metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: transcripts
dtype: string
splits:
- name: train
num_bytes: 3121699647.0752296
num_examples: 36371
- name: test
num_bytes: 379950097.5313095
num_examples: 4547
- name: valid
num_bytes: 377414638.1824603
num_examples: 4546
download_size: 612123731
dataset_size: 3879064382.7889996
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: valid
path: data/valid-*