metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: transcripts
dtype: string
splits:
- name: train
num_bytes: 17609433455.2
num_examples: 99200
- name: test
num_bytes: 2152518506.7
num_examples: 12400
- name: valid
num_bytes: 2326954936.3
num_examples: 12400
download_size: 20778654985
dataset_size: 22088906898.2
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: valid
path: data/valid-*