Thai-Understanding is an open-source repository that provides a solution for speech understanding in the Thai language. This repository includes:
Thai-SUP: The first open-source Thai speech understanding dataset, which includes over 1,000 hours of data across three tasks: Intent Classification (IC), Named Entity Recognition (NER), and Speech Rephrasing (SR).
XLSR-Thai: The first large-scale self-supervised learning (SSL)… See the full description on the dataset page: https://huggingface.co/datasets/mcshao/Thai-understanding.
