metadata
dataset_info:
features:
- name: audio
dtype: audio
- name: sampling_rate
dtype: int64
- name: transcript
dtype: string
- name: lang_spans
dtype: string
- name: dominant_lang
dtype: string
splits:
- name: train
num_bytes: 65813774891.4
num_examples: 205200
- name: test
num_bytes: 7283676534.6
num_examples: 22800
download_size: 69400698467
dataset_size: 73097451426
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*