metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: transcripts
dtype: string
splits:
- name: train
num_bytes: 23928882080.43397
num_examples: 183597
- name: test
num_bytes: 3043184467.253515
num_examples: 22950
- name: valid
num_bytes: 2962297261.103514
num_examples: 22950
download_size: 28492902413
dataset_size: 29934363808.791
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: valid
path: data/valid-*