metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: transcripts
dtype: string
splits:
- name: train
num_bytes: 9837599604.736
num_examples: 36836
- name: test
num_bytes: 1203550279.7205777
num_examples: 4605
- name: valid
num_bytes: 1267133175.147422
num_examples: 4604
download_size: 12155801598
dataset_size: 12308283059.604
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: valid
path: data/valid-*