metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: transcripts
dtype: string
splits:
- name: train
num_bytes: 108047558909.34956
num_examples: 880971
- name: test
num_bytes: 13957212384.713346
num_examples: 110122
- name: valid
num_bytes: 14083017814.714104
num_examples: 110121
download_size: 130808669625
dataset_size: 136087789108.77701
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: valid
path: data/valid-*