2.1.0

Datasets Changes

• New: initial monash time series forecasting by @kashif in #3743
• New: Roman Urdu Hate Speech dataset by @bp-high in #3972
• New: Adversarial GLUE by @jxmorris12 in #3849
• New: MetaShift by @dnaveenr in #3900
• New: GSM8K by @jon-tow in #4103
• New: SBU Captions Photo by @thomasw21 in #4130
• Deprecated: Multilingual Librispeech - deprecate dataset in favor of facebook/multilingual_librispeechby @polinaeterna in #4060
• Update (BREAKING): TIMIT - Redirect users to download data manually from LDC by @lhoestq in #4145
• Update: Wikipedia by @albertvillanova in #3821 and #3989
• Update: conll2012_ontonotesv5 - Support streaming by @albertvillanova in #4002
• Update: daily_dialog - Support streaming by @albertvillanova in #4008
• Update: id_clickbait - Support streaming by @albertvillanova in #4014
• Update: blimp - Support streaming by @albertvillanova in #4016
• Update: scan - Support streaming by @albertvillanova in #4017
• Update: yelp_review_full - Replace data url by @lhoestq in #4018
• Update: yelp_polarity - Support streaming by @lhoestq in #4019
• Update: amazon_polarity - Replace data URL by @lhoestq in #4020
• Update: dbpedia_14 - Replace data url by @lhoestq in #4022
• Update: xtreme - Support streaming dataset for bucc18 config by @albertvillanova in #4026
• Update: yahoo_answers_topics - Replace data url by @lhoestq in #4023* Update: ASSIN 2 dataset - replace broken Google Drive URLS by links on github by @ruanchaves in #4004
• Update: xcopa - Support streaming by @albertvillanova in #4039
• Update: medical_dialog - Add configs with processed data by @albertvillanova in #4127
• Update: xtreme - Support streaming for udpos config by @albertvillanova in #4131
• Update: xtreme - Support streaming for PAWS-X config by @albertvillanova in #4132
• Update: xtreme - Support streaming for PAN-X config by @albertvillanova in #4135
• Update: SQuAD v2 - Use a constant for the articles regex by @bryant1410 in #4030
• Update: HANS - Support streaming by @mariosasko in #4155
• Fix: cats_vs_dogs - fix checksum error dataset by @albertvillanova in #4033
• Fix: xcopa - fix null checksum by @albertvillanova in #4034
• Fix: amazon_us_reviews - fix metadata - 4/4/2022 by @trentonstrong in #4092

Dataset Cards

• Updated annotations for nli_tr dataset by @e-budur in #4058
• Add missing label for emotion description by @lijiazheng99 in #4151
• Remove unncessary 'pylint disable' message in ReadMe by @Datta0 in #3955
• Improve RedCaps dataset card by @mariosasko in #4100
• Fix duplicate key in multi_news by @lhoestq in #4164

Datasets Tags and Search on the Hugging Face Hub

• Tasks alignment with models by @lhoestq in #4066
• Update datasets task tags to align tags with models by @lhoestq in #4067

Metrics Changes

• Xtreme-S Metrics by @patrickvonplaten in #3799
• Fix xtreme s metrics by @patrickvonplaten in #3957
• Avoid info log messages from transformers in FrugalScore metric by @albertvillanova in #3938
• Add exact match metric by @emibaylor in #3899
• Fix comet metric by @lhoestq in #3945
• Add zero_division argument to precision and recall metrics by @albertvillanova in #4035
• Support float data types in pearsonr/spearmanr metrics by @albertvillanova in #4054
• Remove GLEU metric by @emibaylor in #3949

Metric Cards

• Perplexity Metric Card by @emibaylor in #3905
• Create README.md by @sashavor in #3917
• Create README.md for CER metric by @sashavor in #3911
• Create README.md by @sashavor in #3944
• Update README.md by @sashavor in #3933
• Create SARI metric card by @sashavor in #3932
• Create MAUVE metric card by @sashavor in #3934
• Create CoVAL metric card by @sashavor in #3940
• Google BLEU Metric Card by @emibaylor in #3948
• Create metric card for BERTScore by @sashavor in #3966
• Rename wer to cer by @pmgautam in #4012
• Create metric card for XNLI by @sashavor in #4046
• Create metric card for the Code Eval metric by @sashavor in #4049
• Add TER metric card by @emibaylor in #3981
• BLEU metric card by @emibaylor in #3947
• Create metric card for CUAD by @sashavor in #4043
• Create metric card for METEOR by @sashavor in #4065
• Create a metric card for Competition MATH by @sashavor in #4073
• Create metric card for seqeval by @sashavor in #4070
• Create README.md by @sashavor in #3930
• Create metric card for Frugal Score by @sashavor in #4089
• Updating FrugalScore metric card by @sashavor in #4097
• Proposing WikiSplit metric card by @sashavor in #4098
• Fix formatting in BLEU metric card by @mariosasko in #4157

Documentation

• Doc maintenance by @stevhliu in #3926
• [Doc] Don't use v for version tags on GitHub by @sgugger in #3943
• Use templates for doc-builidng jobs by @sgugger in #3914
• Add align_labels_with_mapping docs by @stevhliu in #3931
• Add tip on how to speed up loading with ImageFolder by @mariosasko in #3980
• Fix main_classes docs index by @lhoestq in #3925
• More consistent references in docs by @mariosasko in #3988
• Docs maintenance by @stevhliu in #3999
• Add ROUGE Metric Card by @emibaylor in #4076
• Add chrF(++) Metric Card by @emibaylor in #4082
• Add SacreBLEU Metric Card by @emibaylor in #4083

General improvements and bug fixes

• Fix flatten of complex feature types by @mariosasko in #3723
• Fix flatten of Sequence feature type by @lhoestq in #3962
• Exclude Google Drive tests of the CI by @lhoestq in #3982
• Close PIL.Image file handler in Image.decode_example by @mariosasko in #3995
• Fix Faiss custom_index device by @albertvillanova in #3987
• Fix None issue with Sequence of dict by @lhoestq in #4010
• Update main readme by @lhoestq in #3927
• Fix map remove_columns on empty dataset by @lhoestq in #4021
• Fix Audio.encode_example() when writing an array by @polinaeterna in #3998
• Use audio feature in ASR task template by @lhoestq in #4006
• Improve out of bounds error message by @lhoestq in #4068
• Increase max retries for GitHub metrics by @albertvillanova in #4063
• Fix CLI dummy data generation by @albertvillanova in #4045
• Fix docs on audio feature installation by @albertvillanova in #4028
• Add installation instructions to image_process doc by @mariosasko in #4072
• Fix GithubMetricModuleFactory instantiation with None download_config by @albertvillanova in #4078
• Increase max retries for GitHub datasets by @albertvillanova in #4079
• Close parquet writer properly in push_to_hub by @lhoestq in #4081
• fix typo in rename_column error message by @hunterlang in #4095
• Fix BeamWriter output Parquet file by @albertvillanova in #4087
• Remove unused legacy Beam utils by @albertvillanova in #4088
• Hotfix failing CI tests on Windows by @albertvillanova in #4119
• Update security policy by @albertvillanova in #4111
• Avoid writing empty license files by @albertvillanova in #4090
• Support huggingface_hub 0.5 by @lhoestq in #4106
• Pretty print dataset info files by @mariosasko in #4116
• Add single dataset citations for TweetEval by @gchhablani in #4137
• Adjust path to datasets tutorial in How-To by @NimaBoscarino in #4147
• Applied index-filters on scores in search.py. by @vishalsrao in #3971
• More robust cast_to_python_objects in TypedSequence by @mariosasko in #4128
• Sync Features dictionaries by @mariosasko in #3997
• Avoid rate limit in update hub repositories by @lhoestq in #4167

New Contributors

• @bp-high made their first contribution in #3972
• @ruanchaves made their first contribution in #4004
• @pmgautam made their first contribution in #4012
• @hunterlang made their first contribution in #4095
• @trentonstrong made their first contribution in #4092
• @NimaBoscarino made their first contribution in #4147
• @jon-tow made their first contribution in #4103
• @lijiazheng99 made their first contribution in #4151
• @Datta0 made their first contribution in #3955
• @vishalsrao made their first contribution in #3971

Full Changelog: 2.0.0...2.1.0