metadata
dataset_info:
features:
- name: audio
struct:
- name: bytes
dtype: binary
- name: transcript
dtype: string
splits:
- name: train
num_bytes: 39847651240
num_examples: 54951
- name: test
num_bytes: 4820085781
num_examples: 4956
download_size: 24204265715
dataset_size: 44667737021
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*