metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 24000
- name: transcript
dtype: string
- name: audio_duration
dtype: float32
- name: audio_id
dtype: string
splits:
- name: test_clean
num_bytes: 934011889
num_examples: 87
- name: test_other
num_bytes: 923304320
num_examples: 90
download_size: 3611377651
dataset_size: 1857316209
configs:
- config_name: default
data_files:
- split: test_clean
path: data/test_clean-*
- split: test_other
path: data/test_other-*