metadata
dataset_info:
features:
- name: audio
struct:
- name: bytes
dtype: binary
- name: transcript
dtype: string
splits:
- name: train
num_bytes: 11735583788.402905
num_examples: 21876
download_size: 9144291456
dataset_size: 11735583788.402905
configs:
- config_name: default
data_files:
- split: train
path: data/train-*