]

[]

[ "# Title \nEntSUM: A Data Set for Entity-Centric Extractive Summarization", "# Author list\nMounica Maddela*, Mayank Kulkarni*, Daniel Preotiuc-Pietro", "# Description\nControllable summarization aims to provide summaries that take into account user-specified aspects and preferences to better assist them with their information need, as opposed to the standard summarization setup which build a single generic summary of a document. We introduce a human-annotated data set EntSUM for controllable summarization with a focus on named entities as the aspects to control. We conduct an extensive quantitative analysis to motivate the task of entity-centric summarization and show that existing methods for controllable summarization fail to generate entity-centric summaries. We propose extensions to state-of-the-art summarization approaches that achieve substantially better results on our data set. Our analysis and results show the challenging nature of this task and of the proposed data set.\nAs a part of this zip file, we release the EntSum dataset on which the evaluations are performed. There are three json files, namely, one summary annotation, two summary annotations and a combination of both. Each file contains the document ID from the NYT corpus, the sentence IDs, the summary(s), the salient sentences and summary sentence corresponding to the sentence IDs. Obtaining the source text can be done by downloading the original NYT corpus and mapping the document IDs. The annotation process and pre-processing details are described extensively in the research paper.", "# Language\nEnglish", "# Keywords\nNatural Language Processing, Summarization, Abstractive Summarization, Extractive Summarization", "# Related identifiers \nNYT – is the source that this data set is derived from - URL License (LDC) URL" ]

[ "TAGS\n#region-us \n", "# Title \nEntSUM: A Data Set for Entity-Centric Extractive Summarization", "# Author list\nMounica Maddela*, Mayank Kulkarni*, Daniel Preotiuc-Pietro", "# Description\nControllable summarization aims to provide summaries that take into account user-specified aspects and preferences to better assist them with their information need, as opposed to the standard summarization setup which build a single generic summary of a document. We introduce a human-annotated data set EntSUM for controllable summarization with a focus on named entities as the aspects to control. We conduct an extensive quantitative analysis to motivate the task of entity-centric summarization and show that existing methods for controllable summarization fail to generate entity-centric summaries. We propose extensions to state-of-the-art summarization approaches that achieve substantially better results on our data set. Our analysis and results show the challenging nature of this task and of the proposed data set.\nAs a part of this zip file, we release the EntSum dataset on which the evaluations are performed. There are three json files, namely, one summary annotation, two summary annotations and a combination of both. Each file contains the document ID from the NYT corpus, the sentence IDs, the summary(s), the salient sentences and summary sentence corresponding to the sentence IDs. Obtaining the source text can be done by downloading the original NYT corpus and mapping the document IDs. The annotation process and pre-processing details are described extensively in the research paper.", "# Language\nEnglish", "# Keywords\nNatural Language Processing, Summarization, Abstractive Summarization, Extractive Summarization", "# Related identifiers \nNYT – is the source that this data set is derived from - URL License (LDC) URL" ]

[ "task_categories:translation", "annotations_creators:machine-generated", "annotations_creators:expert-generated", "language_creators:found", "multilinguality:translation", "size_categories:1K<n<10K", "source_datasets:original", "language:en", "language:it", "language:vi", "language:nl", "language:uk", "language:tr", "language:ar", "license:gpl-3.0", "arxiv:2205.12215", "region:us" ]

[ "2205.12215" ]

[ "en", "it", "vi", "nl", "uk", "tr", "ar" ]

[ "### Dataset Summary\n\n\nThis dataset contains the processed 'warmup' and 'main' splits of the DivEMT dataset. A sample of documents extracted from the Flores-101 corpus were either translated from scratch or post-edited from an existing automatic translation by a total of 18 professional translators across six typologically diverse languages (Arabic, Dutch, Italian, Turkish, Ukrainian, Vietnamese). During the translation, behavioral data (keystrokes, pauses, editing times) were collected using the PET platform.\n\n\nWe publicly release the processed dataset including all collected behavioural data, to foster new research on the ability of state-of-the-art NMT systems to generate text in typologically diverse languages.", "### News\n\n\nFebruary, 2023: The DivEMT dataset now contains linguistic annotations ('\\*\\_annotations' fields) computed with Stanza and word-level quality estimation tags ('src\\_wmt22\\_qe', 'mt\\_wmt22\\_qe') obtained using the same scripts adopted for the WMT22 QE Task 2.", "### Languages\n\n\nThe language data of DivEMT is in English (BCP-47 'en'), Italian (BCP-47 'it'), Dutch (BCP-47 'nl'), Arabic (BCP-47 'ar'), Turkish (BCP-47 'tr'), Ukrainian (BCP-47 'uk') and Vietnamese (BCP-47 'vi')\n\n\nDataset Structure\n-----------------", "### Data Instances\n\n\nThe dataset contains two configurations: 'main' and 'warmup'. 'main' contains the full data collected during the main task and analyzed during our experiments. 'warmup' contains the data collected in the verification phase, before the main task begins.", "### Data Fields\n\n\nThe following fields are contained in the training set:", "### Data Splits", "#### Train Split\n\n\nThe 'train' split contains the totality of triplets (or pairs, when translation from scratch is performed) annotated with behavioral data produced during the translation.\n\n\nThe following is an example of the subject 't1' post-editing a machine translation produced by Google Translate (task\\_type 'pe1') taken from the 'train' split for Turkish. The field 'aligned\\_edit' is showed over three lines to provide a visual understanding of its contents.\n\n\nThe text is provided as-is, without further preprocessing or tokenization.", "### Dataset Creation\n\n\nThe dataset was parsed from PET XML files into CSV format using the scripts available in the DivEMT Github repository.\n\n\nThose are adapted from the ones by Antonio Toral found at the following link: URL\n\n\nAdditional Information\n----------------------", "### Dataset Curators\n\n\nFor problems related to this Datasets version, please contact me at g.sarti@URL." ]

[ "TAGS\n#task_categories-translation #annotations_creators-machine-generated #annotations_creators-expert-generated #language_creators-found #multilinguality-translation #size_categories-1K<n<10K #source_datasets-original #language-English #language-Italian #language-Vietnamese #language-Dutch #language-Ukrainian #language-Turkish #language-Arabic #license-gpl-3.0 #arxiv-2205.12215 #region-us \n", "### Dataset Summary\n\n\nThis dataset contains the processed 'warmup' and 'main' splits of the DivEMT dataset. A sample of documents extracted from the Flores-101 corpus were either translated from scratch or post-edited from an existing automatic translation by a total of 18 professional translators across six typologically diverse languages (Arabic, Dutch, Italian, Turkish, Ukrainian, Vietnamese). During the translation, behavioral data (keystrokes, pauses, editing times) were collected using the PET platform.\n\n\nWe publicly release the processed dataset including all collected behavioural data, to foster new research on the ability of state-of-the-art NMT systems to generate text in typologically diverse languages.", "### News\n\n\nFebruary, 2023: The DivEMT dataset now contains linguistic annotations ('\\*\\_annotations' fields) computed with Stanza and word-level quality estimation tags ('src\\_wmt22\\_qe', 'mt\\_wmt22\\_qe') obtained using the same scripts adopted for the WMT22 QE Task 2.", "### Languages\n\n\nThe language data of DivEMT is in English (BCP-47 'en'), Italian (BCP-47 'it'), Dutch (BCP-47 'nl'), Arabic (BCP-47 'ar'), Turkish (BCP-47 'tr'), Ukrainian (BCP-47 'uk') and Vietnamese (BCP-47 'vi')\n\n\nDataset Structure\n-----------------", "### Data Instances\n\n\nThe dataset contains two configurations: 'main' and 'warmup'. 'main' contains the full data collected during the main task and analyzed during our experiments. 'warmup' contains the data collected in the verification phase, before the main task begins.", "### Data Fields\n\n\nThe following fields are contained in the training set:", "### Data Splits", "#### Train Split\n\n\nThe 'train' split contains the totality of triplets (or pairs, when translation from scratch is performed) annotated with behavioral data produced during the translation.\n\n\nThe following is an example of the subject 't1' post-editing a machine translation produced by Google Translate (task\\_type 'pe1') taken from the 'train' split for Turkish. The field 'aligned\\_edit' is showed over three lines to provide a visual understanding of its contents.\n\n\nThe text is provided as-is, without further preprocessing or tokenization.", "### Dataset Creation\n\n\nThe dataset was parsed from PET XML files into CSV format using the scripts available in the DivEMT Github repository.\n\n\nThose are adapted from the ones by Antonio Toral found at the following link: URL\n\n\nAdditional Information\n----------------------", "### Dataset Curators\n\n\nFor problems related to this Datasets version, please contact me at g.sarti@URL." ]

[ "task_categories:audio-classification", "task_ids:audio-emotion-recognition", "annotations_creators:expert-generated", "language_creators:expert-generated", "language_creators:crowdsourced", "multilinguality:monolingual", "size_categories:1K<n<10K", "source_datasets:original", "language:ru", "license:mit", "doi:10.57967/hf/1273", "region:us" ]

[]

[ "ru" ]

[ "# Dataset Card for resd", "## Table of Contents\n- Table of Contents\n- Dataset Description\n - Dataset Summary\n - Supported Tasks and Leaderboards\n - Languages\n- Dataset Structure\n - Data Instances\n - Data Fields\n - Data Splits\n- Dataset Creation\n - Curation Rationale\n - Source Data\n - Annotations\n - Personal and Sensitive Information\n- Considerations for Using the Data\n - Social Impact of Dataset\n - Discussion of Biases\n - Other Known Limitations\n- Additional Information\n - Dataset Curators\n - Licensing Information\n - Citation Information\n - Contributions", "## Dataset Description\n\n- Homepage: URL\n- Repository: URL\n- Paper:\n- Leaderboard:\n- Point of Contact:", "### Dataset Summary\n\nRussian dataset of emotional speech dialogues. This dataset was assembled from ~3.5 hours of live speech by actors who voiced pre-distributed emotions in the dialogue for ~3 minutes each.", "### Supported Tasks and Leaderboards", "### Languages", "## Dataset Structure", "### Data Instances", "### Data Fields", "### Data Splits", "## Dataset Creation", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information", "## Considerations for Using the Data", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations", "## Additional Information", "### Dataset Curators", "### Licensing Information\n\nThis dataset was created by Artem Amentes, Nikita Davidchuk and Ilya Lubenets", "### Contributions\n\nThanks to @Ar4ikov for adding this dataset." ]

[ "TAGS\n#task_categories-audio-classification #task_ids-audio-emotion-recognition #annotations_creators-expert-generated #language_creators-expert-generated #language_creators-crowdsourced #multilinguality-monolingual #size_categories-1K<n<10K #source_datasets-original #language-Russian #license-mit #doi-10.57967/hf/1273 #region-us \n", "# Dataset Card for resd", "## Table of Contents\n- Table of Contents\n- Dataset Description\n - Dataset Summary\n - Supported Tasks and Leaderboards\n - Languages\n- Dataset Structure\n - Data Instances\n - Data Fields\n - Data Splits\n- Dataset Creation\n - Curation Rationale\n - Source Data\n - Annotations\n - Personal and Sensitive Information\n- Considerations for Using the Data\n - Social Impact of Dataset\n - Discussion of Biases\n - Other Known Limitations\n- Additional Information\n - Dataset Curators\n - Licensing Information\n - Citation Information\n - Contributions", "## Dataset Description\n\n- Homepage: URL\n- Repository: URL\n- Paper:\n- Leaderboard:\n- Point of Contact:", "### Dataset Summary\n\nRussian dataset of emotional speech dialogues. This dataset was assembled from ~3.5 hours of live speech by actors who voiced pre-distributed emotions in the dialogue for ~3 minutes each.", "### Supported Tasks and Leaderboards", "### Languages", "## Dataset Structure", "### Data Instances", "### Data Fields", "### Data Splits", "## Dataset Creation", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information", "## Considerations for Using the Data", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations", "## Additional Information", "### Dataset Curators", "### Licensing Information\n\nThis dataset was created by Artem Amentes, Nikita Davidchuk and Ilya Lubenets", "### Contributions\n\nThanks to @Ar4ikov for adding this dataset." ]

[ "task_categories:audio-classification", "task_ids:slot-filling", "annotations_creators:unknown", "language_creators:unknown", "size_categories:100K<n<100M", "source_datasets:unknown", "license:mit", "region:us" ]

[]

[]

[ "# DCASE 2022 Task 3 Data sets: STARSS22 Dataset + Synthetic SELD mixtures\n\nAudio Research Group / Tampere University\nCreative AI Lab/ SONY R&D Center", "## Important\nThis is a copy from the Zenodo Original one\n\nAUTHORS\n\nTampere University\n- Archontis Politis (contact, profile)\n- Parthasaarathy Sudarsanam(contact, profile)\n- Sharath Adavanne (contact, profile)\n- Daniel Krause (contact, profile)\n- Tuomas Virtanen (contact, profile)\n\nSONY\n- Yuki Mitsufuji (contact, profile)\n- Kazuki Shimada (contact, profile)\n- Naoya Takahashi (profile)\n- Yuichiro Koyama\n- Shusuke Takahashi", "# Description\n\nThe Sony-TAu Realistic Spatial Soundscapes 2022 (STARSS22) dataset contains multichannel recordings of sound scenes in various rooms and environments, together with temporal and spatial annotations of prominent events belonging to a set of target classes. The dataset is collected in two different countries, in Tampere, Finland by the Audio Researh Group (ARG) of Tampere University (TAU), and in Tokyo, Japan by SONY, using a similar setup and annotation procedure. The dataset is delivered in two 4-channel spatial recording formats, a microphone array one (MIC), and first-order Ambisonics one (FOA). These recordings serve as the development dataset for the DCASE 2022 Sound Event Localization and Detection Task of the DCASE 2022 Challenge.\n\nContrary to the three previous datasets of synthetic spatial sound scenes of TAU Spatial Sound Events 2019 (development/evaluation), TAU-NIGENS Spatial Sound Events 2020, and TAU-NIGENS Spatial Sound Events 2021 associated with the previous iterations of the DCASE Challenge, the STARSS22 dataset contains recordings of real sound scenes and hence it avoids some of the pitfalls of synthetic generation of scenes. Some such key properties are:\n\n- annotations are based on a combination of human annotators for sound event activity and optical tracking for spatial positions\n- the annotated target event classes are determined by the composition of the real scenes \n- the density, polyphony, occurences and co-occurences of events and sound classes is not random, and it follows actions and interactions of participants in the real scenes \n\nThe recordings were collected between September 2021 and February 2022. Collection of data from the TAU side has received funding from Google.", "# Aim\n\nThe dataset is suitable for training and evaluation of machine-listening models for sound event detection (SED), general sound source localization with diverse sounds or signal-of-interest localization, and joint sound-event-localization-and-detection (SELD). Additionally, the dataset can be used for evaluation of signal processing methods that do not necessarily rely on training, such as acoustic source localization methods and multiple-source acoustic tracking. The dataset allows evaluation of the performance and robustness of the aforementioned applications for diverse types of sounds, and under diverse acoustic conditions.", "# Recording procedure\n\nThe sound scene recordings were captured with a high-channel-count spherical microphone array (Eigenmike em32 by mh Acoustics), simultaneously with a 360° video recording spatially aligned with the spherical array recording (Ricoh Theta V). Additionally, the main sound sources of interest were equipped with tracking markers, which are tracked throughout the recording with an Optitrack Flex 13 system arranged around each scene. All scenes were based on human actors performing some actions, interacting between them and with the objects in the scene, and were by design dynamic. Since the actors were producing most of the sounds in the scene (but not all), they were additionally equipped with DPA Wireless Go II microphones, providing close-miked recordings of the main events. Recording would start and stop according to a scene being acted, usually lasting between 1~5mins. Recording would start in all microphones and tracking devices before the beginning of the scene, and would stop right after. A clapper sound would initiate the acting and it would serve as a reference signal for synchronization between the em32 recording, the Ricoh Theta V video, the DPA wireless microphone recordings, and the Optitrack tracker data. Synchronized clips of all of them would be cropped and stored in the end of each recording session.", "# Annotation procedure\n\nBy combining information from the wireless microphones, the optical tracking data, and the 360° videos, spatiotemporal annotations were extracted semi-automatically, and validated manually. More specifically, the actors were tracked all through each recording session wearing headbands with markers, and the spatial positions of other human-related sources, such as mouth, hands, or footsteps were geometrically extrapolated from those head coordinates. Additional trackers were mounted on other sources of interest (e.g. vacuum cleaner, guitar, water tap, cupboard, door handle, a.o.). Each actor had a wireless microphone mounted on their lapel, providing a clear recording of all sound events produced by that actor, and/or any independent sources closer to that actor than the rest. The temporal annotation was based primarily on those close-miked recordings. The annotators would annotate the sound event activity and label their class during the recording by listening those close-miked signals. Events that were not audible in the overall scene recording of the em32 were not annotated, even if they were audible in the lapel recordings. In ambiguous cases, the annotators could rely on the 360° video to associate an event with a certain actor or source. The final sound event temporal annotations were associated with the tracking data through the class of each sound event and the actor that produced them. All tracked Cartesian coordinates delivered by the tracker were converted to directions-of-arrival (DOAs) with respect to the coordinates of the Eigenmike. Finally, the final class, temporal, and spatial annotations were combined and converted to the challenge format. Validation of the annotations was done by observing videos of the activities of each class visualized as markers positioned at their respective DOAs on the 360° video plane, overlapped with the 360° from the Ricoh Theta V.", "# Recording formats\n\nThe array response of the two recording formats can be considered known. The following theoretical spatial responses (steering vectors) modeling the two formats describe the directional response of each channel to a source incident from direction-of-arrival (DOA) given by azimuth angle $\\phi$ and elevation angle $\\theta$.\n\nFor the first-order ambisonics (FOA):\n\n\\begin{eqnarray}\nH_1(\\phi, \\theta, f) &=& 1 \\\\\nH_2(\\phi, \\theta, f) &=& \\sin(\\phi) * \\cos(\\theta) \\\\\nH_3(\\phi, \\theta, f) &=& \\sin(\\theta) \\\\\nH_4(\\phi, \\theta, f) &=& \\cos(\\phi) * \\cos(\\theta)\n\\end{eqnarray}\nThe (FOA) format is obtained by converting the 32-channel microphone array signals by means of encoding filters based on anechoic measurements of the Eigenmike array response. Note that in the formulas above the encoding format is assumed frequency-independent, something that holds true up to around 9kHz with the specific microphone array, while the actual encoded responses start to deviate gradually at higher frequencies from the ideal ones provided above. \n\nFor the tetrahedral microphone array (MIC):\n\nThe four microphone have the following positions, in spherical coordinates $(\\phi, \\theta, r)$:\n\n\\begin{eqnarray} \nM1: &\\quad(&45^\\circ, &&35^\\circ, &4.2\\mathrm{cm})\\nonumber\\\\\nM2: &\\quad(&-45^\\circ, &-&35^\\circ, &4.2\\mathrm{cm})\\nonumber\\\\\nM3: &\\quad(&135^\\circ, &-&35^\\circ, &4.2\\mathrm{cm})\\nonumber\\\\\nM4: &\\quad(&-135^\\circ, &&35^\\circ, &4.2\\mathrm{cm})\\nonumber\n\\end{eqnarray}\n\nSince the microphones are mounted on an acoustically-hard spherical baffle, an analytical expression for the directional array response is given by the expansion:\n\\begin{equation}\nH_m(\\phi_m, \\theta_m, \\phi, \\theta, \\omega) = \\frac{1}{(\\omega R/c)^2}\\sum_{n=0}^{30} \\frac{i^{n-1}}{h_n'^{(2)}(\\omega R/c)}(2n+1)P_n(\\cos(\\gamma_m))\n\\end{equation}\n\nwhere $m$ is the channel number, $(\\phi_m, \\theta_m)$ are the specific microphone's azimuth and elevation position, $\\omega = 2\\pi f$ is the angular frequency, $R = 0.042$m is the array radius, $c = 343$m/s is the speed of sound, $\\cos(\\gamma_m)$ is the cosine angle between the microphone and the DOA, and $P_n$ is the unnormalized Legendre polynomial of degree $n$, and $h_n'^{(2)}$ is the derivative with respect to the argument of a spherical Hankel function of the second kind. The expansion is limited to 30 terms which provides negligible modeling error up to 20kHz. Example routines that can generate directional frequency and impulse array responses based on the above formula can be found here.", "# Dataset specifications\n\nThe specifications of the dataset can be summarized in the following:\n\n- 70 recording clips of 30 sec ~ 5 min durations, with a total time of ~2hrs, contributed by SONY (development dataset).\n- 51 recording clips of 1 min ~ 5 min durations, with a total time of ~3hrs, contributed by TAU (development dataset).\n- A training-test split is provided for reporting results using the development dataset.\n- 40 recordings contributed by SONY for the training split, captured in 2 rooms (dev-train-sony).\n- 30 recordings contributed by SONY for the testing split, captured in 2 rooms (dev-test-sony).\n- 27 recordings contributed by TAU for the training split, captured in 4 rooms (dev-train-tau).\n- 24 recordings contributed by TAU for the testing split, captured in 3 rooms (dev-test-tau).\n- A total of 11 unique rooms captured in the recordings, 4 from SONY and 7 from TAU (development set).\n- Sampling rate 24kHz.\n- Two 4-channel 3-dimensional recording formats: first-order Ambisonics (FOA) and tetrahedral microphone array (MIC).\n- Recordings are taken in two different countries and two different sites.\n- Each recording clip is part of a recording session happening in a unique room.\n- Groups of participants, sound making props, and scene scenarios are unique for each session (with a few exceptions).\n- To achieve good variability and efficiency in the data, in terms of presence, density, movement, and/or spatial distribution of the sounds events, the scenes are loosely scripted.\n- 13 target classes are identified in the recordings and strongly annotated by humans.\n- Spatial annotations for those active events are captured by an optical tracking system.\n- Sound events out of the target classes are considered as interference.", "# Sound event classes\n\n13 target sound event classes were annotated. The classes follow loosely the Audioset ontology.\n\n 0. Female speech, woman speaking\n 1. Male speech, man speaking\n 2. Clapping\n 3. Telephone\n 4. Laughter\n 5. Domestic sounds\n 6. Walk, footsteps\n 7. Door, open or close\n 8. Music\n 9. Musical instrument\n 10. Water tap, faucet\n 11. Bell\n 12. Knock\n\nThe content of some of these classes corresponds to events of a limited range of Audioset-related subclasses. These are detailed here as additional information on the diversity of those sound events:\n\n - Telephone\n - Mostly traditional _Telephone Bell Ringing_ and _Ringtone_ sounds, without musical ringtones.\n - Domestic sounds\n - Sounds of _Vacuum cleaner_\n - Sounds of water boiler, closer to _Boiling_\n - Sounds of air circulator, closer to _Mechanical fan_\n - Door, open or close\n - Combination of _Door_ and _Cupboard open or close_\n - Music\n - _Background music_ and _Pop music_ played by a loudspeaker in the room.\n - Musical Instrument\n - Acoustic guitar\n - Marimba, xylophone\n - Cowbell\n - Piano\n - Rattle (instrument)\n - Bell\n - Combination of sounds from hotel bell and glass bell, closer to _Bicycle bell_ and single _Chime_.\n\nSome additional notes:\n- The speech classes contain speech in a few different languages.\n- There are occasionally localized sound events that are not annotated and are considered as interferers, with examples such as _computer keyboard_, _shuffling cards_, _dishes, pots, and pans_.\n- There is natural background noise (e.g. HVAC noise) in all recordings, at very low levels in some and at quite high levels in others. Such mostly diffuse background noise should be distinct from other noisy target sources (e.g. vacuum cleaner, mechanical fan) since these are clearly spatially localized.", "# Naming Convention (Development dataset)\n\nThe recordings in the development dataset follow the naming convention:\n\n fold[fold number]_room[room number]_mix[recording number per room].wav\n\nThe fold number at the moment is used only to distinguish between the training and testing split. The room information is provided for the user of the dataset to potentially help understand the performance of their method with respect to different conditions.", "# Reference labels and directions-of-arrival\n\nFor each recording in the development dataset, the labels and DoAs are provided in a plain text CSV file of the same filename as the recording, in the following format:\n\n [frame number (int)], [active class index (int)], [source number index (int)], [azimuth (int)], [elevation (int)]\n\nFrame, class, and source enumeration begins at 0. Frames correspond to a temporal resolution of 100msec. Azimuth and elevation angles are given in degrees, rounded to the closest integer value, with azimuth and elevation being zero at the front, azimuth $\\phi \\in [-180^{\\circ}, 180^{\\circ}]$, and elevation $\\theta \\in [-90^{\\circ}, 90^{\\circ}]$. Note that the azimuth angle is increasing counter-clockwise ($\\phi = 90^{\\circ}$ at the left). \n\nThe source index is a unique integer for each source in the scene, and it is provided only as additional information. Note that each unique actor gets assigned one such identifier, but not individual events produced by the same actor; e.g. a _clapping_ event and a _laughter_ event produced by the same person have the same identifier. Independent sources that are not actors (e.g. a loudspeaker playing music in the room) get a 0 identifier. Note that source identifier information is only included in the development metadata and is not required to be provided by the participants in their results.\n\nOverlapping sound events are indicated with duplicate frame numbers, and can belong to a different or the same class. An example sequence could be as:\n\n 10, 1, 1, -50, 30\n 11, 1, 1, -50, 30\n 11, 1, 2, 10, -20\n 12, 1, 2, 10, -20\n 13, 1, 2, 10, -20\n 13, 8, 0, -40, 0\n\nwhich describes that in frame 10-11, an event of class _male speech_ (_class 1_) belonging to one actor (_source 1_) is active at direction (-50°,30°). However, at frame 11 a second instance of the same class appears simultaneously at a different direction (10°,-20°) belonging to another actor (_source 2_), while at frame 13 an additional event of class _music_ (_class 8_) appears belonging to a non-actor source (_source 0_). Frames that contain no sound events are not included in the sequence.", "# Task setup\n\nThe dataset is associated with the DCASE 2022 Challenge. To have consistent reporting of results between participants on the development set a pre-defined training-testing split is provided. To compare against the challenge baseline and with other participants during the development stage, models should be trained on the training split only, and results should be reported on the testing split only.\n\nNote that even though there are two origins of the data, SONY and TAU, the challenge task considers the dataset as a single entity. Hence models should not be trained separately for each of the two origins, and tested individually on recordings of each of them. Instead, the recordings of the individual training splits (_dev-test-sony_, _dev_test_tau_) and testing splits (_dev-test-sony_, _dev_test_tau_) should be combined (_dev_train_, _dev_test_) and the models should be trained and evaluated in the respective combined splits.\n\nThe evaluation part of the dataset will be published here as a new dataset version, a few weeks before the final challenge submission deadline. The additional evaluation files consist of only audio recordings without any metadata/labels. Participants can decide the training procedure, i.e. the amount of training and validation files in the development dataset, the number of ensemble models etc., and submit the results of the SELD performance on the evaluation dataset.", "# File structure", "# Download\n\ngit clone", "# Example application\n\nAn implementation of a trainable model of a convolutional recurrent neural network, performing joint SELD, trained and evaluated with this dataset is provided here. This implementation will serve as the baseline method in the DCASE 2022 Sound Event Localization and Detection Task.", "# License\n\nThis datast is licensed under the MIT license." ]

[ "TAGS\n#task_categories-audio-classification #task_ids-slot-filling #annotations_creators-unknown #language_creators-unknown #size_categories-100K<n<100M #source_datasets-unknown #license-mit #region-us \n", "# DCASE 2022 Task 3 Data sets: STARSS22 Dataset + Synthetic SELD mixtures\n\nAudio Research Group / Tampere University\nCreative AI Lab/ SONY R&D Center", "## Important\nThis is a copy from the Zenodo Original one\n\nAUTHORS\n\nTampere University\n- Archontis Politis (contact, profile)\n- Parthasaarathy Sudarsanam(contact, profile)\n- Sharath Adavanne (contact, profile)\n- Daniel Krause (contact, profile)\n- Tuomas Virtanen (contact, profile)\n\nSONY\n- Yuki Mitsufuji (contact, profile)\n- Kazuki Shimada (contact, profile)\n- Naoya Takahashi (profile)\n- Yuichiro Koyama\n- Shusuke Takahashi", "# Description\n\nThe Sony-TAu Realistic Spatial Soundscapes 2022 (STARSS22) dataset contains multichannel recordings of sound scenes in various rooms and environments, together with temporal and spatial annotations of prominent events belonging to a set of target classes. The dataset is collected in two different countries, in Tampere, Finland by the Audio Researh Group (ARG) of Tampere University (TAU), and in Tokyo, Japan by SONY, using a similar setup and annotation procedure. The dataset is delivered in two 4-channel spatial recording formats, a microphone array one (MIC), and first-order Ambisonics one (FOA). These recordings serve as the development dataset for the DCASE 2022 Sound Event Localization and Detection Task of the DCASE 2022 Challenge.\n\nContrary to the three previous datasets of synthetic spatial sound scenes of TAU Spatial Sound Events 2019 (development/evaluation), TAU-NIGENS Spatial Sound Events 2020, and TAU-NIGENS Spatial Sound Events 2021 associated with the previous iterations of the DCASE Challenge, the STARSS22 dataset contains recordings of real sound scenes and hence it avoids some of the pitfalls of synthetic generation of scenes. Some such key properties are:\n\n- annotations are based on a combination of human annotators for sound event activity and optical tracking for spatial positions\n- the annotated target event classes are determined by the composition of the real scenes \n- the density, polyphony, occurences and co-occurences of events and sound classes is not random, and it follows actions and interactions of participants in the real scenes \n\nThe recordings were collected between September 2021 and February 2022. Collection of data from the TAU side has received funding from Google.", "# Aim\n\nThe dataset is suitable for training and evaluation of machine-listening models for sound event detection (SED), general sound source localization with diverse sounds or signal-of-interest localization, and joint sound-event-localization-and-detection (SELD). Additionally, the dataset can be used for evaluation of signal processing methods that do not necessarily rely on training, such as acoustic source localization methods and multiple-source acoustic tracking. The dataset allows evaluation of the performance and robustness of the aforementioned applications for diverse types of sounds, and under diverse acoustic conditions.", "# Recording procedure\n\nThe sound scene recordings were captured with a high-channel-count spherical microphone array (Eigenmike em32 by mh Acoustics), simultaneously with a 360° video recording spatially aligned with the spherical array recording (Ricoh Theta V). Additionally, the main sound sources of interest were equipped with tracking markers, which are tracked throughout the recording with an Optitrack Flex 13 system arranged around each scene. All scenes were based on human actors performing some actions, interacting between them and with the objects in the scene, and were by design dynamic. Since the actors were producing most of the sounds in the scene (but not all), they were additionally equipped with DPA Wireless Go II microphones, providing close-miked recordings of the main events. Recording would start and stop according to a scene being acted, usually lasting between 1~5mins. Recording would start in all microphones and tracking devices before the beginning of the scene, and would stop right after. A clapper sound would initiate the acting and it would serve as a reference signal for synchronization between the em32 recording, the Ricoh Theta V video, the DPA wireless microphone recordings, and the Optitrack tracker data. Synchronized clips of all of them would be cropped and stored in the end of each recording session.", "# Annotation procedure\n\nBy combining information from the wireless microphones, the optical tracking data, and the 360° videos, spatiotemporal annotations were extracted semi-automatically, and validated manually. More specifically, the actors were tracked all through each recording session wearing headbands with markers, and the spatial positions of other human-related sources, such as mouth, hands, or footsteps were geometrically extrapolated from those head coordinates. Additional trackers were mounted on other sources of interest (e.g. vacuum cleaner, guitar, water tap, cupboard, door handle, a.o.). Each actor had a wireless microphone mounted on their lapel, providing a clear recording of all sound events produced by that actor, and/or any independent sources closer to that actor than the rest. The temporal annotation was based primarily on those close-miked recordings. The annotators would annotate the sound event activity and label their class during the recording by listening those close-miked signals. Events that were not audible in the overall scene recording of the em32 were not annotated, even if they were audible in the lapel recordings. In ambiguous cases, the annotators could rely on the 360° video to associate an event with a certain actor or source. The final sound event temporal annotations were associated with the tracking data through the class of each sound event and the actor that produced them. All tracked Cartesian coordinates delivered by the tracker were converted to directions-of-arrival (DOAs) with respect to the coordinates of the Eigenmike. Finally, the final class, temporal, and spatial annotations were combined and converted to the challenge format. Validation of the annotations was done by observing videos of the activities of each class visualized as markers positioned at their respective DOAs on the 360° video plane, overlapped with the 360° from the Ricoh Theta V.", "# Recording formats\n\nThe array response of the two recording formats can be considered known. The following theoretical spatial responses (steering vectors) modeling the two formats describe the directional response of each channel to a source incident from direction-of-arrival (DOA) given by azimuth angle $\\phi$ and elevation angle $\\theta$.\n\nFor the first-order ambisonics (FOA):\n\n\\begin{eqnarray}\nH_1(\\phi, \\theta, f) &=& 1 \\\\\nH_2(\\phi, \\theta, f) &=& \\sin(\\phi) * \\cos(\\theta) \\\\\nH_3(\\phi, \\theta, f) &=& \\sin(\\theta) \\\\\nH_4(\\phi, \\theta, f) &=& \\cos(\\phi) * \\cos(\\theta)\n\\end{eqnarray}\nThe (FOA) format is obtained by converting the 32-channel microphone array signals by means of encoding filters based on anechoic measurements of the Eigenmike array response. Note that in the formulas above the encoding format is assumed frequency-independent, something that holds true up to around 9kHz with the specific microphone array, while the actual encoded responses start to deviate gradually at higher frequencies from the ideal ones provided above. \n\nFor the tetrahedral microphone array (MIC):\n\nThe four microphone have the following positions, in spherical coordinates $(\\phi, \\theta, r)$:\n\n\\begin{eqnarray} \nM1: &\\quad(&45^\\circ, &&35^\\circ, &4.2\\mathrm{cm})\\nonumber\\\\\nM2: &\\quad(&-45^\\circ, &-&35^\\circ, &4.2\\mathrm{cm})\\nonumber\\\\\nM3: &\\quad(&135^\\circ, &-&35^\\circ, &4.2\\mathrm{cm})\\nonumber\\\\\nM4: &\\quad(&-135^\\circ, &&35^\\circ, &4.2\\mathrm{cm})\\nonumber\n\\end{eqnarray}\n\nSince the microphones are mounted on an acoustically-hard spherical baffle, an analytical expression for the directional array response is given by the expansion:\n\\begin{equation}\nH_m(\\phi_m, \\theta_m, \\phi, \\theta, \\omega) = \\frac{1}{(\\omega R/c)^2}\\sum_{n=0}^{30} \\frac{i^{n-1}}{h_n'^{(2)}(\\omega R/c)}(2n+1)P_n(\\cos(\\gamma_m))\n\\end{equation}\n\nwhere $m$ is the channel number, $(\\phi_m, \\theta_m)$ are the specific microphone's azimuth and elevation position, $\\omega = 2\\pi f$ is the angular frequency, $R = 0.042$m is the array radius, $c = 343$m/s is the speed of sound, $\\cos(\\gamma_m)$ is the cosine angle between the microphone and the DOA, and $P_n$ is the unnormalized Legendre polynomial of degree $n$, and $h_n'^{(2)}$ is the derivative with respect to the argument of a spherical Hankel function of the second kind. The expansion is limited to 30 terms which provides negligible modeling error up to 20kHz. Example routines that can generate directional frequency and impulse array responses based on the above formula can be found here.", "# Dataset specifications\n\nThe specifications of the dataset can be summarized in the following:\n\n- 70 recording clips of 30 sec ~ 5 min durations, with a total time of ~2hrs, contributed by SONY (development dataset).\n- 51 recording clips of 1 min ~ 5 min durations, with a total time of ~3hrs, contributed by TAU (development dataset).\n- A training-test split is provided for reporting results using the development dataset.\n- 40 recordings contributed by SONY for the training split, captured in 2 rooms (dev-train-sony).\n- 30 recordings contributed by SONY for the testing split, captured in 2 rooms (dev-test-sony).\n- 27 recordings contributed by TAU for the training split, captured in 4 rooms (dev-train-tau).\n- 24 recordings contributed by TAU for the testing split, captured in 3 rooms (dev-test-tau).\n- A total of 11 unique rooms captured in the recordings, 4 from SONY and 7 from TAU (development set).\n- Sampling rate 24kHz.\n- Two 4-channel 3-dimensional recording formats: first-order Ambisonics (FOA) and tetrahedral microphone array (MIC).\n- Recordings are taken in two different countries and two different sites.\n- Each recording clip is part of a recording session happening in a unique room.\n- Groups of participants, sound making props, and scene scenarios are unique for each session (with a few exceptions).\n- To achieve good variability and efficiency in the data, in terms of presence, density, movement, and/or spatial distribution of the sounds events, the scenes are loosely scripted.\n- 13 target classes are identified in the recordings and strongly annotated by humans.\n- Spatial annotations for those active events are captured by an optical tracking system.\n- Sound events out of the target classes are considered as interference.", "# Sound event classes\n\n13 target sound event classes were annotated. The classes follow loosely the Audioset ontology.\n\n 0. Female speech, woman speaking\n 1. Male speech, man speaking\n 2. Clapping\n 3. Telephone\n 4. Laughter\n 5. Domestic sounds\n 6. Walk, footsteps\n 7. Door, open or close\n 8. Music\n 9. Musical instrument\n 10. Water tap, faucet\n 11. Bell\n 12. Knock\n\nThe content of some of these classes corresponds to events of a limited range of Audioset-related subclasses. These are detailed here as additional information on the diversity of those sound events:\n\n - Telephone\n - Mostly traditional _Telephone Bell Ringing_ and _Ringtone_ sounds, without musical ringtones.\n - Domestic sounds\n - Sounds of _Vacuum cleaner_\n - Sounds of water boiler, closer to _Boiling_\n - Sounds of air circulator, closer to _Mechanical fan_\n - Door, open or close\n - Combination of _Door_ and _Cupboard open or close_\n - Music\n - _Background music_ and _Pop music_ played by a loudspeaker in the room.\n - Musical Instrument\n - Acoustic guitar\n - Marimba, xylophone\n - Cowbell\n - Piano\n - Rattle (instrument)\n - Bell\n - Combination of sounds from hotel bell and glass bell, closer to _Bicycle bell_ and single _Chime_.\n\nSome additional notes:\n- The speech classes contain speech in a few different languages.\n- There are occasionally localized sound events that are not annotated and are considered as interferers, with examples such as _computer keyboard_, _shuffling cards_, _dishes, pots, and pans_.\n- There is natural background noise (e.g. HVAC noise) in all recordings, at very low levels in some and at quite high levels in others. Such mostly diffuse background noise should be distinct from other noisy target sources (e.g. vacuum cleaner, mechanical fan) since these are clearly spatially localized.", "# Naming Convention (Development dataset)\n\nThe recordings in the development dataset follow the naming convention:\n\n fold[fold number]_room[room number]_mix[recording number per room].wav\n\nThe fold number at the moment is used only to distinguish between the training and testing split. The room information is provided for the user of the dataset to potentially help understand the performance of their method with respect to different conditions.", "# Reference labels and directions-of-arrival\n\nFor each recording in the development dataset, the labels and DoAs are provided in a plain text CSV file of the same filename as the recording, in the following format:\n\n [frame number (int)], [active class index (int)], [source number index (int)], [azimuth (int)], [elevation (int)]\n\nFrame, class, and source enumeration begins at 0. Frames correspond to a temporal resolution of 100msec. Azimuth and elevation angles are given in degrees, rounded to the closest integer value, with azimuth and elevation being zero at the front, azimuth $\\phi \\in [-180^{\\circ}, 180^{\\circ}]$, and elevation $\\theta \\in [-90^{\\circ}, 90^{\\circ}]$. Note that the azimuth angle is increasing counter-clockwise ($\\phi = 90^{\\circ}$ at the left). \n\nThe source index is a unique integer for each source in the scene, and it is provided only as additional information. Note that each unique actor gets assigned one such identifier, but not individual events produced by the same actor; e.g. a _clapping_ event and a _laughter_ event produced by the same person have the same identifier. Independent sources that are not actors (e.g. a loudspeaker playing music in the room) get a 0 identifier. Note that source identifier information is only included in the development metadata and is not required to be provided by the participants in their results.\n\nOverlapping sound events are indicated with duplicate frame numbers, and can belong to a different or the same class. An example sequence could be as:\n\n 10, 1, 1, -50, 30\n 11, 1, 1, -50, 30\n 11, 1, 2, 10, -20\n 12, 1, 2, 10, -20\n 13, 1, 2, 10, -20\n 13, 8, 0, -40, 0\n\nwhich describes that in frame 10-11, an event of class _male speech_ (_class 1_) belonging to one actor (_source 1_) is active at direction (-50°,30°). However, at frame 11 a second instance of the same class appears simultaneously at a different direction (10°,-20°) belonging to another actor (_source 2_), while at frame 13 an additional event of class _music_ (_class 8_) appears belonging to a non-actor source (_source 0_). Frames that contain no sound events are not included in the sequence.", "# Task setup\n\nThe dataset is associated with the DCASE 2022 Challenge. To have consistent reporting of results between participants on the development set a pre-defined training-testing split is provided. To compare against the challenge baseline and with other participants during the development stage, models should be trained on the training split only, and results should be reported on the testing split only.\n\nNote that even though there are two origins of the data, SONY and TAU, the challenge task considers the dataset as a single entity. Hence models should not be trained separately for each of the two origins, and tested individually on recordings of each of them. Instead, the recordings of the individual training splits (_dev-test-sony_, _dev_test_tau_) and testing splits (_dev-test-sony_, _dev_test_tau_) should be combined (_dev_train_, _dev_test_) and the models should be trained and evaluated in the respective combined splits.\n\nThe evaluation part of the dataset will be published here as a new dataset version, a few weeks before the final challenge submission deadline. The additional evaluation files consist of only audio recordings without any metadata/labels. Participants can decide the training procedure, i.e. the amount of training and validation files in the development dataset, the number of ensemble models etc., and submit the results of the SELD performance on the evaluation dataset.", "# File structure", "# Download\n\ngit clone", "# Example application\n\nAn implementation of a trainable model of a convolutional recurrent neural network, performing joint SELD, trained and evaluated with this dataset is provided here. This implementation will serve as the baseline method in the DCASE 2022 Sound Event Localization and Detection Task.", "# License\n\nThis datast is licensed under the MIT license." ]

[ "license:mit", "region:us" ]

[]

[]

[ "## Synthesized voices from Project Echo on the Skyrim voice datasets." ]

[ "TAGS\n#license-mit #region-us \n", "## Synthesized voices from Project Echo on the Skyrim voice datasets." ]

[ "task_categories:text-generation", "annotations_creators:unknown", "language_creators:unknown", "multilinguality:monolingual", "size_categories:1k<n<10k", "language:en", "license:unknown", "keyphrase-generation", "keyphrase-extraction", "text-mining", "region:us" ]

[]

[ "en" ]

[]

[ "TAGS\n#task_categories-text-generation #annotations_creators-unknown #language_creators-unknown #multilinguality-monolingual #size_categories-1k<n<10k #language-English #license-unknown #keyphrase-generation #keyphrase-extraction #text-mining #region-us \n" ]

[ "task_categories:summarization", "task_ids:news-articles-summarization", "annotations_creators:machine-generated", "language_creators:machine-generated", "multilinguality:translation", "size_categories:1K<n<10K", "language:da", "license:apache-2.0", "region:us" ]

[]

[ "da" ]

[ "# Dataset Card for machine_translated_cnn_dailymail_da_small", "### Dataset Summary\n\nThis dataset is a machine translated subset of the CNN Dailymail Dataset into Danish. The dataset is translated using the Helsinki-NLP/opus-mt-en-da-model. The dataset consists of 2872 articles with summaries with intended usage for Danish text summarisation.", "## Dataset Structure\n\nMachine translated articles ('article') with corresponding summaries ('highlights').", "### Licensing Information\nThe dataset is released under the Apache-2.0 License." ]

[ "TAGS\n#task_categories-summarization #task_ids-news-articles-summarization #annotations_creators-machine-generated #language_creators-machine-generated #multilinguality-translation #size_categories-1K<n<10K #language-Danish #license-apache-2.0 #region-us \n", "# Dataset Card for machine_translated_cnn_dailymail_da_small", "### Dataset Summary\n\nThis dataset is a machine translated subset of the CNN Dailymail Dataset into Danish. The dataset is translated using the Helsinki-NLP/opus-mt-en-da-model. The dataset consists of 2872 articles with summaries with intended usage for Danish text summarisation.", "## Dataset Structure\n\nMachine translated articles ('article') with corresponding summaries ('highlights').", "### Licensing Information\nThe dataset is released under the Apache-2.0 License." ]

[ "region:us" ]

[]

[]

[ "### Languages\n\n\nThe BCP-47 code for the dataset's language is unk.\n\n\nDataset Structure\n-----------------", "### Data Instances\n\n\nA sample from this dataset looks as follows:", "### Dataset Fields\n\n\nThe dataset has the following fields (also called \"features\"):", "### Dataset Splits\n\n\nThis dataset is split into a train and validation split. The split sizes are as follow:" ]

[ "TAGS\n#region-us \n", "### Languages\n\n\nThe BCP-47 code for the dataset's language is unk.\n\n\nDataset Structure\n-----------------", "### Data Instances\n\n\nA sample from this dataset looks as follows:", "### Dataset Fields\n\n\nThe dataset has the following fields (also called \"features\"):", "### Dataset Splits\n\n\nThis dataset is split into a train and validation split. The split sizes are as follow:" ]

[ "region:us" ]

[]

[]

[]

[ "TAGS\n#region-us \n" ]

[ "task_categories:text-classification", "task_ids:sentiment-classification", "annotations_creators:found", "language_creators:found", "multilinguality:monolingual", "size_categories:1K<n<10K", "source_datasets:extended|cedr", "language:ru", "license:mit", "region:us" ]

[]

[ "ru" ]

[ "# Dataset Card for CEDR-M7", "## Table of Contents\n- Table of Contents\n- Dataset Description\n - Dataset Summary\n - Supported Tasks and Leaderboards\n - Languages\n- Dataset Structure\n - Data Instances\n - Data Fields\n - Data Splits\n- Dataset Creation\n - Curation Rationale\n - Source Data\n - Annotations\n - Personal and Sensitive Information\n- Considerations for Using the Data\n - Social Impact of Dataset\n - Discussion of Biases\n - Other Known Limitations\n- Additional Information\n - Dataset Curators\n - Licensing Information\n - Citation Information\n - Contributions", "## Dataset Description\n\n- Homepage:\n- Repository:\n- Paper:\n- Leaderboard:\n- Point of Contact:", "### Dataset Summary", "### Supported Tasks and Leaderboards", "### Languages", "## Dataset Structure", "### Data Instances", "### Data Fields", "### Data Splits", "## Dataset Creation", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information", "## Considerations for Using the Data", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations", "## Additional Information", "### Dataset Curators", "### Licensing Information", "### Contributions\n\nThanks to @toiletsandpaper for adding this dataset." ]

[ "TAGS\n#task_categories-text-classification #task_ids-sentiment-classification #annotations_creators-found #language_creators-found #multilinguality-monolingual #size_categories-1K<n<10K #source_datasets-extended|cedr #language-Russian #license-mit #region-us \n", "# Dataset Card for CEDR-M7", "## Table of Contents\n- Table of Contents\n- Dataset Description\n - Dataset Summary\n - Supported Tasks and Leaderboards\n - Languages\n- Dataset Structure\n - Data Instances\n - Data Fields\n - Data Splits\n- Dataset Creation\n - Curation Rationale\n - Source Data\n - Annotations\n - Personal and Sensitive Information\n- Considerations for Using the Data\n - Social Impact of Dataset\n - Discussion of Biases\n - Other Known Limitations\n- Additional Information\n - Dataset Curators\n - Licensing Information\n - Citation Information\n - Contributions", "## Dataset Description\n\n- Homepage:\n- Repository:\n- Paper:\n- Leaderboard:\n- Point of Contact:", "### Dataset Summary", "### Supported Tasks and Leaderboards", "### Languages", "## Dataset Structure", "### Data Instances", "### Data Fields", "### Data Splits", "## Dataset Creation", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information", "## Considerations for Using the Data", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations", "## Additional Information", "### Dataset Curators", "### Licensing Information", "### Contributions\n\nThanks to @toiletsandpaper for adding this dataset." ]

[ "task_categories:text-classification", "task_ids:sentiment-classification", "multilinguality:monolingual", "size_categories:10K<n<100K", "language:ru", "region:us" ]

[]

[ "ru" ]

[ "### Dataset Summary\nThe dataset contains user reviews about medical facilities.\n\nIn total it contains 70,597 reviews. The detailed distribution on sentiment scale is:\n- 41,419 positive reviews;\n- 29,178 negative reviews.", "### Data Fields\nEach sample contains the following fields:\n- review_id;\n- category category of medical facility (one of 48);\n- title: review title;\n- content: review text;\n- sentiment: sentiment (<em>positive</em> or <em>negative</em>);\n- source_url.", "### Python" ]

[ "TAGS\n#task_categories-text-classification #task_ids-sentiment-classification #multilinguality-monolingual #size_categories-10K<n<100K #language-Russian #region-us \n", "### Dataset Summary\nThe dataset contains user reviews about medical facilities.\n\nIn total it contains 70,597 reviews. The detailed distribution on sentiment scale is:\n- 41,419 positive reviews;\n- 29,178 negative reviews.", "### Data Fields\nEach sample contains the following fields:\n- review_id;\n- category category of medical facility (one of 48);\n- title: review title;\n- content: review text;\n- sentiment: sentiment (<em>positive</em> or <em>negative</em>);\n- source_url.", "### Python" ]

[ "license:mit", "region:us" ]

[]

[]

[]

[ "TAGS\n#license-mit #region-us \n" ]

[ "license:apache-2.0", "region:us" ]

[]

[]

[ "# Dataset Card for GTZAN Collection", "## Table of Contents\n- Dataset Description\n - Dataset Summary\n - Supported Tasks\n - Languages\n- Dataset Structure\n - Data Instances\n - Data Fields\n - Data Splits\n- Dataset Creation\n - Curation Rationale\n - Source Data\n - Annotations\n - Personal and Sensitive Information\n- Considerations for Using the Data\n - Social Impact of Dataset\n - Discussion of Biases\n - Other Known Limitations\n- Additional Information\n - Dataset Curators\n - Licensing Information\n - Citation Information", "## Dataset Description\n\n- Homepage: URL\n- Repository: URL\n- Paper: Musical genre classification of audio signals\n- Leaderboard: \n- Point of Contact:", "### Dataset Summary\n\nThe dataset consists of 1000 audio tracks each 30 seconds long.\nIt contains 10 genres, each represented by 100 tracks.\nThe tracks are all 22050Hz Mono 16-bit audio files in .wav format.\nThe genres are:\n* blues\n* classical\n* country\n* disco\n* hiphop\n* jazz\n* metal\n* pop\n* reggae\n* rock\n\nThis collection includes the following GTZAN variants:\n* raw (original WAV files)\n* melspectrograms (from each WAV file, contiguous 2-second windows at 4 random locations are sampled and transformed to Mel Spectrograms, resulting in 8000 Mel Spectrograms)", "### Supported Tasks and Leaderboards", "### Languages", "## Dataset Structure", "### Data Instances", "### Data Fields", "### Data Splits", "## Dataset Creation", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information", "## Considerations for Using the Data", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations", "## Additional Information", "### Dataset Curators", "### Licensing Information" ]

[ "TAGS\n#license-apache-2.0 #region-us \n", "# Dataset Card for GTZAN Collection", "## Table of Contents\n- Dataset Description\n - Dataset Summary\n - Supported Tasks\n - Languages\n- Dataset Structure\n - Data Instances\n - Data Fields\n - Data Splits\n- Dataset Creation\n - Curation Rationale\n - Source Data\n - Annotations\n - Personal and Sensitive Information\n- Considerations for Using the Data\n - Social Impact of Dataset\n - Discussion of Biases\n - Other Known Limitations\n- Additional Information\n - Dataset Curators\n - Licensing Information\n - Citation Information", "## Dataset Description\n\n- Homepage: URL\n- Repository: URL\n- Paper: Musical genre classification of audio signals\n- Leaderboard: \n- Point of Contact:", "### Dataset Summary\n\nThe dataset consists of 1000 audio tracks each 30 seconds long.\nIt contains 10 genres, each represented by 100 tracks.\nThe tracks are all 22050Hz Mono 16-bit audio files in .wav format.\nThe genres are:\n* blues\n* classical\n* country\n* disco\n* hiphop\n* jazz\n* metal\n* pop\n* reggae\n* rock\n\nThis collection includes the following GTZAN variants:\n* raw (original WAV files)\n* melspectrograms (from each WAV file, contiguous 2-second windows at 4 random locations are sampled and transformed to Mel Spectrograms, resulting in 8000 Mel Spectrograms)", "### Supported Tasks and Leaderboards", "### Languages", "## Dataset Structure", "### Data Instances", "### Data Fields", "### Data Splits", "## Dataset Creation", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information", "## Considerations for Using the Data", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations", "## Additional Information", "### Dataset Curators", "### Licensing Information" ]

[ "task_categories:text-classification", "task_ids:multi-class-classification", "multilinguality:monolingual", "language:en", "license:other", "region:us" ]

[]

[ "en" ]

[ "# Dataset Card for News_Articles_Categorization", "## Table of Contents\n- Dataset Description\n- Languages\n- Dataset Structure\n- Source Data", "## Dataset Description\n\n3722 News Articles classified into different categories namely: World, Politics, Tech, Entertainment, Sport, Business, Health, and Science", "## Languages\n\nThe text in the dataset is in English", "## Dataset Structure\n\nThe dataset consists of two columns namely Text and Category.\nThe Text column consists of the news article and the Category column consists of the class each article belongs to", "## Source Data\n\nThe dataset is scrapped across different news platforms" ]

[ "TAGS\n#task_categories-text-classification #task_ids-multi-class-classification #multilinguality-monolingual #language-English #license-other #region-us \n", "# Dataset Card for News_Articles_Categorization", "## Table of Contents\n- Dataset Description\n- Languages\n- Dataset Structure\n- Source Data", "## Dataset Description\n\n3722 News Articles classified into different categories namely: World, Politics, Tech, Entertainment, Sport, Business, Health, and Science", "## Languages\n\nThe text in the dataset is in English", "## Dataset Structure\n\nThe dataset consists of two columns namely Text and Category.\nThe Text column consists of the news article and the Category column consists of the class each article belongs to", "## Source Data\n\nThe dataset is scrapped across different news platforms" ]

[ "task_categories:token-classification", "task_ids:named-entity-recognition", "multilinguality:monolingual", "size_categories:10K<n<100K", "language:pt", "region:us" ]

[]

[ "pt" ]

[ "# Dataset Card for UlyssesNER-Br", "## Table of Contents\n- Table of Contents\n- Dataset Description\n - Dataset Summary\n - Supported Tasks and Leaderboards\n - Languages\n- Dataset Structure\n - Data Instances\n - Data Fields\n - Data Splits\n- Dataset Creation\n - Curation Rationale\n - Source Data\n - Annotations\n - Personal and Sensitive Information\n- Considerations for Using the Data\n - Social Impact of Dataset\n - Discussion of Biases\n - Other Known Limitations\n- Additional Information\n - Dataset Curators\n - Licensing Information\n - Citation Information\n - Contributions", "## Dataset Description\n\n- Homepage: Convenio-Camara-dos-Deputados/ulyssesner-br-propor\n- Repository: Convenio-Camara-dos-Deputados/ulyssesner-br-propor\n- Paper: UlyssesNER-Br: a Corpus of Brazilian Legislative Documents for Named Entity Recognition\n- Leaderboard:\n- Point of Contact:", "### Dataset Summary", "### Supported Tasks and Leaderboards", "### Languages\n\nPortuguese (Brazil).", "## Dataset Structure", "### Data Instances", "### Data Fields", "### Data Splits", "## Dataset Creation", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information", "## Considerations for Using the Data", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations", "## Additional Information", "### Dataset Curators", "### Licensing Information", "### Contributions\n\nThanks to @augusnunes for adding this dataset." ]

[ "TAGS\n#task_categories-token-classification #task_ids-named-entity-recognition #multilinguality-monolingual #size_categories-10K<n<100K #language-Portuguese #region-us \n", "# Dataset Card for UlyssesNER-Br", "## Table of Contents\n- Table of Contents\n- Dataset Description\n - Dataset Summary\n - Supported Tasks and Leaderboards\n - Languages\n- Dataset Structure\n - Data Instances\n - Data Fields\n - Data Splits\n- Dataset Creation\n - Curation Rationale\n - Source Data\n - Annotations\n - Personal and Sensitive Information\n- Considerations for Using the Data\n - Social Impact of Dataset\n - Discussion of Biases\n - Other Known Limitations\n- Additional Information\n - Dataset Curators\n - Licensing Information\n - Citation Information\n - Contributions", "## Dataset Description\n\n- Homepage: Convenio-Camara-dos-Deputados/ulyssesner-br-propor\n- Repository: Convenio-Camara-dos-Deputados/ulyssesner-br-propor\n- Paper: UlyssesNER-Br: a Corpus of Brazilian Legislative Documents for Named Entity Recognition\n- Leaderboard:\n- Point of Contact:", "### Dataset Summary", "### Supported Tasks and Leaderboards", "### Languages\n\nPortuguese (Brazil).", "## Dataset Structure", "### Data Instances", "### Data Fields", "### Data Splits", "## Dataset Creation", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information", "## Considerations for Using the Data", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations", "## Additional Information", "### Dataset Curators", "### Licensing Information", "### Contributions\n\nThanks to @augusnunes for adding this dataset." ]

[ "region:us" ]

[]

[]

[ "# Dataset Card for GamePhysics_Grand_Theft_Auto_V", "## Table of Contents\n- Dataset Description\n - Dataset Summary\n - Supported Tasks\n - Languages\n- Dataset Structure\n - Data Instances\n - Data Fields\n - Data Splits\n- Dataset Creation\n - Curation Rationale\n - Source Data\n - Annotations\n - Personal and Sensitive Information\n- Considerations for Using the Data\n - Social Impact of Dataset\n - Discussion of Biases\n - Other Known Limitations\n- Additional Information\n - Dataset Curators\n - Licensing Information\n - Citation Information", "## Dataset Description\n\n- Homepage: URL\n- Repository: URL\n- Paper: CLIP meets GamePhysics\n- Leaderboard: [N/A]\n- Point of Contact: Mohammad Reza Taesiri", "### Dataset Summary\n\nThe GamePhysics Grand Theft Auto V dataset is a small video dataset of buggy gameplay videos of Grand Theft Auto V game, collected from GamePhysics subrredit", "### Supported Tasks and Leaderboards\n\n[N/A]", "### Languages", "## Dataset Structure", "### Data Instances", "### Data Fields", "### Data Splits", "## Dataset Creation", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information", "## Considerations for Using the Data", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations", "## Additional Information", "### Dataset Curators", "### Licensing Information" ]

[ "TAGS\n#region-us \n", "# Dataset Card for GamePhysics_Grand_Theft_Auto_V", "## Table of Contents\n- Dataset Description\n - Dataset Summary\n - Supported Tasks\n - Languages\n- Dataset Structure\n - Data Instances\n - Data Fields\n - Data Splits\n- Dataset Creation\n - Curation Rationale\n - Source Data\n - Annotations\n - Personal and Sensitive Information\n- Considerations for Using the Data\n - Social Impact of Dataset\n - Discussion of Biases\n - Other Known Limitations\n- Additional Information\n - Dataset Curators\n - Licensing Information\n - Citation Information", "## Dataset Description\n\n- Homepage: URL\n- Repository: URL\n- Paper: CLIP meets GamePhysics\n- Leaderboard: [N/A]\n- Point of Contact: Mohammad Reza Taesiri", "### Dataset Summary\n\nThe GamePhysics Grand Theft Auto V dataset is a small video dataset of buggy gameplay videos of Grand Theft Auto V game, collected from GamePhysics subrredit", "### Supported Tasks and Leaderboards\n\n[N/A]", "### Languages", "## Dataset Structure", "### Data Instances", "### Data Fields", "### Data Splits", "## Dataset Creation", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information", "## Considerations for Using the Data", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations", "## Additional Information", "### Dataset Curators", "### Licensing Information" ]

[ "language:de", "language:en", "language:ja", "arxiv:2104.06893", "region:us" ]

[ "2104.06893" ]

[ "de", "en", "ja" ]

[ "# Amazon Multilingual Counterfactual Dataset \n\nThe dataset contains sentences from Amazon customer reviews (sampled from Amazon product review dataset) annotated for counterfactual detection (CFD) binary classification. Counterfactual statements describe events that did not or cannot take place. Counterfactual statements may be identified as statements of the form – If p was true, then q would be true (i.e. assertions whose antecedent (p) and consequent (q) are known or assumed to be false).\n\nThe key features of this dataset are:\n\n* The dataset is multilingual and contains sentences in English, German, and Japanese.\n* The labeling was done by professional linguists and high quality was ensured.\n* The dataset is supplemented with the annotation guidelines and definitions, which were worked out by professional linguists. We also provide the clue word lists, which are typical for counterfactual sentences and were used for initial data filtering. The clue word lists were also compiled by professional linguists.\n\nPlease see the paper for the data statistics, detailed description of data collection and annotation.\n\n\nGitHub repo URL: URL", "## Usage\n\nYou can load each of the languages as follows:" ]

[ "TAGS\n#language-German #language-English #language-Japanese #arxiv-2104.06893 #region-us \n", "# Amazon Multilingual Counterfactual Dataset \n\nThe dataset contains sentences from Amazon customer reviews (sampled from Amazon product review dataset) annotated for counterfactual detection (CFD) binary classification. Counterfactual statements describe events that did not or cannot take place. Counterfactual statements may be identified as statements of the form – If p was true, then q would be true (i.e. assertions whose antecedent (p) and consequent (q) are known or assumed to be false).\n\nThe key features of this dataset are:\n\n* The dataset is multilingual and contains sentences in English, German, and Japanese.\n* The labeling was done by professional linguists and high quality was ensured.\n* The dataset is supplemented with the annotation guidelines and definitions, which were worked out by professional linguists. We also provide the clue word lists, which are typical for counterfactual sentences and were used for initial data filtering. The clue word lists were also compiled by professional linguists.\n\nPlease see the paper for the data statistics, detailed description of data collection and annotation.\n\n\nGitHub repo URL: URL", "## Usage\n\nYou can load each of the languages as follows:" ]

[ "region:us" ]

[]

[]

[]

[ "TAGS\n#region-us \n" ]

[ "license:afl-3.0", "region:us" ]

[]

[]

[]

[ "TAGS\n#license-afl-3.0 #region-us \n" ]

[ "region:us" ]

[]

[]

[]

[ "TAGS\n#region-us \n" ]

[ "task_categories:text-classification", "task_categories:text-generation", "task_categories:text2text-generation", "task_categories:sentence-similarity", "task_ids:semantic-similarity-scoring", "annotations_creators:crowdsourced", "annotations_creators:expert-generated", "annotations_creators:machine-generated", "language_creators:crowdsourced", "multilinguality:monolingual", "size_categories:1M<n<10M", "source_datasets:original", "language:ru", "license:mit", "region:us" ]

[]

[ "ru" ]

[ "# Dataset Card for ParaPhraser", "### Dataset Summary\n\nParaPhraser is a news headlines corpus annotated according to the following schema:\n\n\nThe _Plus_ part is also available.\nIt contains clusters of news headline paraphrases labeled automatically by a fine-tuned paraphrase detection BERT model. \nIn order to load it:", "## Dataset Structure", "### Contributions\n\nDataset maintainer:\nVadim Gudkov: @merionum" ]

[ "TAGS\n#task_categories-text-classification #task_categories-text-generation #task_categories-text2text-generation #task_categories-sentence-similarity #task_ids-semantic-similarity-scoring #annotations_creators-crowdsourced #annotations_creators-expert-generated #annotations_creators-machine-generated #language_creators-crowdsourced #multilinguality-monolingual #size_categories-1M<n<10M #source_datasets-original #language-Russian #license-mit #region-us \n", "# Dataset Card for ParaPhraser", "### Dataset Summary\n\nParaPhraser is a news headlines corpus annotated according to the following schema:\n\n\nThe _Plus_ part is also available.\nIt contains clusters of news headline paraphrases labeled automatically by a fine-tuned paraphrase detection BERT model. \nIn order to load it:", "## Dataset Structure", "### Contributions\n\nDataset maintainer:\nVadim Gudkov: @merionum" ]

[ "language:en", "region:us" ]

[]

[ "en" ]

[ "# Toxic Conversation\nThis is a version of the Jigsaw Unintended Bias in Toxicity Classification dataset. It contains comments from the Civil Comments platform together with annotations if the comment is toxic or not.\n\nThis dataset just contains the first 50k training examples.\n\n10 annotators annotated each example and, as recommended in the task page, set a comment as toxic when target >= 0.5\n\nThe dataset is inbalanced, with only about 8% of the comments marked as toxic." ]

[ "TAGS\n#language-English #region-us \n", "# Toxic Conversation\nThis is a version of the Jigsaw Unintended Bias in Toxicity Classification dataset. It contains comments from the Civil Comments platform together with annotations if the comment is toxic or not.\n\nThis dataset just contains the first 50k training examples.\n\n10 annotators annotated each example and, as recommended in the task page, set a comment as toxic when target >= 0.5\n\nThe dataset is inbalanced, with only about 8% of the comments marked as toxic." ]

[ "task_categories:audio-classification", "task_ids:audio-emotion-recognition", "annotations_creators:crowdsourced", "language_creators:expert-generated", "language_creators:crowdsourced", "multilinguality:monolingual", "size_categories:1K<n<10K", "source_datasets:original", "language:ru", "license:mit", "region:us" ]

[]

[ "ru" ]

[]

[ "TAGS\n#task_categories-audio-classification #task_ids-audio-emotion-recognition #annotations_creators-crowdsourced #language_creators-expert-generated #language_creators-crowdsourced #multilinguality-monolingual #size_categories-1K<n<10K #source_datasets-original #language-Russian #license-mit #region-us \n" ]

[ "task_categories:audio-classification", "task_ids:audio-emotion-recognition", "annotations_creators:crowdsourced", "language_creators:expert-generated", "language_creators:crowdsourced", "multilinguality:monolingual", "size_categories:1K<n<10K", "source_datasets:original", "language:ru", "license:mit", "region:us" ]

[]

[ "ru" ]

[]

[ "TAGS\n#task_categories-audio-classification #task_ids-audio-emotion-recognition #annotations_creators-crowdsourced #language_creators-expert-generated #language_creators-crowdsourced #multilinguality-monolingual #size_categories-1K<n<10K #source_datasets-original #language-Russian #license-mit #region-us \n" ]

[ "language_creators:Digital Umuganda", "size_categories:3K<n<4K", "size_categories:~6hours", "language:rw", "license:cc-by-4.0", "region:us" ]

[]

[ "rw" ]

[ "# Kinyarwanda TTS dataset\n\nThe dataset consists of 3992 clips of Kinyarwanda TTS corpus recorded in a studio using a voice actress, it was collected in the mbaza project", "## Data structure", "## Language\nThe corresponding dataset is in the Kinyarwanda Language", "## Dataset Creation\n- Text collected had to include Kinyarwanda syllabes, which is made by a combination of a consonant or a group of consonats (e.g. Nyw) and a vowel.\n- Text were reviewed by a linguist to ensure the text fit kinyarwanda standards\n- The voice were recorded in a studio albeit in a semi-professional settings (i.e. some of the audio contains reverbs)" ]

[ "TAGS\n#language_creators-Digital Umuganda #size_categories-3K<n<4K #size_categories-~6hours #language-Kinyarwanda #license-cc-by-4.0 #region-us \n", "# Kinyarwanda TTS dataset\n\nThe dataset consists of 3992 clips of Kinyarwanda TTS corpus recorded in a studio using a voice actress, it was collected in the mbaza project", "## Data structure", "## Language\nThe corresponding dataset is in the Kinyarwanda Language", "## Dataset Creation\n- Text collected had to include Kinyarwanda syllabes, which is made by a combination of a consonant or a group of consonats (e.g. Nyw) and a vowel.\n- Text were reviewed by a linguist to ensure the text fit kinyarwanda standards\n- The voice were recorded in a studio albeit in a semi-professional settings (i.e. some of the audio contains reverbs)" ]

[ "task_categories:multiple-choice", "task_categories:question-answering", "task_ids:multiple-choice-qa", "task_ids:open-domain-qa", "annotations_creators:expert-generated", "language_creators:crowdsourced", "multilinguality:monolingual", "size_categories:n<1K", "source_datasets:original", "language:en", "license:apache-2.0", "movie-recommendation", "collaborative-filtering", "movielens", "film", "doi:10.57967/hf/0257", "region:us" ]

[]

[ "en" ]

[ "# Dataset for evaluation of (zero-shot) recommendation with language models\n\nWe showed that pretrained large language models can act as a recommender system, and compare few-shot learning results to matrix factorization baselines.\nThis is the BIG-Bench version of our language-based movie recommendation dataset.\n\n<URL\n\nGPT-2 has a 48.8% accuracy, chance is 25%.\nHuman accuracy is 60.4%." ]

[ "TAGS\n#task_categories-multiple-choice #task_categories-question-answering #task_ids-multiple-choice-qa #task_ids-open-domain-qa #annotations_creators-expert-generated #language_creators-crowdsourced #multilinguality-monolingual #size_categories-n<1K #source_datasets-original #language-English #license-apache-2.0 #movie-recommendation #collaborative-filtering #movielens #film #doi-10.57967/hf/0257 #region-us \n", "# Dataset for evaluation of (zero-shot) recommendation with language models\n\nWe showed that pretrained large language models can act as a recommender system, and compare few-shot learning results to matrix factorization baselines.\nThis is the BIG-Bench version of our language-based movie recommendation dataset.\n\n<URL\n\nGPT-2 has a 48.8% accuracy, chance is 25%.\nHuman accuracy is 60.4%." ]

[ "task_categories:question-answering", "task_categories:multiple-choice", "task_ids:open-domain-qa", "task_ids:multiple-choice-qa", "annotations_creators:expert-generated", "language_creators:crowdsourced", "multilinguality:monolingual", "size_categories:n<1K", "source_datasets:original", "language:en", "license:apache-2.0", "region:us" ]

[]

[ "en" ]

[ "# Dataset for evaluation of (zero-shot) discourse marker prediction with language models\n\nThis is the Big-Bench version of our discourse marker prediction dataset, Discovery\n\nDesign considerations:\n<URL\n\nGPT2 has to zero-shot 15% accuracy with on this multiple-choice task based on language modeling perplexity. As a comparison, a fully supervised model, trained with 10k examples per marker with ROBERTA and default hyperparameters with one epoch, leads to an accuracy of 30% with 174 possible markers. This shows that this task is hard for GPT2 and that the model didn't memorize the discourse markers, but that high accuracies are still possible." ]

[ "TAGS\n#task_categories-question-answering #task_categories-multiple-choice #task_ids-open-domain-qa #task_ids-multiple-choice-qa #annotations_creators-expert-generated #language_creators-crowdsourced #multilinguality-monolingual #size_categories-n<1K #source_datasets-original #language-English #license-apache-2.0 #region-us \n", "# Dataset for evaluation of (zero-shot) discourse marker prediction with language models\n\nThis is the Big-Bench version of our discourse marker prediction dataset, Discovery\n\nDesign considerations:\n<URL\n\nGPT2 has to zero-shot 15% accuracy with on this multiple-choice task based on language modeling perplexity. As a comparison, a fully supervised model, trained with 10k examples per marker with ROBERTA and default hyperparameters with one epoch, leads to an accuracy of 30% with 174 possible markers. This shows that this task is hard for GPT2 and that the model didn't memorize the discourse markers, but that high accuracies are still possible." ]

[ "task_categories:text-classification", "task_ids:sentiment-classification", "multilinguality:monolingual", "size_categories:10K<n<100K", "language:ru", "region:us" ]

[]

[ "ru" ]

[ "### Dataset Summary\nThe dataset contains user reviews about medical institutions.\n\nIn total it contains 12,036 reviews. A review tagged with the <em>general</em> sentiment and sentiments on 5 aspects: <em>quality, service, equipment, food, location</em>.", "### Data Fields\nEach sample contains the following fields:\n- review_id;\n- content: review text;\n- general;\n- quality;\n- service;\n- equipment;\n- food;\n- location.", "### Python" ]

[ "TAGS\n#task_categories-text-classification #task_ids-sentiment-classification #multilinguality-monolingual #size_categories-10K<n<100K #language-Russian #region-us \n", "### Dataset Summary\nThe dataset contains user reviews about medical institutions.\n\nIn total it contains 12,036 reviews. A review tagged with the <em>general</em> sentiment and sentiments on 5 aspects: <em>quality, service, equipment, food, location</em>.", "### Data Fields\nEach sample contains the following fields:\n- review_id;\n- content: review text;\n- general;\n- quality;\n- service;\n- equipment;\n- food;\n- location.", "### Python" ]

[ "task_categories:text-classification", "language:ar", "region:us" ]

[]

[ "ar" ]

[ "### Languages\n\n\nThe BCP-47 code for the dataset's language is ar.\n\n\nDataset Structure\n-----------------", "### Data Instances\n\n\nA sample from this dataset looks as follows:", "### Dataset Fields\n\n\nThe dataset has the following fields (also called \"features\"):", "### Dataset Splits\n\n\nThis dataset is split into a train and validation split. The split sizes are as follow:" ]

[ "TAGS\n#task_categories-text-classification #language-Arabic #region-us \n", "### Languages\n\n\nThe BCP-47 code for the dataset's language is ar.\n\n\nDataset Structure\n-----------------", "### Data Instances\n\n\nA sample from this dataset looks as follows:", "### Dataset Fields\n\n\nThe dataset has the following fields (also called \"features\"):", "### Dataset Splits\n\n\nThis dataset is split into a train and validation split. The split sizes are as follow:" ]

[ "region:us" ]

[]

[]

[]

[ "TAGS\n#region-us \n" ]

[ "language:de", "license:other", "region:us" ]

[]

[ "de" ]

[ "# Inclusive words in German ️‍ 🇩🇪\n\nPairs of words and phrases in exclusive language and alternative words and phrases in inclusive language.\n\nInclusivity aims to comprehend all dimensions of diversity (age, ethnic background and nationality, gender and gender identity, physical and mental abilities, religion and worldview, sexual orientation, social background, and more); but currently focuses almost exclusively on gender inclusion, since gender exclusion is very dominant in German language.", "## Dataset structure\n\nTrain/test split: There is no train/test split, just a \"train\" dataset.\n\n- 'exclusive': Exclusive words and phrases in the singular. For the dimension of gender, these are certain words and phrases in the grammatical masculine. Note that the grammatical masculine is only exclusive if it is used in a _generic_ sense: \"Die Doktoren\" may be accurately used to describe three male doctors, but the same phrase is exclusive when it intends to refer to a group that also (potentially) includes women and nonbinary people. The relation between exclusive and inclusive phrases is n-to-n: An exclusive phrase may occur in multiple rows with various inclusive phrases associated, and vice versa.\n\n- 'inclusive': Corresponding inclusive word or phrase that can replace the exclusive phrase. It may be applicable only in a certain context and not in others. Usually in the singular; where 'number' is plural, it may be either in the singular or plural. The relation between exclusive and inclusive phrases is n-to-n: An inclusive phrase may occur in multiple rows with various exclusive phrases associated, and vice versa.\n\n- 'applicable': One of 'in_singular', 'in_plural', or 'always'. Specifies the grammatical number that the inclusive phrase must be found in such that it can be replaced by the inclusive phrase given in this entry.\n\n - _Special case:_ Some singular words (such as \"Management\" as a replacement for \"Manager\") occur in two rows, once with the attribute 'always', once with the attribute 'plural'. The first means that \"Manager\"(singular) can be replaced with \"Management\" (singular) and \"Manager\" (plural) can be replaced with \"Managements\" (plural); the second means that \"Manager\" (plural) can (also) be replaced with \"Management\" (singular).\n\n- 'gender_of_inclusive': Whether the inclusive phrase is semantically 'neutral' or 'female'. If it is female, it is not by itself inclusive but has to be combined with the male phrase (and potentially a character such as the gender star for representing nonbinary persons) to form a neutral phrase. (Since the male phrase is already given by the 'exclusive' column, it is not repeated in the 'inclusive' column due to potentially questionable ideological beliefs about data normalization.)\n\n- 'source': The origin of the entry.\n\n - _geschicktgendern_: The entry has been copied from the _Genderwörterbuch_ by _Geschickt Gendern_. These entries are under a CC-BY-NC-SA 4.0 International License (c) Johanna Usinger, URL.\n\n - _dereko_: The entry has been extracted from the German reference corpus DeReKo. Since these are single words only, copyright does not apply and the entries are under the CC-0 license.\n\n - _diversifix_: Entries added by ourselves or our community, also under the CC-0 license.", "## Bias\n\nThe entries from the 'dereko' source have been extracted according to their frequency in the corpus. This means, for example, that there are words referring to people from larger countries but not from some smaller countries; or, more accurately, countries that are considered important from the perspective of German-speaking journalism are more prevalent in the dataset.", "## License\n\nMixed license. All data is open, but a part of it only noncommercially. See the description for the 'source' column above for details.", "## See also\n\n- Other data sources on inclusive German.\n- retext-equality ️‍ 🇬🇧" ]

[ "TAGS\n#language-German #license-other #region-us \n", "# Inclusive words in German ️‍ 🇩🇪\n\nPairs of words and phrases in exclusive language and alternative words and phrases in inclusive language.\n\nInclusivity aims to comprehend all dimensions of diversity (age, ethnic background and nationality, gender and gender identity, physical and mental abilities, religion and worldview, sexual orientation, social background, and more); but currently focuses almost exclusively on gender inclusion, since gender exclusion is very dominant in German language.", "## Dataset structure\n\nTrain/test split: There is no train/test split, just a \"train\" dataset.\n\n- 'exclusive': Exclusive words and phrases in the singular. For the dimension of gender, these are certain words and phrases in the grammatical masculine. Note that the grammatical masculine is only exclusive if it is used in a _generic_ sense: \"Die Doktoren\" may be accurately used to describe three male doctors, but the same phrase is exclusive when it intends to refer to a group that also (potentially) includes women and nonbinary people. The relation between exclusive and inclusive phrases is n-to-n: An exclusive phrase may occur in multiple rows with various inclusive phrases associated, and vice versa.\n\n- 'inclusive': Corresponding inclusive word or phrase that can replace the exclusive phrase. It may be applicable only in a certain context and not in others. Usually in the singular; where 'number' is plural, it may be either in the singular or plural. The relation between exclusive and inclusive phrases is n-to-n: An inclusive phrase may occur in multiple rows with various exclusive phrases associated, and vice versa.\n\n- 'applicable': One of 'in_singular', 'in_plural', or 'always'. Specifies the grammatical number that the inclusive phrase must be found in such that it can be replaced by the inclusive phrase given in this entry.\n\n - _Special case:_ Some singular words (such as \"Management\" as a replacement for \"Manager\") occur in two rows, once with the attribute 'always', once with the attribute 'plural'. The first means that \"Manager\"(singular) can be replaced with \"Management\" (singular) and \"Manager\" (plural) can be replaced with \"Managements\" (plural); the second means that \"Manager\" (plural) can (also) be replaced with \"Management\" (singular).\n\n- 'gender_of_inclusive': Whether the inclusive phrase is semantically 'neutral' or 'female'. If it is female, it is not by itself inclusive but has to be combined with the male phrase (and potentially a character such as the gender star for representing nonbinary persons) to form a neutral phrase. (Since the male phrase is already given by the 'exclusive' column, it is not repeated in the 'inclusive' column due to potentially questionable ideological beliefs about data normalization.)\n\n- 'source': The origin of the entry.\n\n - _geschicktgendern_: The entry has been copied from the _Genderwörterbuch_ by _Geschickt Gendern_. These entries are under a CC-BY-NC-SA 4.0 International License (c) Johanna Usinger, URL.\n\n - _dereko_: The entry has been extracted from the German reference corpus DeReKo. Since these are single words only, copyright does not apply and the entries are under the CC-0 license.\n\n - _diversifix_: Entries added by ourselves or our community, also under the CC-0 license.", "## Bias\n\nThe entries from the 'dereko' source have been extracted according to their frequency in the corpus. This means, for example, that there are words referring to people from larger countries but not from some smaller countries; or, more accurately, countries that are considered important from the perspective of German-speaking journalism are more prevalent in the dataset.", "## License\n\nMixed license. All data is open, but a part of it only noncommercially. See the description for the 'source' column above for details.", "## See also\n\n- Other data sources on inclusive German.\n- retext-equality ️‍ 🇬🇧" ]

[ "license:apache-2.0", "arxiv:2208.03219", "arxiv:2209.09450", "region:us" ]

[ "2208.03219", "2209.09450" ]

[]

[]

[ "TAGS\n#license-apache-2.0 #arxiv-2208.03219 #arxiv-2209.09450 #region-us \n" ]

[ "task_categories:conversational", "task_ids:dialogue-generation", "annotations_creators:other", "language_creators:other", "multilinguality:monolingual", "size_categories:10M<n<100M", "source_datasets:original", "language:zh", "license:mit", "dialogue-response-retrieval", "arxiv:2008.03946", "region:us" ]

[ "2008.03946" ]

[ "zh" ]

[ "### Dataset Summary\n\n\nlccc: Large-scale Cleaned Chinese Conversation corpus (LCCC) is a large Chinese dialogue corpus originate from Chinese social medias. A rigorous data cleaning pipeline is designed to ensure the quality of the corpus. This pipeline involves a set of rules and several classifier-based filters. Noises such as offensive or sensitive words, special symbols, emojis, grammatically incorrect sentences, and incoherent conversations are filtered.\n\n\nlccc是一套来自于中文社交媒体的对话数据，我们设计了一套严格的数据过滤流程来确保该数据集中对话数据的质量。 这一数据过滤流程中包括一系列手工规则以及若干基于机器学习算法所构建的分类器。 我们所过滤掉的噪声包括：脏字脏词、特殊字符、颜表情、语法不通的语句、上下文不相关的对话等。", "### Supported Tasks and Leaderboards\n\n\n* dialogue-generation: The dataset can be used to train a model for generating dialogue responses.\n* response-retrieval: The dataset can be used to train a reranker model that can be used to implement a retrieval-based dialogue model.", "### Languages\n\n\nLCCC is in Chinese\n\n\nLCCC中的对话是中文的\n\n\nDataset Structure\n-----------------", "### Data Instances\n\n\n[\"火锅 我 在 重庆 成都 吃 了 七八 顿 火锅\", \"哈哈哈哈 ！ 那 我 的 嘴巴 可能 要 烂掉 ！\", \"不会 的 就是 好 油腻\"]", "### Data Fields\n\n\nEach line is a list of utterances that consist a dialogue.\nNote that the LCCC dataset provided in our original Github page is in json format,\nhowever, we are providing LCCC in jsonl format here.", "### Data Splits\n\n\nWe do not provide the offical split for LCCC-large.\nBut we provide a split for LCCC-base:\n\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations\n\n\nAdditional Information\n----------------------", "### Dataset Curators", "### Licensing Information\n\n\nPlease cite the following paper if you find this dataset useful:" ]

[ "TAGS\n#task_categories-conversational #task_ids-dialogue-generation #annotations_creators-other #language_creators-other #multilinguality-monolingual #size_categories-10M<n<100M #source_datasets-original #language-Chinese #license-mit #dialogue-response-retrieval #arxiv-2008.03946 #region-us \n", "### Dataset Summary\n\n\nlccc: Large-scale Cleaned Chinese Conversation corpus (LCCC) is a large Chinese dialogue corpus originate from Chinese social medias. A rigorous data cleaning pipeline is designed to ensure the quality of the corpus. This pipeline involves a set of rules and several classifier-based filters. Noises such as offensive or sensitive words, special symbols, emojis, grammatically incorrect sentences, and incoherent conversations are filtered.\n\n\nlccc是一套来自于中文社交媒体的对话数据，我们设计了一套严格的数据过滤流程来确保该数据集中对话数据的质量。 这一数据过滤流程中包括一系列手工规则以及若干基于机器学习算法所构建的分类器。 我们所过滤掉的噪声包括：脏字脏词、特殊字符、颜表情、语法不通的语句、上下文不相关的对话等。", "### Supported Tasks and Leaderboards\n\n\n* dialogue-generation: The dataset can be used to train a model for generating dialogue responses.\n* response-retrieval: The dataset can be used to train a reranker model that can be used to implement a retrieval-based dialogue model.", "### Languages\n\n\nLCCC is in Chinese\n\n\nLCCC中的对话是中文的\n\n\nDataset Structure\n-----------------", "### Data Instances\n\n\n[\"火锅 我 在 重庆 成都 吃 了 七八 顿 火锅\", \"哈哈哈哈 ！ 那 我 的 嘴巴 可能 要 烂掉 ！\", \"不会 的 就是 好 油腻\"]", "### Data Fields\n\n\nEach line is a list of utterances that consist a dialogue.\nNote that the LCCC dataset provided in our original Github page is in json format,\nhowever, we are providing LCCC in jsonl format here.", "### Data Splits\n\n\nWe do not provide the offical split for LCCC-large.\nBut we provide a split for LCCC-base:\n\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations\n\n\nAdditional Information\n----------------------", "### Dataset Curators", "### Licensing Information\n\n\nPlease cite the following paper if you find this dataset useful:" ]

[ "task_categories:conversational", "task_ids:dialogue-generation", "annotations_creators:no-annotation", "language_creators:found", "multilinguality:monolingual", "size_categories:10M<n<100M", "source_datasets:original", "language:zh", "license:other", "arxiv:2108.07154", "arxiv:2008.03946", "region:us" ]

[ "2108.07154", "2008.03946" ]

[ "zh" ]

[ "### Dataset Summary\n\n\nMMChat is a large-scale dialogue dataset that contains image-grounded dialogues in Chinese. Each dialogue in MMChat is associated with one or more images (maximum 9 images per dialogue). We design various strategies to ensure the quality of the dialogues in MMChat.\n\n\nMMChat comes with 4 different versions:\n\n\n* 'mmchat': The MMChat dataset used in our paper.\n* 'mmchat\\_hf': Contains human annotation on 100K sessions of dialogues.\n* 'mmchat\\_raw': Raw dialogues used to construct MMChat.\n'mmchat\\_lccc\\_filtered': Raw dialogues filtered using the LCCC dataset.\n\n\nIf you what to use high quality multi-modal dialogues that are closed related to the given images, I suggest you to use the 'mmchat\\_hf' version.\nIf you only care about the quality of dialogue texts, I suggest you to use the 'mmchat\\_lccc\\_filtered' version.", "### Supported Tasks and Leaderboards\n\n\n* dialogue-generation: The dataset can be used to train a model for generating dialogue responses.\n* response-retrieval: The dataset can be used to train a reranker model that can be used to implement a retrieval-based dialogue model.", "### Languages\n\n\nMMChat is in Chinese\n\n\nMMChat中的对话是中文的\n\n\nDataset Structure\n-----------------", "### Data Instances\n\n\nSeveral versions of MMChat are available. For 'mmchat', 'mmchat\\_raw', 'mmchat\\_lccc\\_filtered', the following instance applies:\n\n\nFor 'mmchat\\_hf', the following instance applies:", "### Data Fields\n\n\n* 'dialog' (list of strings): List of utterances consisting of a dialogue.\n* 'weibo\\_content' (string): Weibo content of the dialogue.\n* 'imgs' (list of strings): List of URLs of images.\n* 'labels' (dict): Human-annotated labels of the dialogue.\n* 'image\\_qualified' (bool): Whether the image is of high quality.\n* 'dialog\\_qualified' (bool): Whether the dialogue is of high quality.\n* 'dialog\\_image\\_related' (bool): Whether the dialogue is related to the image.", "### Data Splits\n\n\nFor 'mmchat', we provide the following splits:\n\n\n\nFor other versions, we do not provide the offical split.\nMore stastics are listed here:\n\n\n\n\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations\n\n\nAdditional Information\n----------------------", "### Dataset Curators", "### Licensing Information\n\n\nother-weibo\n\n\nThis dataset is collected from Weibo.\nYou can refer to the detailed policy required to use this dataset.\nPlease restrict the usage of this dataset to non-commerical purposes.", "### Contributions\n\n\nThanks to Yinhe Zheng for adding this dataset." ]

[ "TAGS\n#task_categories-conversational #task_ids-dialogue-generation #annotations_creators-no-annotation #language_creators-found #multilinguality-monolingual #size_categories-10M<n<100M #source_datasets-original #language-Chinese #license-other #arxiv-2108.07154 #arxiv-2008.03946 #region-us \n", "### Dataset Summary\n\n\nMMChat is a large-scale dialogue dataset that contains image-grounded dialogues in Chinese. Each dialogue in MMChat is associated with one or more images (maximum 9 images per dialogue). We design various strategies to ensure the quality of the dialogues in MMChat.\n\n\nMMChat comes with 4 different versions:\n\n\n* 'mmchat': The MMChat dataset used in our paper.\n* 'mmchat\\_hf': Contains human annotation on 100K sessions of dialogues.\n* 'mmchat\\_raw': Raw dialogues used to construct MMChat.\n'mmchat\\_lccc\\_filtered': Raw dialogues filtered using the LCCC dataset.\n\n\nIf you what to use high quality multi-modal dialogues that are closed related to the given images, I suggest you to use the 'mmchat\\_hf' version.\nIf you only care about the quality of dialogue texts, I suggest you to use the 'mmchat\\_lccc\\_filtered' version.", "### Supported Tasks and Leaderboards\n\n\n* dialogue-generation: The dataset can be used to train a model for generating dialogue responses.\n* response-retrieval: The dataset can be used to train a reranker model that can be used to implement a retrieval-based dialogue model.", "### Languages\n\n\nMMChat is in Chinese\n\n\nMMChat中的对话是中文的\n\n\nDataset Structure\n-----------------", "### Data Instances\n\n\nSeveral versions of MMChat are available. For 'mmchat', 'mmchat\\_raw', 'mmchat\\_lccc\\_filtered', the following instance applies:\n\n\nFor 'mmchat\\_hf', the following instance applies:", "### Data Fields\n\n\n* 'dialog' (list of strings): List of utterances consisting of a dialogue.\n* 'weibo\\_content' (string): Weibo content of the dialogue.\n* 'imgs' (list of strings): List of URLs of images.\n* 'labels' (dict): Human-annotated labels of the dialogue.\n* 'image\\_qualified' (bool): Whether the image is of high quality.\n* 'dialog\\_qualified' (bool): Whether the dialogue is of high quality.\n* 'dialog\\_image\\_related' (bool): Whether the dialogue is related to the image.", "### Data Splits\n\n\nFor 'mmchat', we provide the following splits:\n\n\n\nFor other versions, we do not provide the offical split.\nMore stastics are listed here:\n\n\n\n\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations\n\n\nAdditional Information\n----------------------", "### Dataset Curators", "### Licensing Information\n\n\nother-weibo\n\n\nThis dataset is collected from Weibo.\nYou can refer to the detailed policy required to use this dataset.\nPlease restrict the usage of this dataset to non-commerical purposes.", "### Contributions\n\n\nThanks to Yinhe Zheng for adding this dataset." ]

[ "task_categories:conversational", "task_ids:dialogue-generation", "annotations_creators:no-annotation", "language_creators:found", "multilinguality:monolingual", "size_categories:10M<n<100M", "source_datasets:original", "language:zh", "license:other", "arxiv:1901.09672", "region:us" ]

[ "1901.09672" ]

[ "zh" ]

[ "### Dataset Summary\n\n\nThe PersonalDialog dataset is a large-scale multi-turn Chinese dialogue dataset containing various traits from a large number of speakers.\nWe are releasing about 5M sessions of carefully filtered dialogues.\nEach utterance in PersonalDialog is associated with a speaker marked with traits like Gender, Location, Interest Tags.", "### Supported Tasks and Leaderboards\n\n\n* dialogue-generation: The dataset can be used to train a model for generating dialogue responses.\n* response-retrieval: The dataset can be used to train a reranker model that can be used to implement a retrieval-based dialogue model.", "### Languages\n\n\nPersonalDialog is in Chinese\n\n\nPersonalDialog中的对话是中文的\n\n\nDataset Structure\n-----------------", "### Data Instances\n\n\n'train' split:\n\n\n'dev' and 'test' split:", "### Data Fields\n\n\n* 'dialog' (list of strings): List of utterances consisting of a dialogue.\n* 'profile' (list of dicts): List of profiles associated with each speaker.\n* 'tag' (list of strings): List of tags associated with each speaker.\n* 'loc' (string): Location of each speaker.\n* 'gender' (string): Gender of each speaker.\n* 'uid' (list of int): Speaker id for each utterance in the dialogue.\n* 'responder\\_profile' (dict): Profile of the responder. (Only available in 'dev' and 'test' split)\n* 'golden\\_response' (str): Response of the responder. (Only available in 'dev' and 'test' split)\n* 'id\\_biased' (bool): Whether the dialogue is guranteed to be persona related or not. (Only available in 'dev' and 'test' split)", "### Data Splits\n\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations\n\n\nAdditional Information\n----------------------", "### Dataset Curators", "### Licensing Information\n\n\nother-weibo\n\n\nThis dataset is collected from Weibo.\nYou can refer to the detailed policy required to use this dataset.\nPlease restrict the usage of this dataset to non-commerical purposes.", "### Contributions\n\n\nThanks to Yinhe Zheng for adding this dataset." ]

[ "TAGS\n#task_categories-conversational #task_ids-dialogue-generation #annotations_creators-no-annotation #language_creators-found #multilinguality-monolingual #size_categories-10M<n<100M #source_datasets-original #language-Chinese #license-other #arxiv-1901.09672 #region-us \n", "### Dataset Summary\n\n\nThe PersonalDialog dataset is a large-scale multi-turn Chinese dialogue dataset containing various traits from a large number of speakers.\nWe are releasing about 5M sessions of carefully filtered dialogues.\nEach utterance in PersonalDialog is associated with a speaker marked with traits like Gender, Location, Interest Tags.", "### Supported Tasks and Leaderboards\n\n\n* dialogue-generation: The dataset can be used to train a model for generating dialogue responses.\n* response-retrieval: The dataset can be used to train a reranker model that can be used to implement a retrieval-based dialogue model.", "### Languages\n\n\nPersonalDialog is in Chinese\n\n\nPersonalDialog中的对话是中文的\n\n\nDataset Structure\n-----------------", "### Data Instances\n\n\n'train' split:\n\n\n'dev' and 'test' split:", "### Data Fields\n\n\n* 'dialog' (list of strings): List of utterances consisting of a dialogue.\n* 'profile' (list of dicts): List of profiles associated with each speaker.\n* 'tag' (list of strings): List of tags associated with each speaker.\n* 'loc' (string): Location of each speaker.\n* 'gender' (string): Gender of each speaker.\n* 'uid' (list of int): Speaker id for each utterance in the dialogue.\n* 'responder\\_profile' (dict): Profile of the responder. (Only available in 'dev' and 'test' split)\n* 'golden\\_response' (str): Response of the responder. (Only available in 'dev' and 'test' split)\n* 'id\\_biased' (bool): Whether the dialogue is guranteed to be persona related or not. (Only available in 'dev' and 'test' split)", "### Data Splits\n\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations\n\n\nAdditional Information\n----------------------", "### Dataset Curators", "### Licensing Information\n\n\nother-weibo\n\n\nThis dataset is collected from Weibo.\nYou can refer to the detailed policy required to use this dataset.\nPlease restrict the usage of this dataset to non-commerical purposes.", "### Contributions\n\n\nThanks to Yinhe Zheng for adding this dataset." ]

[ "task_categories:summarization", "annotations_creators:crowd-sourced", "language_creators:unknown", "multilinguality:unknown", "size_categories:unknown", "source_datasets:original", "language:en", "license:cc-by-4.0", "arxiv:2205.11465", "arxiv:2112.07637", "arxiv:2104.05938", "region:us" ]

[ "2205.11465", "2112.07637", "2104.05938" ]

[ "en" ]

[ "# Dataset Card for GEM/squality", "## Dataset Description\n\n- Homepage: URL\n- Repository: URL\n- Paper: URL\n- Leaderboard: N/A\n- Point of Contact: Alex Wang", "### Link to Main Data Card\n\nYou can find the main data card on the GEM Website.", "### Dataset Summary \n\nSQuALITY (Summarization-format QUestion Answering with Long Input Texts, Yes!) is a summarization dataset that is:\n* Abstractive\n* Long-input: The input document are short stories between 3000--6000 words.\n* Question-focused: Each story is associated with multiple question-summary pairs.\n* Multi-reference: Each question is paired with 4 summaries.\n* High-quality: The summaries are crowdsourced from skilled and trained writers.\n\nYou can load the dataset via:\n\nThe data loader can be found here.", "#### website\nGithub", "#### paper\nArXiv", "#### authors\nAlex Wang (NYU); Angelica Chen (NYU); Richard Yuanzhe Pang (NYU); Nitish Joshi (NYU); Samuel R. Bowman (NYU)", "## Dataset Overview", "### Where to find the Data and its Documentation", "#### Webpage\n\n\n\nGithub", "#### Download\n\n\n\nGithub", "#### Paper\n\n\n\nArXiv", "#### BibTex", "#### Contact Name\n\n\n\n\nAlex Wang", "#### Contact Email\n\n\n\nwangalexc@URL", "#### Has a Leaderboard?\n\n\n\nno", "### Languages and Intended Use", "#### Multilingual?\n\n\n\n\nno", "#### Covered Dialects\n\n\n\nstories: 1930--1970 American English\nsummaries: modern American English", "#### Covered Languages\n\n\n\n\n'English'", "#### Whose Language?\n\n\n\nstories: 1930--1970 American science fiction writers (predominantly American men)\nsummaries: Upwork writers (college-educated, native-English) and NYU undergraduates (English-fluent college students)", "#### License\n\n\n\n\ncc-by-4.0: Creative Commons Attribution 4.0 International", "#### Intended Use\n\n\n\nsummarization research", "#### Primary Task\n\n\n\nSummarization", "#### Communicative Goal\n\n\n\n\nGiven a question about a particular high-level aspect of a short story, provide a summary about that aspect in the story (e.g., plot, character relationships, setting, theme, etc.).", "### Credit", "#### Curation Organization Type(s)\n\n\n\n'academic'", "#### Curation Organization(s)\n\n\n\nNew York University", "#### Dataset Creators\n\n\n\nAlex Wang (NYU); Angelica Chen (NYU); Richard Yuanzhe Pang (NYU); Nitish Joshi (NYU); Samuel R. Bowman (NYU)", "#### Funding\n\n\n\nEric and Wendy Schmidt; Apple; NSF", "#### Who added the Dataset to GEM?\n\n\n\nAlex Wang (NYU)", "### Dataset Structure", "#### Data Fields\n\n\n\n* metadata: Project Gutenberg ID, internal UID, Project Gutenberg license\n* document: the story\n* questions: a list where each element contains\n * question text: the question\n * question number: the order in which workers answered the question\n * responses: a list where each element contains\n * worker ID: anonymous\n * internal UID\n * response text: the response", "#### Reason for Structure\n\n\n\nThe dataset is arranged with responses grouped by question (for ease of multi-reference training and evaluation) and questions grouped by story (to avoid duplicating the story in the dataset)", "#### Example Instance", "#### Data Splits\n\n\n\ntrain, dev, test", "#### Splitting Criteria\n\n\n\nStories that appear in both SQuALITY and QuALITY are assigned to the same split in both datasets.", "## Dataset in GEM", "### Rationale for Inclusion in GEM", "#### Why is the Dataset in GEM?\n\n\n\nThe summaries in the dataset were crowdsourced, allowing us to use input documents that are easily understood by crowdworkers (as opposed to technical domains, such as scientific papers). Additionally, there is no lede bias in stories, as is typically in news articles used in benchmark summarization datasets like CNN/DM and XSum.\n\nAdditionally, the dataset is multi-reference and the references for each task are highly diverse. Having a diverse set of references better represents the set of acceptable summaries for an input, and opens the door for creative evaluation methodologies using these multiple references.", "#### Similar Datasets\n\n\n\nyes", "#### Unique Language Coverage\n\n\n\nno", "#### Difference from other GEM datasets\n\n\n\nThe inputs (story-question pairs) are multi-reference. The questions are high-level and are written to draw from multiple parts of the story, instead of a single section of the story.", "### GEM-Specific Curation", "#### Modificatied for GEM?\n\n\n\nno", "#### Additional Splits?\n\n\n\nno", "### Getting Started with the Task", "#### Pointers to Resources\n\n\n\n* original paper\n* modeling question-focused summarization\n* similar task format but different domain", "## Previous Results", "### Previous Results", "#### Metrics\n\n\n\n'ROUGE', 'BERT-Score'", "#### Proposed Evaluation\n\n\n\nFollowing norms in summarization, we have evaluated with automatic evaluation metrics like ROUGE and BERTScore, but these metrics do not correlate with human judgments of summary quality when comparing model summaries (see paper for details). \n\nWe highly recommend users of the benchmark use human evaluation as the primary method for evaluating systems. We present one example of such in the paper in which we ask Upwork workers to read the short story and then rate sets of three responses to each question. While this is close to the gold standard in how we would want to evaluate systems on this task, we recognize that finding workers who will read the whole story (~30m) is difficult and expensive, and doing efficient human evaluation for long document tasks is an open problem.", "#### Previous results available?\n\n\n\nyes", "#### Other Evaluation Approaches\n\n\n\nHuman evaluation", "#### Relevant Previous Results\n\n\n\nSee paper (URL", "## Dataset Curation", "### Original Curation", "#### Sourced from Different Sources\n\n\n\nno", "### Language Data", "#### How was Language Data Obtained?\n\n\n\n'Crowdsourced'", "#### Where was it crowdsourced?\n\n\n\n'Other crowdworker platform'", "#### Language Producers\n\n\n\nUpwork: US-born, native English speakers with backgrounds in the humanities and copywriting\n\nNYU undergraduates: English-fluent undergraduates from a diverse set of nationalities and majors", "#### Topics Covered\n\n\n\nThe short stories are primarily science fiction and from the 1930s -- 1970s.", "#### Data Validation\n\n\n\nvalidated by crowdworker", "#### Was Data Filtered?\n\n\n\nnot filtered", "### Structured Annotations", "#### Additional Annotations?\n\n\n\n\ncrowd-sourced", "#### Number of Raters\n\n\n\n11<n<50", "#### Rater Qualifications\n\n\n\nEnglish-fluent, with experience reading and writing about literature", "#### Raters per Training Example\n\n\n\n4", "#### Raters per Test Example\n\n\n\n4", "#### Annotation Service?\n\n\n\nno", "#### Any Quality Control?\n\n\n\nvalidated by another rater", "#### Quality Control Details\n\n\n\nEach response was reviewed by three reviewers, who ranked the response (against two other responses), highlighted errors in the response, and provided feedback to the original response writer.", "### Consent", "#### Any Consent Policy?\n\n\n\nyes", "#### Consent Policy Details\n\n\n\nWriters were informed that their writing and reviewing would be used in the development of AI.", "### Private Identifying Information (PII)", "#### Contains PII?\n\n\n\n\nunlikely", "#### Any PII Identification?\n\n\n\nno identification", "### Maintenance", "#### Any Maintenance Plan?\n\n\n\nno", "## Broader Social Context", "### Previous Work on the Social Impact of the Dataset", "#### Usage of Models based on the Data\n\n\n\nno", "### Impact on Under-Served Communities", "#### Addresses needs of underserved Communities?\n\n\n\nno", "### Discussion of Biases", "#### Any Documented Social Biases?\n\n\n\nyes", "## Considerations for Using the Data", "### PII Risks and Liability", "### Licenses", "#### Copyright Restrictions on the Dataset\n\n\n\n'open license - commercial use allowed'", "#### Copyright Restrictions on the Language Data\n\n\n\n'public domain'", "### Known Technical Limitations", "#### Unsuited Applications\n\n\n\nThe stories in the dataset are from the 1930--1970s and may contain harmful stances on topics like race and gender. Models trained on the stories may reproduce these stances in their outputs.", "#### Discouraged Use Cases\n\n\n\nThe proposed automatic metrics for this dataset (ROUGE, BERTScore) are not sensitive to factual errors in summaries, and have been shown to not correlate well with human judgments of summary quality along a number of axes." ]

[ "TAGS\n#task_categories-summarization #annotations_creators-crowd-sourced #language_creators-unknown #multilinguality-unknown #size_categories-unknown #source_datasets-original #language-English #license-cc-by-4.0 #arxiv-2205.11465 #arxiv-2112.07637 #arxiv-2104.05938 #region-us \n", "# Dataset Card for GEM/squality", "## Dataset Description\n\n- Homepage: URL\n- Repository: URL\n- Paper: URL\n- Leaderboard: N/A\n- Point of Contact: Alex Wang", "### Link to Main Data Card\n\nYou can find the main data card on the GEM Website.", "### Dataset Summary \n\nSQuALITY (Summarization-format QUestion Answering with Long Input Texts, Yes!) is a summarization dataset that is:\n* Abstractive\n* Long-input: The input document are short stories between 3000--6000 words.\n* Question-focused: Each story is associated with multiple question-summary pairs.\n* Multi-reference: Each question is paired with 4 summaries.\n* High-quality: The summaries are crowdsourced from skilled and trained writers.\n\nYou can load the dataset via:\n\nThe data loader can be found here.", "#### website\nGithub", "#### paper\nArXiv", "#### authors\nAlex Wang (NYU); Angelica Chen (NYU); Richard Yuanzhe Pang (NYU); Nitish Joshi (NYU); Samuel R. Bowman (NYU)", "## Dataset Overview", "### Where to find the Data and its Documentation", "#### Webpage\n\n\n\nGithub", "#### Download\n\n\n\nGithub", "#### Paper\n\n\n\nArXiv", "#### BibTex", "#### Contact Name\n\n\n\n\nAlex Wang", "#### Contact Email\n\n\n\nwangalexc@URL", "#### Has a Leaderboard?\n\n\n\nno", "### Languages and Intended Use", "#### Multilingual?\n\n\n\n\nno", "#### Covered Dialects\n\n\n\nstories: 1930--1970 American English\nsummaries: modern American English", "#### Covered Languages\n\n\n\n\n'English'", "#### Whose Language?\n\n\n\nstories: 1930--1970 American science fiction writers (predominantly American men)\nsummaries: Upwork writers (college-educated, native-English) and NYU undergraduates (English-fluent college students)", "#### License\n\n\n\n\ncc-by-4.0: Creative Commons Attribution 4.0 International", "#### Intended Use\n\n\n\nsummarization research", "#### Primary Task\n\n\n\nSummarization", "#### Communicative Goal\n\n\n\n\nGiven a question about a particular high-level aspect of a short story, provide a summary about that aspect in the story (e.g., plot, character relationships, setting, theme, etc.).", "### Credit", "#### Curation Organization Type(s)\n\n\n\n'academic'", "#### Curation Organization(s)\n\n\n\nNew York University", "#### Dataset Creators\n\n\n\nAlex Wang (NYU); Angelica Chen (NYU); Richard Yuanzhe Pang (NYU); Nitish Joshi (NYU); Samuel R. Bowman (NYU)", "#### Funding\n\n\n\nEric and Wendy Schmidt; Apple; NSF", "#### Who added the Dataset to GEM?\n\n\n\nAlex Wang (NYU)", "### Dataset Structure", "#### Data Fields\n\n\n\n* metadata: Project Gutenberg ID, internal UID, Project Gutenberg license\n* document: the story\n* questions: a list where each element contains\n * question text: the question\n * question number: the order in which workers answered the question\n * responses: a list where each element contains\n * worker ID: anonymous\n * internal UID\n * response text: the response", "#### Reason for Structure\n\n\n\nThe dataset is arranged with responses grouped by question (for ease of multi-reference training and evaluation) and questions grouped by story (to avoid duplicating the story in the dataset)", "#### Example Instance", "#### Data Splits\n\n\n\ntrain, dev, test", "#### Splitting Criteria\n\n\n\nStories that appear in both SQuALITY and QuALITY are assigned to the same split in both datasets.", "## Dataset in GEM", "### Rationale for Inclusion in GEM", "#### Why is the Dataset in GEM?\n\n\n\nThe summaries in the dataset were crowdsourced, allowing us to use input documents that are easily understood by crowdworkers (as opposed to technical domains, such as scientific papers). Additionally, there is no lede bias in stories, as is typically in news articles used in benchmark summarization datasets like CNN/DM and XSum.\n\nAdditionally, the dataset is multi-reference and the references for each task are highly diverse. Having a diverse set of references better represents the set of acceptable summaries for an input, and opens the door for creative evaluation methodologies using these multiple references.", "#### Similar Datasets\n\n\n\nyes", "#### Unique Language Coverage\n\n\n\nno", "#### Difference from other GEM datasets\n\n\n\nThe inputs (story-question pairs) are multi-reference. The questions are high-level and are written to draw from multiple parts of the story, instead of a single section of the story.", "### GEM-Specific Curation", "#### Modificatied for GEM?\n\n\n\nno", "#### Additional Splits?\n\n\n\nno", "### Getting Started with the Task", "#### Pointers to Resources\n\n\n\n* original paper\n* modeling question-focused summarization\n* similar task format but different domain", "## Previous Results", "### Previous Results", "#### Metrics\n\n\n\n'ROUGE', 'BERT-Score'", "#### Proposed Evaluation\n\n\n\nFollowing norms in summarization, we have evaluated with automatic evaluation metrics like ROUGE and BERTScore, but these metrics do not correlate with human judgments of summary quality when comparing model summaries (see paper for details). \n\nWe highly recommend users of the benchmark use human evaluation as the primary method for evaluating systems. We present one example of such in the paper in which we ask Upwork workers to read the short story and then rate sets of three responses to each question. While this is close to the gold standard in how we would want to evaluate systems on this task, we recognize that finding workers who will read the whole story (~30m) is difficult and expensive, and doing efficient human evaluation for long document tasks is an open problem.", "#### Previous results available?\n\n\n\nyes", "#### Other Evaluation Approaches\n\n\n\nHuman evaluation", "#### Relevant Previous Results\n\n\n\nSee paper (URL", "## Dataset Curation", "### Original Curation", "#### Sourced from Different Sources\n\n\n\nno", "### Language Data", "#### How was Language Data Obtained?\n\n\n\n'Crowdsourced'", "#### Where was it crowdsourced?\n\n\n\n'Other crowdworker platform'", "#### Language Producers\n\n\n\nUpwork: US-born, native English speakers with backgrounds in the humanities and copywriting\n\nNYU undergraduates: English-fluent undergraduates from a diverse set of nationalities and majors", "#### Topics Covered\n\n\n\nThe short stories are primarily science fiction and from the 1930s -- 1970s.", "#### Data Validation\n\n\n\nvalidated by crowdworker", "#### Was Data Filtered?\n\n\n\nnot filtered", "### Structured Annotations", "#### Additional Annotations?\n\n\n\n\ncrowd-sourced", "#### Number of Raters\n\n\n\n11<n<50", "#### Rater Qualifications\n\n\n\nEnglish-fluent, with experience reading and writing about literature", "#### Raters per Training Example\n\n\n\n4", "#### Raters per Test Example\n\n\n\n4", "#### Annotation Service?\n\n\n\nno", "#### Any Quality Control?\n\n\n\nvalidated by another rater", "#### Quality Control Details\n\n\n\nEach response was reviewed by three reviewers, who ranked the response (against two other responses), highlighted errors in the response, and provided feedback to the original response writer.", "### Consent", "#### Any Consent Policy?\n\n\n\nyes", "#### Consent Policy Details\n\n\n\nWriters were informed that their writing and reviewing would be used in the development of AI.", "### Private Identifying Information (PII)", "#### Contains PII?\n\n\n\n\nunlikely", "#### Any PII Identification?\n\n\n\nno identification", "### Maintenance", "#### Any Maintenance Plan?\n\n\n\nno", "## Broader Social Context", "### Previous Work on the Social Impact of the Dataset", "#### Usage of Models based on the Data\n\n\n\nno", "### Impact on Under-Served Communities", "#### Addresses needs of underserved Communities?\n\n\n\nno", "### Discussion of Biases", "#### Any Documented Social Biases?\n\n\n\nyes", "## Considerations for Using the Data", "### PII Risks and Liability", "### Licenses", "#### Copyright Restrictions on the Dataset\n\n\n\n'open license - commercial use allowed'", "#### Copyright Restrictions on the Language Data\n\n\n\n'public domain'", "### Known Technical Limitations", "#### Unsuited Applications\n\n\n\nThe stories in the dataset are from the 1930--1970s and may contain harmful stances on topics like race and gender. Models trained on the stories may reproduce these stances in their outputs.", "#### Discouraged Use Cases\n\n\n\nThe proposed automatic metrics for this dataset (ROUGE, BERTScore) are not sensitive to factual errors in summaries, and have been shown to not correlate well with human judgments of summary quality along a number of axes." ]

[ "license:cc-by-4.0", "region:us" ]

[]

[]

[]

[ "TAGS\n#license-cc-by-4.0 #region-us \n" ]

[ "license:afl-3.0", "region:us" ]

[]

[]

[]

[ "TAGS\n#license-afl-3.0 #region-us \n" ]

[ "task_categories:question-answering", "task_ids:extractive-qa", "annotations_creators:expert-generated", "language_creators:expert-generated", "multilinguality:multilingual", "size_categories:unknown", "source_datasets:extended|squad", "language:en", "language:es", "language:de", "language:el", "language:hi", "language:th", "language:ru", "language:tr", "language:ar", "language:vi", "language:zh", "language:ro", "license:cc-by-sa-4.0", "arxiv:1910.11856", "region:us" ]

[ "1910.11856" ]

[ "en", "es", "de", "el", "hi", "th", "ru", "tr", "ar", "vi", "zh", "ro" ]

[ "### Dataset Summary\n\n\nXQuAD (Cross-lingual Question Answering Dataset) is a benchmark dataset for evaluating cross-lingual question answering\nperformance. The dataset consists of a subset of 240 paragraphs and 1190 question-answer pairs from the development set\nof SQuAD v1.1 (Rajpurkar et al., 2016) together with their professional translations into ten language: Spanish, German,\nGreek, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, Hindi and Romanian. Consequently, the dataset is entirely parallel across 12 languages.\n\n\nWe also include \"translate-train\", \"translate-dev\", and \"translate-test\"\nsplits for each non-English language from XTREME (Hu et al., 2020). These can be used to run XQuAD in the \"translate-train\" or \"translate-test\" settings. URL", "### Supported Tasks and Leaderboards", "### Languages\n\n\nDataset Structure\n-----------------", "### Data Instances", "#### ar\n\n\n* Size of downloaded dataset files: 12.68 MB\n* Size of the generated dataset: 1.64 MB\n* Total amount of disk used: 14.33 MB\n\n\nAn example of 'test' looks as follows.", "#### de\n\n\n* Size of downloaded dataset files: 12.68 MB\n* Size of the generated dataset: 1.23 MB\n* Total amount of disk used: 13.91 MB\n\n\nAn example of 'test' looks as follows.", "#### el\n\n\n* Size of downloaded dataset files: 12.68 MB\n* Size of the generated dataset: 2.11 MB\n* Total amount of disk used: 14.79 MB\n\n\nAn example of 'test' looks as follows.", "#### en\n\n\n* Size of downloaded dataset files: 12.68 MB\n* Size of the generated dataset: 1.07 MB\n* Total amount of disk used: 13.75 MB\n\n\nAn example of 'test' looks as follows.", "#### es\n\n\n* Size of downloaded dataset files: 12.68 MB\n* Size of the generated dataset: 1.22 MB\n* Total amount of disk used: 13.90 MB\n\n\nAn example of 'test' looks as follows.", "### Data Fields\n\n\nThe data fields are the same among all splits.", "#### ar\n\n\n* 'id': a 'string' feature.\n* 'context': a 'string' feature.\n* 'question': a 'string' feature.\n* 'answers': a dictionary feature containing:\n\t+ 'text': a 'string' feature.\n\t+ 'answer\\_start': a 'int32' feature.", "#### de\n\n\n* 'id': a 'string' feature.\n* 'context': a 'string' feature.\n* 'question': a 'string' feature.\n* 'answers': a dictionary feature containing:\n\t+ 'text': a 'string' feature.\n\t+ 'answer\\_start': a 'int32' feature.", "#### el\n\n\n* 'id': a 'string' feature.\n* 'context': a 'string' feature.\n* 'question': a 'string' feature.\n* 'answers': a dictionary feature containing:\n\t+ 'text': a 'string' feature.\n\t+ 'answer\\_start': a 'int32' feature.", "#### en\n\n\n* 'id': a 'string' feature.\n* 'context': a 'string' feature.\n* 'question': a 'string' feature.\n* 'answers': a dictionary feature containing:\n\t+ 'text': a 'string' feature.\n\t+ 'answer\\_start': a 'int32' feature.", "#### es\n\n\n* 'id': a 'string' feature.\n* 'context': a 'string' feature.\n* 'question': a 'string' feature.\n* 'answers': a dictionary feature containing:\n\t+ 'text': a 'string' feature.\n\t+ 'answer\\_start': a 'int32' feature.", "### Data Splits\n\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations\n\n\nAdditional Information\n----------------------", "### Dataset Curators", "### Licensing Information", "### Contributions\n\n\nThanks to @lewtun, @patrickvonplaten, @thomwolf for adding this dataset." ]

[ "TAGS\n#task_categories-question-answering #task_ids-extractive-qa #annotations_creators-expert-generated #language_creators-expert-generated #multilinguality-multilingual #size_categories-unknown #source_datasets-extended|squad #language-English #language-Spanish #language-German #language-Modern Greek (1453-) #language-Hindi #language-Thai #language-Russian #language-Turkish #language-Arabic #language-Vietnamese #language-Chinese #language-Romanian #license-cc-by-sa-4.0 #arxiv-1910.11856 #region-us \n", "### Dataset Summary\n\n\nXQuAD (Cross-lingual Question Answering Dataset) is a benchmark dataset for evaluating cross-lingual question answering\nperformance. The dataset consists of a subset of 240 paragraphs and 1190 question-answer pairs from the development set\nof SQuAD v1.1 (Rajpurkar et al., 2016) together with their professional translations into ten language: Spanish, German,\nGreek, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, Hindi and Romanian. Consequently, the dataset is entirely parallel across 12 languages.\n\n\nWe also include \"translate-train\", \"translate-dev\", and \"translate-test\"\nsplits for each non-English language from XTREME (Hu et al., 2020). These can be used to run XQuAD in the \"translate-train\" or \"translate-test\" settings. URL", "### Supported Tasks and Leaderboards", "### Languages\n\n\nDataset Structure\n-----------------", "### Data Instances", "#### ar\n\n\n* Size of downloaded dataset files: 12.68 MB\n* Size of the generated dataset: 1.64 MB\n* Total amount of disk used: 14.33 MB\n\n\nAn example of 'test' looks as follows.", "#### de\n\n\n* Size of downloaded dataset files: 12.68 MB\n* Size of the generated dataset: 1.23 MB\n* Total amount of disk used: 13.91 MB\n\n\nAn example of 'test' looks as follows.", "#### el\n\n\n* Size of downloaded dataset files: 12.68 MB\n* Size of the generated dataset: 2.11 MB\n* Total amount of disk used: 14.79 MB\n\n\nAn example of 'test' looks as follows.", "#### en\n\n\n* Size of downloaded dataset files: 12.68 MB\n* Size of the generated dataset: 1.07 MB\n* Total amount of disk used: 13.75 MB\n\n\nAn example of 'test' looks as follows.", "#### es\n\n\n* Size of downloaded dataset files: 12.68 MB\n* Size of the generated dataset: 1.22 MB\n* Total amount of disk used: 13.90 MB\n\n\nAn example of 'test' looks as follows.", "### Data Fields\n\n\nThe data fields are the same among all splits.", "#### ar\n\n\n* 'id': a 'string' feature.\n* 'context': a 'string' feature.\n* 'question': a 'string' feature.\n* 'answers': a dictionary feature containing:\n\t+ 'text': a 'string' feature.\n\t+ 'answer\\_start': a 'int32' feature.", "#### de\n\n\n* 'id': a 'string' feature.\n* 'context': a 'string' feature.\n* 'question': a 'string' feature.\n* 'answers': a dictionary feature containing:\n\t+ 'text': a 'string' feature.\n\t+ 'answer\\_start': a 'int32' feature.", "#### el\n\n\n* 'id': a 'string' feature.\n* 'context': a 'string' feature.\n* 'question': a 'string' feature.\n* 'answers': a dictionary feature containing:\n\t+ 'text': a 'string' feature.\n\t+ 'answer\\_start': a 'int32' feature.", "#### en\n\n\n* 'id': a 'string' feature.\n* 'context': a 'string' feature.\n* 'question': a 'string' feature.\n* 'answers': a dictionary feature containing:\n\t+ 'text': a 'string' feature.\n\t+ 'answer\\_start': a 'int32' feature.", "#### es\n\n\n* 'id': a 'string' feature.\n* 'context': a 'string' feature.\n* 'question': a 'string' feature.\n* 'answers': a dictionary feature containing:\n\t+ 'text': a 'string' feature.\n\t+ 'answer\\_start': a 'int32' feature.", "### Data Splits\n\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations\n\n\nAdditional Information\n----------------------", "### Dataset Curators", "### Licensing Information", "### Contributions\n\n\nThanks to @lewtun, @patrickvonplaten, @thomwolf for adding this dataset." ]

[ "region:us" ]

[]

[]

[ "### Dataset Summary\n\n\nThis dataset is intended as a test case for fine-grain classification tasks (10 different kinds of monkey species). The dataset consists of almost 1400 JPEG images grouped into two splits - training and validation. Each split contains 10 categories labeled as n0~n9, each corresponding a species from Wikipedia's monkey cladogram. Images were downloaded with help of the googliser open source code.\n\n\n\nThis collection includes the following GTZAN variants:\n\n\n* original (images are 400x300 px or larger; ~550 MB)\n* downsized (images are downsized to 224x224 px; ~40 MB)", "### Supported Tasks and Leaderboards", "### Languages\n\n\nDataset Structure\n-----------------", "### Data Instances", "### Data Fields", "### Data Splits\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations\n\n\nAdditional Information\n----------------------", "### Dataset Curators", "### Licensing Information" ]

[ "TAGS\n#region-us \n", "### Dataset Summary\n\n\nThis dataset is intended as a test case for fine-grain classification tasks (10 different kinds of monkey species). The dataset consists of almost 1400 JPEG images grouped into two splits - training and validation. Each split contains 10 categories labeled as n0~n9, each corresponding a species from Wikipedia's monkey cladogram. Images were downloaded with help of the googliser open source code.\n\n\n\nThis collection includes the following GTZAN variants:\n\n\n* original (images are 400x300 px or larger; ~550 MB)\n* downsized (images are downsized to 224x224 px; ~40 MB)", "### Supported Tasks and Leaderboards", "### Languages\n\n\nDataset Structure\n-----------------", "### Data Instances", "### Data Fields", "### Data Splits\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations\n\n\nAdditional Information\n----------------------", "### Dataset Curators", "### Licensing Information" ]

[ "task_categories:fill-mask", "task_categories:token-classification", "task_categories:text-classification", "task_ids:masked-language-modeling", "task_ids:named-entity-recognition", "task_ids:part-of-speech", "task_ids:lemmatization", "task_ids:parsing", "task_ids:entity-linking-classification", "annotations_creators:machine-generated", "language_creators:machine-generated", "multilinguality:monolingual", "size_categories:1M<n<10M", "source_datasets:extended|wikitext", "language:en", "license:cc-by-sa-4.0", "arxiv:1609.07843", "region:us" ]

[ "1609.07843" ]

[ "en" ]

[ "### Dataset Summary\n\n\nThe WikiText language modeling dataset is a collection of over 100 million tokens extracted from\nthe set of verified Good and Featured articles on Wikipedia. Dependency Relations, POS, NER tags\nare marked with trankit, entities are linked with\nentity-fishing, which also tags another field\nof NER tags. The dataset is available under the Creative Commons Attribution-ShareAlike License.\n\n\nCompared to the preprocessed version of Penn Treebank (PTB), WikiText-2 is over 2 times larger and\nWikiText-103 is over 110 times larger. The WikiText dataset also features a far larger vocabulary\nand retains the original case, punctuation and numbers - all of which are removed in PTB. As it is\ncomposed of full articles, the dataset is well suited for models that can take advantage of long\nterm dependencies.", "### Supported Tasks and Leaderboards\n\n\n* masked-language-modeling\n* named-entity-recognition\n* part-of-speech\n* lemmatization\n* parsing\n* entity-linking-classification", "### Languages\n\n\nEnglish.\n\n\nDataset Structure\n-----------------", "### Data Instances", "#### wikitext2\n\n\n* Size of downloaded dataset files: 27.3 MB\n* Size of the generated dataset: 197.2 MB\n* Total amount of disk used: 197.2 MB\n\n\nAn example of 'validation' looks as follows.", "#### wikitext103\n\n\n* Size of downloaded dataset files: 1.11 GB\n* Size of the generated dataset: 7.82 GB\n* Total amount of disk used: 7.82 GB\n\n\nAn example of 'train' looks as follows.\n\n\nUse following code to print the examples nicely:", "### Data Fields\n\n\nThe data fields are the same among all splits.\n\n\n* text: string feature.\n* original\\_id: int feature. Mapping to index within original wikitext dataset.\n* tok\\_span: sequence of (int, int) tuples. Denotes token spans (start inclusive, end exclusive)\nwithin each sentence.\nNote that each sentence includes an artificial root node to align dependency relations.\n* tok\\_upos: string feature. Universal Dependency POS tag\ntags. Aligned with tok\\_span. Root node has tag \"root\".\n* tok\\_xpos: string geature. XPOS POS tag.\nAligned with tok\\_span. Root node has tag \"root\".\n* tok\\_dephead: int feature.\nUniversal Dependency Head Node. Int refers\nto tokens in tok\\_span. Root node has head '0' (itself).\n* tok\\_deprel: Universal Dependency Relation Description.\nRefers to the relation between this token and head token. Aligned with tok\\_span. Root node has\ndependency relation \"root\" to itself.\n* tok\\_lemma: string feature. Lemma of token. Aligend with tok\\_span.\n* tok\\_ner: string feature. NER tag of token. Marked in BIOS schema (e.g. S-MISC, B-LOC, ...)\nAligned with tok\\_span. Root node has NER tag 'None'.\n* ent\\_span: sequence of (int, int) tuples. Denotes entities found by entity-fishing\n(start inclusive, end exclusive).\n* ent\\_wikipedia\\_external\\_ref: string feature. External Reference to wikipedia page. You can\naccess the wikipedia page via the url 'URL\nAligend with ent\\_span. All entities either have this field, or the 'ent\\_ner' field, but not both.\nAn empty field is denoted by the string 'None'. Aligned with ent\\_span.\n* ent\\_ner: string feature. Denotes NER tags. An empty field is denoted by the string 'None'.\nAligned with ent\\_span.\n\"ent\\_domains\": sequence of string. Denotes domains of entity. Can be empty sequence. Aligned with\nent\\_span.", "### Data Splits\n\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization\n\n\nURL", "#### Who are the source language producers?", "### Annotations", "#### Annotation process\n\n\n1. Started with 'wikitext2-raw-v1' and 'wikitext103-raw-v1' from wikitext\n2. Ran datasets through Trankit. Marked all fields starting with 'tok'.\n\n\nIn this step, the texts have been split into sentences. To retain the original text sections\nyou can accumulate over 'original\\_id' (examples are in order).\n\n\n3. Ran datasets through entity-fishing. Marked all fields starting with 'ent'.", "#### Who are the annotators?\n\n\nMachines powered by DFKI.", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations\n\n\nAdditional Information\n----------------------", "### Dataset Curators", "### Licensing Information\n\n\nCreative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0)\n\n\nPlease cite the original creators of wikitext, and the great people\ndeveloping trankit and entity-fishing.", "### Contributions\n\n\nThanks to @GabrielKP for adding this dataset." ]

[ "TAGS\n#task_categories-fill-mask #task_categories-token-classification #task_categories-text-classification #task_ids-masked-language-modeling #task_ids-named-entity-recognition #task_ids-part-of-speech #task_ids-lemmatization #task_ids-parsing #task_ids-entity-linking-classification #annotations_creators-machine-generated #language_creators-machine-generated #multilinguality-monolingual #size_categories-1M<n<10M #source_datasets-extended|wikitext #language-English #license-cc-by-sa-4.0 #arxiv-1609.07843 #region-us \n", "### Dataset Summary\n\n\nThe WikiText language modeling dataset is a collection of over 100 million tokens extracted from\nthe set of verified Good and Featured articles on Wikipedia. Dependency Relations, POS, NER tags\nare marked with trankit, entities are linked with\nentity-fishing, which also tags another field\nof NER tags. The dataset is available under the Creative Commons Attribution-ShareAlike License.\n\n\nCompared to the preprocessed version of Penn Treebank (PTB), WikiText-2 is over 2 times larger and\nWikiText-103 is over 110 times larger. The WikiText dataset also features a far larger vocabulary\nand retains the original case, punctuation and numbers - all of which are removed in PTB. As it is\ncomposed of full articles, the dataset is well suited for models that can take advantage of long\nterm dependencies.", "### Supported Tasks and Leaderboards\n\n\n* masked-language-modeling\n* named-entity-recognition\n* part-of-speech\n* lemmatization\n* parsing\n* entity-linking-classification", "### Languages\n\n\nEnglish.\n\n\nDataset Structure\n-----------------", "### Data Instances", "#### wikitext2\n\n\n* Size of downloaded dataset files: 27.3 MB\n* Size of the generated dataset: 197.2 MB\n* Total amount of disk used: 197.2 MB\n\n\nAn example of 'validation' looks as follows.", "#### wikitext103\n\n\n* Size of downloaded dataset files: 1.11 GB\n* Size of the generated dataset: 7.82 GB\n* Total amount of disk used: 7.82 GB\n\n\nAn example of 'train' looks as follows.\n\n\nUse following code to print the examples nicely:", "### Data Fields\n\n\nThe data fields are the same among all splits.\n\n\n* text: string feature.\n* original\\_id: int feature. Mapping to index within original wikitext dataset.\n* tok\\_span: sequence of (int, int) tuples. Denotes token spans (start inclusive, end exclusive)\nwithin each sentence.\nNote that each sentence includes an artificial root node to align dependency relations.\n* tok\\_upos: string feature. Universal Dependency POS tag\ntags. Aligned with tok\\_span. Root node has tag \"root\".\n* tok\\_xpos: string geature. XPOS POS tag.\nAligned with tok\\_span. Root node has tag \"root\".\n* tok\\_dephead: int feature.\nUniversal Dependency Head Node. Int refers\nto tokens in tok\\_span. Root node has head '0' (itself).\n* tok\\_deprel: Universal Dependency Relation Description.\nRefers to the relation between this token and head token. Aligned with tok\\_span. Root node has\ndependency relation \"root\" to itself.\n* tok\\_lemma: string feature. Lemma of token. Aligend with tok\\_span.\n* tok\\_ner: string feature. NER tag of token. Marked in BIOS schema (e.g. S-MISC, B-LOC, ...)\nAligned with tok\\_span. Root node has NER tag 'None'.\n* ent\\_span: sequence of (int, int) tuples. Denotes entities found by entity-fishing\n(start inclusive, end exclusive).\n* ent\\_wikipedia\\_external\\_ref: string feature. External Reference to wikipedia page. You can\naccess the wikipedia page via the url 'URL\nAligend with ent\\_span. All entities either have this field, or the 'ent\\_ner' field, but not both.\nAn empty field is denoted by the string 'None'. Aligned with ent\\_span.\n* ent\\_ner: string feature. Denotes NER tags. An empty field is denoted by the string 'None'.\nAligned with ent\\_span.\n\"ent\\_domains\": sequence of string. Denotes domains of entity. Can be empty sequence. Aligned with\nent\\_span.", "### Data Splits\n\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization\n\n\nURL", "#### Who are the source language producers?", "### Annotations", "#### Annotation process\n\n\n1. Started with 'wikitext2-raw-v1' and 'wikitext103-raw-v1' from wikitext\n2. Ran datasets through Trankit. Marked all fields starting with 'tok'.\n\n\nIn this step, the texts have been split into sentences. To retain the original text sections\nyou can accumulate over 'original\\_id' (examples are in order).\n\n\n3. Ran datasets through entity-fishing. Marked all fields starting with 'ent'.", "#### Who are the annotators?\n\n\nMachines powered by DFKI.", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations\n\n\nAdditional Information\n----------------------", "### Dataset Curators", "### Licensing Information\n\n\nCreative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0)\n\n\nPlease cite the original creators of wikitext, and the great people\ndeveloping trankit and entity-fishing.", "### Contributions\n\n\nThanks to @GabrielKP for adding this dataset." ]

[ "region:us" ]

[]

[]

[]

[ "TAGS\n#region-us \n" ]

[ "region:us" ]

[]

[]

[]

[ "TAGS\n#region-us \n" ]

[ "language:ar", "language:de", "language:en", "language:es", "language:fr", "language:it", "language:pl", "language:ru", "language:tr", "language:zh", "region:us" ]

[]

[ "ar", "de", "en", "es", "fr", "it", "pl", "ru", "tr", "zh" ]

[]

[ "TAGS\n#language-Arabic #language-German #language-English #language-Spanish #language-French #language-Italian #language-Polish #language-Russian #language-Turkish #language-Chinese #region-us \n" ]

[ "license:cc", "region:us" ]

[]

[]

[]

[ "TAGS\n#license-cc #region-us \n" ]

[ "license:mit", "arxiv:2203.13926", "arxiv:1710.03957", "arxiv:1902.00164", "arxiv:2004.04494", "region:us" ]

[ "2203.13926", "1710.03957", "1902.00164", "2004.04494" ]

[]

[ "# Dataset Card for CICERO", "## Description\n\n- Homepage: URL\n- Repository: URL\n- Paper: URL\n- arXiv: URL", "### Summary\n\nCICERO is a new dataset for dialogue reasoning with contextualized commonsense inference. It contains 53K inferences for five commonsense dimensions – cause, subsequent event, prerequisite, motivation, and emotional reaction collected from 5.6K dialogues. We design several generative and multi-choice answer selection tasks to show the usefulness of CICERO in dialogue reasoning.", "### Supported Tasks\n\nInference generation (NLG) and multi-choice answer selection (QA).", "### Languages\n\nThe text in the dataset is in English. The associated BCP-47 code is en.", "## Dataset Structure", "### Data Fields\n\n- ID: Dialogue ID with dataset indicator.\n- Dialogue: Utterances of the dialogue in a list.\n- Target: Target utterance.\n- Question: One of the five questions (inference types).\n- Choices: Five possible answer choices in a list. One of the answers is human written. The other four answers are machine-generated and selected through the Adversarial Filtering (AF) algorithm.\n- Human Written Answer: Index of the human written answer in a single element list. Index starts from 0.\n- Correct Answers: List of all correct answers indicated as plausible or speculatively correct by the human annotators. Includes the index of the human written answer.", "### Data Instances\n\nAn instance of the dataset is as the following:", "### Data Splits\n\nThe dataset contains 31,418 instances for training, 10,888 instances for validation and 10,898 instances for testing.", "## Dataset Creation", "### Curation Rationale\n\nThe annotation process of CICERO can be found in the paper.", "### Source Data\n\nThe dialogues in CICERO are collected from three datasets - DailyDialog, DREAM, and MuTual" ]

[ "TAGS\n#license-mit #arxiv-2203.13926 #arxiv-1710.03957 #arxiv-1902.00164 #arxiv-2004.04494 #region-us \n", "# Dataset Card for CICERO", "## Description\n\n- Homepage: URL\n- Repository: URL\n- Paper: URL\n- arXiv: URL", "### Summary\n\nCICERO is a new dataset for dialogue reasoning with contextualized commonsense inference. It contains 53K inferences for five commonsense dimensions – cause, subsequent event, prerequisite, motivation, and emotional reaction collected from 5.6K dialogues. We design several generative and multi-choice answer selection tasks to show the usefulness of CICERO in dialogue reasoning.", "### Supported Tasks\n\nInference generation (NLG) and multi-choice answer selection (QA).", "### Languages\n\nThe text in the dataset is in English. The associated BCP-47 code is en.", "## Dataset Structure", "### Data Fields\n\n- ID: Dialogue ID with dataset indicator.\n- Dialogue: Utterances of the dialogue in a list.\n- Target: Target utterance.\n- Question: One of the five questions (inference types).\n- Choices: Five possible answer choices in a list. One of the answers is human written. The other four answers are machine-generated and selected through the Adversarial Filtering (AF) algorithm.\n- Human Written Answer: Index of the human written answer in a single element list. Index starts from 0.\n- Correct Answers: List of all correct answers indicated as plausible or speculatively correct by the human annotators. Includes the index of the human written answer.", "### Data Instances\n\nAn instance of the dataset is as the following:", "### Data Splits\n\nThe dataset contains 31,418 instances for training, 10,888 instances for validation and 10,898 instances for testing.", "## Dataset Creation", "### Curation Rationale\n\nThe annotation process of CICERO can be found in the paper.", "### Source Data\n\nThe dialogues in CICERO are collected from three datasets - DailyDialog, DREAM, and MuTual" ]

[ "region:us" ]

[]

[]

[]

[ "TAGS\n#region-us \n" ]

[ "task_categories:other", "annotations_creators:expert-generated", "language_creators:expert-generated", "multilinguality:monolingual", "size_categories:10K<n<100K", "source_datasets:original", "language:en", "license:other", "head-pose-estimation", "region:us" ]

[]

[ "en" ]

[ "# Dataset Card for Biwi Kinect Head Pose Database", "## Table of Contents\n- Dataset Description\n - Dataset Summary\n - Supported Tasks\n - Languages\n- Dataset Structure\n - Data Instances\n - Data Fields\n - Data Splits\n- Dataset Creation\n - Curation Rationale\n - Source Data\n - Annotations\n - Personal and Sensitive Information\n- Considerations for Using the Data\n - Social Impact of Dataset\n - Discussion of Biases\n - Other Known Limitations\n- Additional Information\n - Dataset Curators\n - Licensing Information\n - Citation Information", "## Dataset Description\n\n- Homepage: Biwi Kinect Head Pose homepage\n- Repository: \n- Paper: Biwi Kinect Head Pose paper\n- Leaderboard: \n- Point of Contact: Gabriele Fanelli", "### Dataset Summary\n\nThe Biwi Kinect Head Pose Database is acquired with the Microsoft Kinect sensor, a structured IR light device.It contains 15K images of 20 people with 6 females and 14 males where 4 people were recorded twice.\n\nFor each frame, there is :\n- a depth image,\n- a corresponding rgb image (both 640x480 pixels),\n- annotation\n\nThe head pose range covers about +-75 degrees yaw and +-60 degrees pitch. The ground truth is the 3D location of the head and its rotation.", "### Data Processing\n\nExample code for reading a compressed binary depth image file provided by the authors.\n\n<details>\n <summary> View C++ Code </summary>\n\n\n\n</details>", "### Supported Tasks and Leaderboards\n\nBiwi Kinect Head Pose Database supports the following tasks :\n- Head pose estimation\n- Pose estimation\n- Face verification", "### Languages", "## Dataset Structure", "### Data Instances\n\nA sample from the Biwi Kinect Head Pose dataset is provided below:", "### Data Fields\n\n- 'sequence_number' : This refers to the sequence number in the dataset. There are a total of 24 sequences.\n- 'subject_id' : This refers to the subjects in the dataset. There are a total of 20 people with 6 females and 14 males where 4 people were recorded twice.\n- 'rgb' : List of png frames containing the poses.\n- 'rgb_cal': Contains calibration information for the color camera which includes intrinsic matrix, \nglobal rotation and translation.\n- 'depth' : List of depth frames for the poses.\n- 'depth_cal': Contains calibration information for the depth camera which includes intrinsic matrix, global rotation and translation.\n- 'head_pose_gt' : Contains ground truth information, i.e., the location of the center of the head in 3D and the head rotation, encoded as a 3x3 rotation matrix.", "### Data Splits\n\nAll the data is contained in the training set.", "## Dataset Creation", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization\n\nThe Biwi Kinect Head Pose Database is acquired with the Microsoft Kinect sensor, a structured IR light device.", "#### Who are the source language producers?", "### Annotations", "#### Annotation process\n\nFrom Dataset's README : \n> The database contains 24 sequences acquired with a Kinect sensor. 20 people (some were recorded twice - 6 women and 14 men) were recorded while turning their heads, sitting in front of the sensor, at roughly one meter of distance.", "#### Who are the annotators?", "### Personal and Sensitive Information", "## Considerations for Using the Data", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations", "## Additional Information", "### Dataset Curators", "### Licensing Information\n\nFrom Dataset's README : \n> This database is made available for non-commercial use such as university research and education.", "### Contributions\n\nThanks to @dnaveenr for adding this dataset." ]

[ "TAGS\n#task_categories-other #annotations_creators-expert-generated #language_creators-expert-generated #multilinguality-monolingual #size_categories-10K<n<100K #source_datasets-original #language-English #license-other #head-pose-estimation #region-us \n", "# Dataset Card for Biwi Kinect Head Pose Database", "## Table of Contents\n- Dataset Description\n - Dataset Summary\n - Supported Tasks\n - Languages\n- Dataset Structure\n - Data Instances\n - Data Fields\n - Data Splits\n- Dataset Creation\n - Curation Rationale\n - Source Data\n - Annotations\n - Personal and Sensitive Information\n- Considerations for Using the Data\n - Social Impact of Dataset\n - Discussion of Biases\n - Other Known Limitations\n- Additional Information\n - Dataset Curators\n - Licensing Information\n - Citation Information", "## Dataset Description\n\n- Homepage: Biwi Kinect Head Pose homepage\n- Repository: \n- Paper: Biwi Kinect Head Pose paper\n- Leaderboard: \n- Point of Contact: Gabriele Fanelli", "### Dataset Summary\n\nThe Biwi Kinect Head Pose Database is acquired with the Microsoft Kinect sensor, a structured IR light device.It contains 15K images of 20 people with 6 females and 14 males where 4 people were recorded twice.\n\nFor each frame, there is :\n- a depth image,\n- a corresponding rgb image (both 640x480 pixels),\n- annotation\n\nThe head pose range covers about +-75 degrees yaw and +-60 degrees pitch. The ground truth is the 3D location of the head and its rotation.", "### Data Processing\n\nExample code for reading a compressed binary depth image file provided by the authors.\n\n<details>\n <summary> View C++ Code </summary>\n\n\n\n</details>", "### Supported Tasks and Leaderboards\n\nBiwi Kinect Head Pose Database supports the following tasks :\n- Head pose estimation\n- Pose estimation\n- Face verification", "### Languages", "## Dataset Structure", "### Data Instances\n\nA sample from the Biwi Kinect Head Pose dataset is provided below:", "### Data Fields\n\n- 'sequence_number' : This refers to the sequence number in the dataset. There are a total of 24 sequences.\n- 'subject_id' : This refers to the subjects in the dataset. There are a total of 20 people with 6 females and 14 males where 4 people were recorded twice.\n- 'rgb' : List of png frames containing the poses.\n- 'rgb_cal': Contains calibration information for the color camera which includes intrinsic matrix, \nglobal rotation and translation.\n- 'depth' : List of depth frames for the poses.\n- 'depth_cal': Contains calibration information for the depth camera which includes intrinsic matrix, global rotation and translation.\n- 'head_pose_gt' : Contains ground truth information, i.e., the location of the center of the head in 3D and the head rotation, encoded as a 3x3 rotation matrix.", "### Data Splits\n\nAll the data is contained in the training set.", "## Dataset Creation", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization\n\nThe Biwi Kinect Head Pose Database is acquired with the Microsoft Kinect sensor, a structured IR light device.", "#### Who are the source language producers?", "### Annotations", "#### Annotation process\n\nFrom Dataset's README : \n> The database contains 24 sequences acquired with a Kinect sensor. 20 people (some were recorded twice - 6 women and 14 men) were recorded while turning their heads, sitting in front of the sensor, at roughly one meter of distance.", "#### Who are the annotators?", "### Personal and Sensitive Information", "## Considerations for Using the Data", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations", "## Additional Information", "### Dataset Curators", "### Licensing Information\n\nFrom Dataset's README : \n> This database is made available for non-commercial use such as university research and education.", "### Contributions\n\nThanks to @dnaveenr for adding this dataset." ]

[ "task_categories:text-classification", "task_ids:sentiment-classification", "multilinguality:monolingual", "size_categories:10K<n<100K", "language:ru", "region:us" ]

[]

[ "ru" ]

[ "### Dataset Summary\nThe dataset contains user reviews about restaurants.\nIn total it contains 47,139 reviews. A review tagged with the <em>general</em> sentiment and sentiments on 3 aspects: <em>food, interior, service</em>.", "### Data Fields\nEach sample contains the following fields:\n- review_id;\n- general;\n- food;\n- interior;\n- service;\n- text review text.", "### Python" ]

[ "TAGS\n#task_categories-text-classification #task_ids-sentiment-classification #multilinguality-monolingual #size_categories-10K<n<100K #language-Russian #region-us \n", "### Dataset Summary\nThe dataset contains user reviews about restaurants.\nIn total it contains 47,139 reviews. A review tagged with the <em>general</em> sentiment and sentiments on 3 aspects: <em>food, interior, service</em>.", "### Data Fields\nEach sample contains the following fields:\n- review_id;\n- general;\n- food;\n- interior;\n- service;\n- text review text.", "### Python" ]

[ "license:cc-by-sa-4.0", "region:us" ]

[]

[]

[ "# Discursos Perón\nDiscursos completos pronunciados por el ex Presidente Juan Domingo Perón entre 1ro de diciembre de 1943 y el 19 de septiembre de 1955. \n\nLos documentos, con excepción de los correspondientes al año 1949, fueron suministrados por el historiador Enrique de Alzáa, quien lideró un equipo que transcribió a formato digital editable los originales en papel que se encuentran en el Archivo General de la Nación. Los discursos del año 1949 fueron tomados de Perón (2016) en formato PDF. \n\nDado que este trabajo se realizó hace varios años y en distintas épocas, los documentos recibidos corresponden a tres versiones diferentes de\ndocumentos de Microsoft Word. Los discursos del año 1949 fueron tomados de Perón (2016)^1 en formato PDF.n La variedad y tipo de formatos de los documentos originales requirió un extenso trabajo de manipulación, limpieza y ordenamiento de los datos. Para más información sobre el preprocesamiento referirse aquí.", "# Información de licenciamiento\n\nEste conjunto de datos está licenciado bajo la licencia internacional Creative Commons Attribution-ShareAlike 4.0 CC BY-SA 4.0.", "# Información de citado\n\n\n\n---\n^1: Perón, J. D. (2016). Discursos, mensajes, correspondencia y escritos: 1949 / Perón (Tomos I y II). Buenos Aires, Argentina: Biblioteca del Congreso de la Nación." ]

[ "TAGS\n#license-cc-by-sa-4.0 #region-us \n", "# Discursos Perón\nDiscursos completos pronunciados por el ex Presidente Juan Domingo Perón entre 1ro de diciembre de 1943 y el 19 de septiembre de 1955. \n\nLos documentos, con excepción de los correspondientes al año 1949, fueron suministrados por el historiador Enrique de Alzáa, quien lideró un equipo que transcribió a formato digital editable los originales en papel que se encuentran en el Archivo General de la Nación. Los discursos del año 1949 fueron tomados de Perón (2016) en formato PDF. \n\nDado que este trabajo se realizó hace varios años y en distintas épocas, los documentos recibidos corresponden a tres versiones diferentes de\ndocumentos de Microsoft Word. Los discursos del año 1949 fueron tomados de Perón (2016)^1 en formato PDF.n La variedad y tipo de formatos de los documentos originales requirió un extenso trabajo de manipulación, limpieza y ordenamiento de los datos. Para más información sobre el preprocesamiento referirse aquí.", "# Información de licenciamiento\n\nEste conjunto de datos está licenciado bajo la licencia internacional Creative Commons Attribution-ShareAlike 4.0 CC BY-SA 4.0.", "# Información de citado\n\n\n\n---\n^1: Perón, J. D. (2016). Discursos, mensajes, correspondencia y escritos: 1949 / Perón (Tomos I y II). Buenos Aires, Argentina: Biblioteca del Congreso de la Nación." ]

[ "annotations_creators:crowdsourced", "language_creators:crowdsourced", "multilinguality:monolingual", "language:tg", "license:cc-by-4.0", "doi:10.57967/hf/0061", "region:us" ]

[]

[ "tg" ]

[]

[ "TAGS\n#annotations_creators-crowdsourced #language_creators-crowdsourced #multilinguality-monolingual #language-Tajik #license-cc-by-4.0 #doi-10.57967/hf/0061 #region-us \n" ]

[ "task_categories:text-classification", "task_ids:sentiment-classification", "annotations_creators:expert-generated", "language_creators:expert-generated", "multilinguality:monolingual", "size_categories:10K<n<100K", "source_datasets:extended|imdb", "language:en", "license:mit", "region:us" ]

[]

[ "en" ]

[ "# Dataset Card for 'reviews_with_drift'", "## Table of Contents\n- Table of Contents\n- Dataset Description\n - Dataset Summary\n - Supported Tasks and Leaderboards\n - Languages\n- Dataset Structure\n - Data Instances\n - Data Fields\n - Data Splits\n- Dataset Creation\n - Curation Rationale\n - Source Data\n - Annotations\n - Personal and Sensitive Information\n- Considerations for Using the Data\n - Social Impact of Dataset\n - Discussion of Biases\n - Other Known Limitations\n- Additional Information\n - Dataset Curators\n - Licensing Information\n - Citation Information\n - Contributions", "## Dataset Description", "### Dataset Summary\n\nThis dataset was crafted to be used in our tutorial [Link to the tutorial when ready]. It consists on a large Movie Review Dataset mixed with some reviews from a Hotel Review Dataset. The training/validation set are purely obtained from the Movie Review Dataset while the production set is mixed. Some other features have been added ('age', 'gender', 'context') as well as a made up timestamp 'prediction_ts' of when the inference took place.", "### Supported Tasks and Leaderboards\n\n'text-classification', 'sentiment-classification': The dataset is mainly used for text classification: given the text, predict the sentiment (positive or negative).", "### Languages\n\nText is mainly written in english.", "## Dataset Structure", "### Data Instances", "### Data Fields", "### Data Splits", "## Dataset Creation", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information", "## Considerations for Using the Data", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations", "## Additional Information", "### Dataset Curators", "### Licensing Information", "### Contributions\n\nThanks to @fjcasti1 for adding this dataset." ]

[ "TAGS\n#task_categories-text-classification #task_ids-sentiment-classification #annotations_creators-expert-generated #language_creators-expert-generated #multilinguality-monolingual #size_categories-10K<n<100K #source_datasets-extended|imdb #language-English #license-mit #region-us \n", "# Dataset Card for 'reviews_with_drift'", "## Table of Contents\n- Table of Contents\n- Dataset Description\n - Dataset Summary\n - Supported Tasks and Leaderboards\n - Languages\n- Dataset Structure\n - Data Instances\n - Data Fields\n - Data Splits\n- Dataset Creation\n - Curation Rationale\n - Source Data\n - Annotations\n - Personal and Sensitive Information\n- Considerations for Using the Data\n - Social Impact of Dataset\n - Discussion of Biases\n - Other Known Limitations\n- Additional Information\n - Dataset Curators\n - Licensing Information\n - Citation Information\n - Contributions", "## Dataset Description", "### Dataset Summary\n\nThis dataset was crafted to be used in our tutorial [Link to the tutorial when ready]. It consists on a large Movie Review Dataset mixed with some reviews from a Hotel Review Dataset. The training/validation set are purely obtained from the Movie Review Dataset while the production set is mixed. Some other features have been added ('age', 'gender', 'context') as well as a made up timestamp 'prediction_ts' of when the inference took place.", "### Supported Tasks and Leaderboards\n\n'text-classification', 'sentiment-classification': The dataset is mainly used for text classification: given the text, predict the sentiment (positive or negative).", "### Languages\n\nText is mainly written in english.", "## Dataset Structure", "### Data Instances", "### Data Fields", "### Data Splits", "## Dataset Creation", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information", "## Considerations for Using the Data", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations", "## Additional Information", "### Dataset Curators", "### Licensing Information", "### Contributions\n\nThanks to @fjcasti1 for adding this dataset." ]

[ "region:us" ]

[]

[]

[]

[ "TAGS\n#region-us \n" ]

[ "task_categories:fill-mask", "task_categories:text-generation", "task_ids:language-modeling", "task_ids:masked-language-modeling", "annotations_creators:no-annotation", "language_creators:found", "multilinguality:monolingual", "size_categories:10K<n<100K", "source_datasets:original", "language:en", "license:mit", "region:us" ]

[]

[ "en" ]

[ "### Dataset Summary\n\n\nThe enwik8 dataset is the first 100,000,000 (100M) bytes of the English Wikipedia XML dump on Mar. 3, 2006 and is typically used to measure a model's ability to compress data.", "### Supported Tasks and Leaderboards\n\n\nA leaderboard for byte-level causal language modelling can be found on paperswithcode", "### Languages\n\n\nen\n\n\nDataset Structure\n-----------------", "### Data Instances\n\n\n* Size of downloaded dataset files: 36.45 MB\n* Size of the generated dataset: 102.38 MB\n* Total amount of disk used: 138.83 MB", "### Data Fields\n\n\nThe data fields are the same among all sets.", "#### enwik8\n\n\n* 'text': a 'string' feature.", "#### enwik8-raw\n\n\n* 'text': a 'string' feature.", "### Data Splits\n\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization\n\n\nThe data is just English Wikipedia XML dump on Mar. 3, 2006 split by line for enwik8 and not split by line for enwik8-raw.", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations\n\n\nAdditional Information\n----------------------", "### Dataset Curators", "### Licensing Information\n\n\nDataset is not part of a publication, and can therefore not be cited.", "### Contributions\n\n\nThanks to @HallerPatrick for adding this dataset and @mtanghu for updating it." ]

[ "TAGS\n#task_categories-fill-mask #task_categories-text-generation #task_ids-language-modeling #task_ids-masked-language-modeling #annotations_creators-no-annotation #language_creators-found #multilinguality-monolingual #size_categories-10K<n<100K #source_datasets-original #language-English #license-mit #region-us \n", "### Dataset Summary\n\n\nThe enwik8 dataset is the first 100,000,000 (100M) bytes of the English Wikipedia XML dump on Mar. 3, 2006 and is typically used to measure a model's ability to compress data.", "### Supported Tasks and Leaderboards\n\n\nA leaderboard for byte-level causal language modelling can be found on paperswithcode", "### Languages\n\n\nen\n\n\nDataset Structure\n-----------------", "### Data Instances\n\n\n* Size of downloaded dataset files: 36.45 MB\n* Size of the generated dataset: 102.38 MB\n* Total amount of disk used: 138.83 MB", "### Data Fields\n\n\nThe data fields are the same among all sets.", "#### enwik8\n\n\n* 'text': a 'string' feature.", "#### enwik8-raw\n\n\n* 'text': a 'string' feature.", "### Data Splits\n\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization\n\n\nThe data is just English Wikipedia XML dump on Mar. 3, 2006 split by line for enwik8 and not split by line for enwik8-raw.", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations\n\n\nAdditional Information\n----------------------", "### Dataset Curators", "### Licensing Information\n\n\nDataset is not part of a publication, and can therefore not be cited.", "### Contributions\n\n\nThanks to @HallerPatrick for adding this dataset and @mtanghu for updating it." ]

[ "task_ids:intent-classification", "task_ids:multi-class-classification", "task_ids:sentiment-classification", "annotations_creators:diana_logan", "multilinguality:monolingual", "source_datasets:other-generated-datasets", "language:en", "license:apache-2.0", "arxiv:2010.12421", "region:us" ]

[ "2010.12421" ]

[ "en" ]

[ "### Dataset Summary\n\n\nTweetEval consists of seven heterogenous tasks in Twitter, all framed as multi-class tweet classification. The tasks include - irony, hate, offensive, stance, emoji, emotion, and sentiment. All tasks have been unified into the same benchmark, with each dataset presented in the same format and with fixed training, validation and test splits.", "### Supported Tasks and Leaderboards\n\n\n* 'text\\_classification': The dataset can be trained using a SentenceClassification model from HuggingFace transformers.", "### Languages\n\n\nThe text in the dataset is in English, as spoken by Twitter users.\n\n\nDataset Structure\n-----------------", "### Data Instances\n\n\nAn instance from 'emoji' config:\n\n\nAn instance from 'emotion' config:\n\n\nAn instance from 'hate' config:\n\n\nAn instance from 'irony' config:\n\n\nAn instance from 'offensive' config:\n\n\nAn instance from 'sentiment' config:\n\n\nAn instance from 'stance\\_abortion' config:\n\n\nAn instance from 'stance\\_atheism' config:\n\n\nAn instance from 'stance\\_climate' config:\n\n\nAn instance from 'stance\\_feminist' config:\n\n\nAn instance from 'stance\\_hillary' config:", "### Data Fields\n\n\nFor 'emoji' config:\n\n\n* 'text': a 'string' feature containing the tweet.\n* 'label': an 'int' classification label with the following mapping:\n'0':\n'1':\n'2':\n'3':\n'4':\n'5':\n'6':\n'7':\n'8':\n'9':\n'10':\n'11': 🇺🇸\n'12':\n'13':\n'14':\n'15':\n'16':\n'17':\n'18':\n'19':\nFor 'emotion' config:\n* 'text': a 'string' feature containing the tweet.\n* 'label': an 'int' classification label with the following mapping:\n'0': anger\n'1': joy\n'2': optimism\n'3': sadness\nFor 'hate' config:\n* 'text': a 'string' feature containing the tweet.\n* 'label': an 'int' classification label with the following mapping:\n'0': non-hate\n'1': hate\nFor 'irony' config:\n* 'text': a 'string' feature containing the tweet.\n* 'label': an 'int' classification label with the following mapping:\n'0': non\\_irony\n'1': irony\nFor 'offensive' config:\n* 'text': a 'string' feature containing the tweet.\n* 'label': an 'int' classification label with the following mapping:\n'0': non-offensive\n'1': offensive\nFor 'sentiment' config:\n* 'text': a 'string' feature containing the tweet.\n* 'label': an 'int' classification label with the following mapping:\n'0': negative\n'1': neutral\n'2': positive\nFor 'stance\\_abortion' config:\n* 'text': a 'string' feature containing the tweet.\n* 'label': an 'int' classification label with the following mapping:\n'0': none\n'1': against\n'2': favor\nFor 'stance\\_atheism' config:\n* 'text': a 'string' feature containing the tweet.\n* 'label': an 'int' classification label with the following mapping:\n'0': none\n'1': against\n'2': favor\nFor 'stance\\_climate' config:\n* 'text': a 'string' feature containing the tweet.\n* 'label': an 'int' classification label with the following mapping:\n'0': none\n'1': against\n'2': favor\nFor 'stance\\_feminist' config:\n* 'text': a 'string' feature containing the tweet.\n* 'label': an 'int' classification label with the following mapping:\n'0': none\n'1': against\n'2': favor\nFor 'stance\\_hillary' config:\n* 'text': a 'string' feature containing the tweet.\n* 'label': an 'int' classification label with the following mapping:\n'0': none\n'1': against\n'2': favor", "### Data Splits\n\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations\n\n\nAdditional Information\n----------------------", "### Dataset Curators\n\n\nFrancesco Barbieri, Jose Camacho-Collados, Luis Espiinosa-Anke and Leonardo Neves through Cardiff NLP.", "### Licensing Information\n\n\nThis is not a single dataset, therefore each subset has its own license (the collection itself does not have additional restrictions).\nAll of the datasets require complying with Twitter Terms Of Service and Twitter API Terms Of Service\nAdditionally the license are:\n\n\n* emoji: Undefined\n* emotion(EmoInt): Undefined\n* hate (HateEval): Need permission here\n* irony: Undefined\n* Offensive: Undefined\n* Sentiment: Creative Commons Attribution 3.0 Unported License\n* Stance: Undefined\n\n\nIf you use any of the TweetEval datasets, please cite their original publications:", "#### Emotion Recognition:", "#### Emoji Prediction:", "#### Irony Detection:", "#### Hate Speech Detection:", "#### Offensive Language Identification:", "#### Sentiment Analysis:", "#### Stance Detection:" ]

[ "TAGS\n#task_ids-intent-classification #task_ids-multi-class-classification #task_ids-sentiment-classification #annotations_creators-diana_logan #multilinguality-monolingual #source_datasets-other-generated-datasets #language-English #license-apache-2.0 #arxiv-2010.12421 #region-us \n", "### Dataset Summary\n\n\nTweetEval consists of seven heterogenous tasks in Twitter, all framed as multi-class tweet classification. The tasks include - irony, hate, offensive, stance, emoji, emotion, and sentiment. All tasks have been unified into the same benchmark, with each dataset presented in the same format and with fixed training, validation and test splits.", "### Supported Tasks and Leaderboards\n\n\n* 'text\\_classification': The dataset can be trained using a SentenceClassification model from HuggingFace transformers.", "### Languages\n\n\nThe text in the dataset is in English, as spoken by Twitter users.\n\n\nDataset Structure\n-----------------", "### Data Instances\n\n\nAn instance from 'emoji' config:\n\n\nAn instance from 'emotion' config:\n\n\nAn instance from 'hate' config:\n\n\nAn instance from 'irony' config:\n\n\nAn instance from 'offensive' config:\n\n\nAn instance from 'sentiment' config:\n\n\nAn instance from 'stance\\_abortion' config:\n\n\nAn instance from 'stance\\_atheism' config:\n\n\nAn instance from 'stance\\_climate' config:\n\n\nAn instance from 'stance\\_feminist' config:\n\n\nAn instance from 'stance\\_hillary' config:", "### Data Fields\n\n\nFor 'emoji' config:\n\n\n* 'text': a 'string' feature containing the tweet.\n* 'label': an 'int' classification label with the following mapping:\n'0':\n'1':\n'2':\n'3':\n'4':\n'5':\n'6':\n'7':\n'8':\n'9':\n'10':\n'11': 🇺🇸\n'12':\n'13':\n'14':\n'15':\n'16':\n'17':\n'18':\n'19':\nFor 'emotion' config:\n* 'text': a 'string' feature containing the tweet.\n* 'label': an 'int' classification label with the following mapping:\n'0': anger\n'1': joy\n'2': optimism\n'3': sadness\nFor 'hate' config:\n* 'text': a 'string' feature containing the tweet.\n* 'label': an 'int' classification label with the following mapping:\n'0': non-hate\n'1': hate\nFor 'irony' config:\n* 'text': a 'string' feature containing the tweet.\n* 'label': an 'int' classification label with the following mapping:\n'0': non\\_irony\n'1': irony\nFor 'offensive' config:\n* 'text': a 'string' feature containing the tweet.\n* 'label': an 'int' classification label with the following mapping:\n'0': non-offensive\n'1': offensive\nFor 'sentiment' config:\n* 'text': a 'string' feature containing the tweet.\n* 'label': an 'int' classification label with the following mapping:\n'0': negative\n'1': neutral\n'2': positive\nFor 'stance\\_abortion' config:\n* 'text': a 'string' feature containing the tweet.\n* 'label': an 'int' classification label with the following mapping:\n'0': none\n'1': against\n'2': favor\nFor 'stance\\_atheism' config:\n* 'text': a 'string' feature containing the tweet.\n* 'label': an 'int' classification label with the following mapping:\n'0': none\n'1': against\n'2': favor\nFor 'stance\\_climate' config:\n* 'text': a 'string' feature containing the tweet.\n* 'label': an 'int' classification label with the following mapping:\n'0': none\n'1': against\n'2': favor\nFor 'stance\\_feminist' config:\n* 'text': a 'string' feature containing the tweet.\n* 'label': an 'int' classification label with the following mapping:\n'0': none\n'1': against\n'2': favor\nFor 'stance\\_hillary' config:\n* 'text': a 'string' feature containing the tweet.\n* 'label': an 'int' classification label with the following mapping:\n'0': none\n'1': against\n'2': favor", "### Data Splits\n\n\n\nDataset Creation\n----------------", "### Curation Rationale", "### Source Data", "#### Initial Data Collection and Normalization", "#### Who are the source language producers?", "### Annotations", "#### Annotation process", "#### Who are the annotators?", "### Personal and Sensitive Information\n\n\nConsiderations for Using the Data\n---------------------------------", "### Social Impact of Dataset", "### Discussion of Biases", "### Other Known Limitations\n\n\nAdditional Information\n----------------------", "### Dataset Curators\n\n\nFrancesco Barbieri, Jose Camacho-Collados, Luis Espiinosa-Anke and Leonardo Neves through Cardiff NLP.", "### Licensing Information\n\n\nThis is not a single dataset, therefore each subset has its own license (the collection itself does not have additional restrictions).\nAll of the datasets require complying with Twitter Terms Of Service and Twitter API Terms Of Service\nAdditionally the license are:\n\n\n* emoji: Undefined\n* emotion(EmoInt): Undefined\n* hate (HateEval): Need permission here\n* irony: Undefined\n* Offensive: Undefined\n* Sentiment: Creative Commons Attribution 3.0 Unported License\n* Stance: Undefined\n\n\nIf you use any of the TweetEval datasets, please cite their original publications:", "#### Emotion Recognition:", "#### Emoji Prediction:", "#### Irony Detection:", "#### Hate Speech Detection:", "#### Offensive Language Identification:", "#### Sentiment Analysis:", "#### Stance Detection:" ]

[ "task_categories:text-generation", "task_ids:language-modeling", "multilinguality:monolingual", "size_categories:10K<n<100K", "source_datasets:subjqa", "language:en", "license:cc-by-4.0", "question-generation", "arxiv:2210.03992", "region:us" ]

[ "2210.03992" ]

[ "en" ]

[ "### Dataset Summary\n\n\nThis is a subset of QG-Bench, a unified question generation benchmark proposed in\n\"Generative Language Models for Paragraph-Level Question Generation: A Unified Benchmark and Evaluation, EMNLP 2022 main conference\".\nModified version of SQuADShifts for question generation (QG) task.", "### Supported Tasks and Leaderboards\n\n\n* 'question-generation': The dataset can be used to train a model for question generation.\nSuccess on this task is typically measured by achieving a high BLEU4/METEOR/ROUGE-L/BERTScore/MoverScore (see our paper for more in detail).", "### Languages\n\n\nEnglish (en)\n\n\nDataset Structure\n-----------------\n\n\nAn example of 'train' looks as follows.\n\n\nThe data fields are the same among all splits.\n\n\n* 'question': a 'string' feature.\n* 'paragraph': a 'string' feature.\n* 'answer': a 'string' feature.\n* 'sentence': a 'string' feature.\n* 'paragraph\\_answer': a 'string' feature, which is same as the paragraph but the answer is highlighted by a special token ''.\n* 'paragraph\\_sentence': a 'string' feature, which is same as the paragraph but a sentence containing the answer is highlighted by a special token ''.\n* 'sentence\\_answer': a 'string' feature, which is same as the sentence but the answer is highlighted by a special token ''.\n\n\nEach of 'paragraph\\_answer', 'paragraph\\_sentence', and 'sentence\\_answer' feature is assumed to be used to train a question generation model,\nbut with different information. The 'paragraph\\_answer' and 'sentence\\_answer' features are for answer-aware question generation and\n'paragraph\\_sentence' feature is for sentence-aware question generation.", "### Data Splits" ]

[ "TAGS\n#task_categories-text-generation #task_ids-language-modeling #multilinguality-monolingual #size_categories-10K<n<100K #source_datasets-subjqa #language-English #license-cc-by-4.0 #question-generation #arxiv-2210.03992 #region-us \n", "### Dataset Summary\n\n\nThis is a subset of QG-Bench, a unified question generation benchmark proposed in\n\"Generative Language Models for Paragraph-Level Question Generation: A Unified Benchmark and Evaluation, EMNLP 2022 main conference\".\nModified version of SQuADShifts for question generation (QG) task.", "### Supported Tasks and Leaderboards\n\n\n* 'question-generation': The dataset can be used to train a model for question generation.\nSuccess on this task is typically measured by achieving a high BLEU4/METEOR/ROUGE-L/BERTScore/MoverScore (see our paper for more in detail).", "### Languages\n\n\nEnglish (en)\n\n\nDataset Structure\n-----------------\n\n\nAn example of 'train' looks as follows.\n\n\nThe data fields are the same among all splits.\n\n\n* 'question': a 'string' feature.\n* 'paragraph': a 'string' feature.\n* 'answer': a 'string' feature.\n* 'sentence': a 'string' feature.\n* 'paragraph\\_answer': a 'string' feature, which is same as the paragraph but the answer is highlighted by a special token ''.\n* 'paragraph\\_sentence': a 'string' feature, which is same as the paragraph but a sentence containing the answer is highlighted by a special token ''.\n* 'sentence\\_answer': a 'string' feature, which is same as the sentence but the answer is highlighted by a special token ''.\n\n\nEach of 'paragraph\\_answer', 'paragraph\\_sentence', and 'sentence\\_answer' feature is assumed to be used to train a question generation model,\nbut with different information. The 'paragraph\\_answer' and 'sentence\\_answer' features are for answer-aware question generation and\n'paragraph\\_sentence' feature is for sentence-aware question generation.", "### Data Splits" ]

[ "region:us" ]

[]

[]

[ "# gets role for executing training job\nrole = sagemaker.get_execution_role()\nhyperparameters = {\n\t'model_name_or_path':'etmckinley/BERFALTER',\n\t'output_dir':'/opt/ml/model'\n\t# add your remaining hyperparameters\n\t# more info here URL\n}", "# git configuration to download our fine-tuning script\ngit_config = {'repo': 'URL 'v4.17.0'}", "# creates Hugging Face estimator\nhuggingface_estimator = HuggingFace(\n\tentry_point='run_qa.py',\n\tsource_dir='./examples/pytorch/question-answering',\n\tinstance_type='ml.p3.2xlarge',\n\tinstance_count=1,\n\trole=role,\n\tgit_config=git_config,\n\ttransformers_version='4.17.0',\n\tpytorch_version='1.10.2',\n\tpy_version='py38',\n\thyperparameters = hyperparameters\n)", "# starting the train job\nhuggingface_estimator.fit()" ]

[ "TAGS\n#region-us \n", "# gets role for executing training job\nrole = sagemaker.get_execution_role()\nhyperparameters = {\n\t'model_name_or_path':'etmckinley/BERFALTER',\n\t'output_dir':'/opt/ml/model'\n\t# add your remaining hyperparameters\n\t# more info here URL\n}", "# git configuration to download our fine-tuning script\ngit_config = {'repo': 'URL 'v4.17.0'}", "# creates Hugging Face estimator\nhuggingface_estimator = HuggingFace(\n\tentry_point='run_qa.py',\n\tsource_dir='./examples/pytorch/question-answering',\n\tinstance_type='ml.p3.2xlarge',\n\tinstance_count=1,\n\trole=role,\n\tgit_config=git_config,\n\ttransformers_version='4.17.0',\n\tpytorch_version='1.10.2',\n\tpy_version='py38',\n\thyperparameters = hyperparameters\n)", "# starting the train job\nhuggingface_estimator.fit()" ]

[ "task_categories:text-generation", "task_ids:language-modeling", "multilinguality:monolingual", "size_categories:10K<n<100K", "source_datasets:squad_es", "language:es", "license:cc-by-4.0", "question-generation", "arxiv:2210.03992", "region:us" ]

[ "2210.03992" ]

[ "es" ]

[ "### Dataset Summary\n\n\nThis is a subset of QG-Bench, a unified question generation benchmark proposed in\n\"Generative Language Models for Paragraph-Level Question Generation: A Unified Benchmark and Evaluation, EMNLP 2022 main conference\".\nThis is a modified version of SQuAD-es for question generation (QG) task.\nSince the original dataset only contains training/validation set, we manually sample test set from training set, which\nhas no overlap in terms of the paragraph with the training set.", "### Supported Tasks and Leaderboards\n\n\n* 'question-generation': The dataset is assumed to be used to train a model for question generation.\nSuccess on this task is typically measured by achieving a high BLEU4/METEOR/ROUGE-L/BERTScore/MoverScore (see our paper for more in detail).", "### Languages\n\n\nSpanish (es)\n\n\nDataset Structure\n-----------------\n\n\nAn example of 'train' looks as follows.\n\n\nThe data fields are the same among all splits.\n\n\n* 'question': a 'string' feature.\n* 'paragraph': a 'string' feature.\n* 'answer': a 'string' feature.\n* 'sentence': a 'string' feature.\n* 'paragraph\\_answer': a 'string' feature, which is same as the paragraph but the answer is highlighted by a special token ''.\n* 'paragraph\\_sentence': a 'string' feature, which is same as the paragraph but a sentence containing the answer is highlighted by a special token ''.\n* 'sentence\\_answer': a 'string' feature, which is same as the sentence but the answer is highlighted by a special token ''.\n\n\nEach of 'paragraph\\_answer', 'paragraph\\_sentence', and 'sentence\\_answer' feature is assumed to be used to train a question generation model,\nbut with different information. The 'paragraph\\_answer' and 'sentence\\_answer' features are for answer-aware question generation and\n'paragraph\\_sentence' feature is for sentence-aware question generation.\n\n\nData Splits\n-----------" ]

[ "TAGS\n#task_categories-text-generation #task_ids-language-modeling #multilinguality-monolingual #size_categories-10K<n<100K #source_datasets-squad_es #language-Spanish #license-cc-by-4.0 #question-generation #arxiv-2210.03992 #region-us \n", "### Dataset Summary\n\n\nThis is a subset of QG-Bench, a unified question generation benchmark proposed in\n\"Generative Language Models for Paragraph-Level Question Generation: A Unified Benchmark and Evaluation, EMNLP 2022 main conference\".\nThis is a modified version of SQuAD-es for question generation (QG) task.\nSince the original dataset only contains training/validation set, we manually sample test set from training set, which\nhas no overlap in terms of the paragraph with the training set.", "### Supported Tasks and Leaderboards\n\n\n* 'question-generation': The dataset is assumed to be used to train a model for question generation.\nSuccess on this task is typically measured by achieving a high BLEU4/METEOR/ROUGE-L/BERTScore/MoverScore (see our paper for more in detail).", "### Languages\n\n\nSpanish (es)\n\n\nDataset Structure\n-----------------\n\n\nAn example of 'train' looks as follows.\n\n\nThe data fields are the same among all splits.\n\n\n* 'question': a 'string' feature.\n* 'paragraph': a 'string' feature.\n* 'answer': a 'string' feature.\n* 'sentence': a 'string' feature.\n* 'paragraph\\_answer': a 'string' feature, which is same as the paragraph but the answer is highlighted by a special token ''.\n* 'paragraph\\_sentence': a 'string' feature, which is same as the paragraph but a sentence containing the answer is highlighted by a special token ''.\n* 'sentence\\_answer': a 'string' feature, which is same as the sentence but the answer is highlighted by a special token ''.\n\n\nEach of 'paragraph\\_answer', 'paragraph\\_sentence', and 'sentence\\_answer' feature is assumed to be used to train a question generation model,\nbut with different information. The 'paragraph\\_answer' and 'sentence\\_answer' features are for answer-aware question generation and\n'paragraph\\_sentence' feature is for sentence-aware question generation.\n\n\nData Splits\n-----------" ]

[ "task_categories:text-generation", "task_ids:language-modeling", "multilinguality:monolingual", "size_categories:10K<n<100K", "source_datasets:squad_es", "language:ko", "license:cc-by-4.0", "question-generation", "arxiv:2210.03992", "region:us" ]

[ "2210.03992" ]

[ "ko" ]

[ "### Dataset Summary\n\n\nThis is a subset of QG-Bench, a unified question generation benchmark proposed in\n\"Generative Language Models for Paragraph-Level Question Generation: A Unified Benchmark and Evaluation, EMNLP 2022 main conference\".\nThis is a modified version of KorQuAD for question generation (QG) task.\nSince the original dataset only contains training/validation set, we manually sample test set from training set, which\nhas no overlap in terms of the paragraph with the training set.", "### Supported Tasks and Leaderboards\n\n\n* 'question-generation': The dataset is assumed to be used to train a model for question generation.\nSuccess on this task is typically measured by achieving a high BLEU4/METEOR/ROUGE-L/BERTScore/MoverScore (see our paper for more in detail).", "### Languages\n\n\nKorean (ko)\n\n\nDataset Structure\n-----------------\n\n\nAn example of 'train' looks as follows.\n\n\nThe data fields are the same among all splits.\n\n\n* 'question': a 'string' feature.\n* 'paragraph': a 'string' feature.\n* 'answer': a 'string' feature.\n* 'sentence': a 'string' feature.\n* 'paragraph\\_answer': a 'string' feature, which is same as the paragraph but the answer is highlighted by a special token ''.\n* 'paragraph\\_sentence': a 'string' feature, which is same as the paragraph but a sentence containing the answer is highlighted by a special token ''.\n* 'sentence\\_answer': a 'string' feature, which is same as the sentence but the answer is highlighted by a special token ''.\n\n\nEach of 'paragraph\\_answer', 'paragraph\\_sentence', and 'sentence\\_answer' feature is assumed to be used to train a question generation model,\nbut with different information. The 'paragraph\\_answer' and 'sentence\\_answer' features are for answer-aware question generation and\n'paragraph\\_sentence' feature is for sentence-aware question generation.\n\n\nData Splits\n-----------" ]

[ "TAGS\n#task_categories-text-generation #task_ids-language-modeling #multilinguality-monolingual #size_categories-10K<n<100K #source_datasets-squad_es #language-Korean #license-cc-by-4.0 #question-generation #arxiv-2210.03992 #region-us \n", "### Dataset Summary\n\n\nThis is a subset of QG-Bench, a unified question generation benchmark proposed in\n\"Generative Language Models for Paragraph-Level Question Generation: A Unified Benchmark and Evaluation, EMNLP 2022 main conference\".\nThis is a modified version of KorQuAD for question generation (QG) task.\nSince the original dataset only contains training/validation set, we manually sample test set from training set, which\nhas no overlap in terms of the paragraph with the training set.", "### Supported Tasks and Leaderboards\n\n\n* 'question-generation': The dataset is assumed to be used to train a model for question generation.\nSuccess on this task is typically measured by achieving a high BLEU4/METEOR/ROUGE-L/BERTScore/MoverScore (see our paper for more in detail).", "### Languages\n\n\nKorean (ko)\n\n\nDataset Structure\n-----------------\n\n\nAn example of 'train' looks as follows.\n\n\nThe data fields are the same among all splits.\n\n\n* 'question': a 'string' feature.\n* 'paragraph': a 'string' feature.\n* 'answer': a 'string' feature.\n* 'sentence': a 'string' feature.\n* 'paragraph\\_answer': a 'string' feature, which is same as the paragraph but the answer is highlighted by a special token ''.\n* 'paragraph\\_sentence': a 'string' feature, which is same as the paragraph but a sentence containing the answer is highlighted by a special token ''.\n* 'sentence\\_answer': a 'string' feature, which is same as the sentence but the answer is highlighted by a special token ''.\n\n\nEach of 'paragraph\\_answer', 'paragraph\\_sentence', and 'sentence\\_answer' feature is assumed to be used to train a question generation model,\nbut with different information. The 'paragraph\\_answer' and 'sentence\\_answer' features are for answer-aware question generation and\n'paragraph\\_sentence' feature is for sentence-aware question generation.\n\n\nData Splits\n-----------" ]