metadata
dataset_info:
features:
- name: audio
dtype: audio
- name: score
dtype: float64
- name: transcripts
dtype: string
splits:
- name: train
num_bytes: 21426678953.212
num_examples: 146916
download_size: 19358465696
dataset_size: 21426678953.212
configs:
- config_name: default
data_files:
- split: train
path: data/train-*