Filtered Multitask-National-Speech-Corpus-v1
The original dataset at https://huggingface.co/datasets/MERaLiON/Multitask-National-Speech-Corpus-v1
how to prepare the dataset
The dataset already split among zip files, so you can unzip distributely,
# Speech
huggingface-cli download --repo-type dataset \
--include 'part*.zip' \
--local-dir './' \
Revolab/filter-Multitask-National-Speech-Corpus-v1
# originally from https://gist.github.com/huseinzol05/2e26de4f3b29d99e993b349864ab6c10
wget https://gist.githubusercontent.com/huseinzol05/2e26de4f3b29d99e993b349864ab6c10/raw/33c840ccab1da41f4500652c8ee2decd09a085cf/unzip.py
python3 unzip.py
# Labels
huggingface-cli download --repo-type dataset \
--include 'data/*.parquet' \
--local-dir './' \
Revolab/filter-Multitask-National-Speech-Corpus-v1
Source code
Source code at https://github.com/Revolab-Sdn-Bhd/Speech-To-Text/tree/ML-23-filter-multitask-national-speech-corpus-v-1-dataset
- Downloads last month
- 5