metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: transcription
dtype: string
- name: source
dtype: string
- name: length_s
dtype: float64
- name: cleaned_transcription
dtype: string
splits:
- name: train
num_bytes: 13028688980
num_examples: 88712
- name: test
num_bytes: 330934008.084
num_examples: 1708
download_size: 12907433193
dataset_size: 13359622988.084
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*