metadata
language:
- ms
- zh
dataset_info:
features:
- name: audio_filename
dtype: string
- name: prompt
dtype: string
- name: transcription
dtype: string
splits:
- name: train
num_bytes: 115507240
num_examples: 646010
- name: test
num_bytes: 7253
num_examples: 40
download_size: 55523846
dataset_size: 115514493
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
This dataset in HuggingFace format combination of multiple speakers from https://huggingface.co/datasets/mesolitica/TTS
Audio files compressed at https://huggingface.co/datasets/mesolitica/TTS/tree/main/processed