Datasets

Published Datasets

• AfroBench published at Under review for 64 African languages.

• GlobalMMLU published at Under review for 42 languages.

• GlobalMMLU published at Under review for 42 languages on MMLU.

• Uhura published at Under review for QA on six African languages.

• TransWebEdu published at Under review for language modeling of 10 languages (machine translated).

• BRIGHTER published at Under review for emotion detection of 28 languages.

• INJONGO published at Under review for slot filling and intent detection of 16 African languages.

• AFRIDOC-MT published at Under review for document level MT for five African languages.

• WorldCuisines published at NAACL 2025 for food visual QA in 30 languages.

• IrokoBench published at NAACL 2025 for MMLU, math reasoning and NLI for 16 African languages.

• Warri published at NAACL Findings 2025 for machine translation of Nigerian-Pidgin.

• CVQA published at NeurIPS 2024 for visual QA for 31 languages.

• YORULECT published at EMNLP 2024 for three Yoruba dialects (Speech and MT).

• SIB-200 published at EACL 2024 for topic classification for 200+ languages.

• AfriMTE published at NAACL 2024 for machine translation evaluation of 13 African language pairs.

• EkoHate published at WOAH Workshop @NAACL 2024 for abusive and hate speech detection of Lagos election tweets.

• NaijaRC published at AfricaNLP Workshop @ICLR 2024 for reading comprehension QA for three Nigerian languages.

• AfriHG published at AfricaNLP Workshop @ICLR 2024 for news headline generation for 16 languages.

• YAD published at AfricaNLP Workshop @ICLR 2024 for diacritics restoration of Yoruba language.

• Wura published at EMNLP 2024 for language modelling of 16 African languages.

• IroyinSpeech published at LREC-COLING 2024 for ASR and TTS for Yoruba.

• XTREME-UP published at EMNLP Findings 2023 , a benchmark for 88 under-represented languages.

• AfriQA published at EMNLP 2023 for cross-lingual open retrieval QA of 10 African languages.

• NollySenti published at ACL 2023 for movie sentiment classification of 5 Nigerian languages.

• MasakhaPOS published at ACL 2023 for parts of speech tagging of 20 African languages.

• [ε ku ](https://github.com/ajesujoba/yoruba_greetings/tree/main/data) published at [C3NLP at EACL 2021](https://aclanthology.org/2023.c3nlp-1.1/) for cultural greetings in machine translation for Yoruba.

• MasakhaNEWS published at IJCNLP-AACL for topic classification of 16 African languages (Area Chair Award).

• MphayaNER published at AfricaNLP Workshop at ICLR 2023 for named entity recognition of Tshivenda.

• AfriSenti published at EMNLP 2023 for sentiment classification of 14 African languages

• MasakhaNER 2.0 published at EMNLP 2022 for named entity recognition of 20 African languages

• ANTC published at COLING 2022 for topic classification of 5 African languages (Best Paper Award (Grand Challenges))

• BibleTTS published at Interspeech 2022 for text-to-speech of 10 African languages

• MAFAND published at NAACL 2022 for machine translation of 21 African languages

• MasakhaNER published at TACL 2021 for named entity recognition of 10 African languages

• Menyo-20k published at MT Summit 2021 for machine translation of Yoruba

• Hausa-VOA-NER, Hausa-VOA-Topics, and Yoruba-BBC-Topics published at EMNLP 2020 for named entity recognition and topic classification of Hausa and Yoruba

• Yoruba-NER published at LREC 2020 for named entity recognition of Yoruba