metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: transcripts
dtype: string
splits:
- name: train
num_bytes: 18234945060
num_examples: 101236
- name: test
num_bytes: 2410514563.634364
num_examples: 12655
- name: valid
num_bytes: 2300143023.7636356
num_examples: 12654
download_size: 21259055664
dataset_size: 22945602647.398003
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: valid
path: data/valid-*