Create dataset loader for SEACrowd Instruct Multi-task Collection #723
Description
Dataloader name: seacrowd_instruct/seacrowd_instruct.py
DataCatalogue: http://seacrowd.github.io/seacrowd-catalogue/card.html?seacrowd_instruct
| Dataset | seacrowd_instruct |
|---|---|
| Description | The SEACrowd Instruct Multi-task Collection is a multi-task, instruction-formatted compilation of 29 preexisting datasets from SEACrowd. The collection comprises 332,040 question-answer pairs, to integrate the permissively licensed datasets in SEACrowd into the Data Provenance Explorer tool as well as multi-task instruction fine-tuning. |
| Subsets | Bhinneka Korpus - Translated, CebuaNER - Translated, EMoTES-3K - Translated, Fake News Filipino - Translated, Filipino Hate Speech Tiktok - Text - Translated, Filipino Slang Spelling Normalization - Translated, id-vaccines-tweets - Translated, identifikasi-bahasa - Translated, IndoNER-Tourism - Translated, Indonesia-Chinese-MTRobustEval - Translated, LimeSoda - Translated, MABL - Translated, MKQA - Translated, Multilabel Multiclass Sentiment and Emotion Dataset from Indonesian Mobile Application Review - Translated, Myanmar (Burmese) Name Romanization with Alignment on Grapheme-Level - Translated, NTREX-128 - Translated, SPAMID-PAIR - Translated, Tagalog Profanity Dataset - Translated, Thai Depression Dataset - Translated, Thai Toxicity Tweet Corpus - Translated, Typhoon Yolanda Tweets - Translated, UIT-ViCTSD - Translated, UIT-ViOCD - Translated, Vietnamese Social Media Emotion Corpus (UIT-VSMEC) - Translated, ViHealthQA - Translated, Wisesight Thai Sentiment Corpus - Translated, Wongnai Reviews - Translated, XNLI - Translated, XStoryCloze - Translated |
| Languages | eng, cmn, ind, mya, vie, tha, fil, tgl, khg, hmv, hmf, hnj, lao, zlm, por, tam, yue, khm, jav, abs, ceb, day, xdy, aoz |
| Tasks | Instruction Tuning |
| License | Unknown (unknown) |
| Homepage | https://github.com/Data-Provenance-Initiative/Data-Provenance-Collection/blob/main/data_summaries/Seacrowd.json |
| HF URL | https://huggingface.co/datasets/minnieliang5/seacrowd |
| Paper URL | - |