[TODO] Todo: Add link

[wannaphong]

ASR

• ASR2K: Speech Recognition for Around 2000 Languages without Audio https://arxiv.org/abs/2209.02842
• Whisper: Whisper is a general-purpose speech recognition model. https://github.com/openai/whisper

Dataset

• FLEURS: Fleurs is the speech version of the FLoRes machine translation benchmark https://huggingface.co/datasets/google/fleurs

LM

• mLUKE https://huggingface.co/studio-ousia/mluke-base

Speech

• CharsiuG2P: Multilingual G2P in 100 languages https://github.com/lingjzhu/CharsiuG2P

Text corpus

• Multilingual Open Text (MOT) https://github.com/bltlab/mot/
• Thai depression detection dataset and baseline models https://zenodo.org/record/4734552

Coreference resolution

• 🪿 Han-Coref: Thai Coreference resolution by PyThaiNLP GitHub

[wannaphong]

mOSCAR: A Large-scale Multilingual and Multimodal Document-level Corpus

Paper: https://arxiv.org/abs/2406.08707
Dataset: https://huggingface.co/datasets/oscar-corpus/mOSCAR