Multitask-National-Speech-Corpus-v1

Multitask-National-Speech-Corpus (MNSC v1) is derived from IMDA's NSC Corpus. MNSC is a multitask speech understanding dataset derived and further annotated from IMDA NSC Corpus. It focuses on the knowledge of Singapore's local accent, localised terms, and code-switching. ASR: Automatic Speech Recognition SQA: Speech Question Answering SDS: Spoken Dialogue Summarization PQA: Paralinguistic Question Answering from datasets import load_dataset data =… See the full description on the dataset page: https://huggingface.co/datasets/MERaLiON/Multitask-National-Speech-Corpus-v1.