metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: transcripts
dtype: string
splits:
- name: train
num_bytes: 295905745877.416
num_examples: 3549172
- name: test
num_bytes: 36606988802.17158
num_examples: 443647
- name: valid
num_bytes: 36726015922.83043
num_examples: 443646
download_size: 297049626999
dataset_size: 369238750602.418
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: valid
path: data/valid-*