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: 2056345033.625
num_examples: 13867
download_size: 1976130547
dataset_size: 2056345033.625
configs:
- config_name: default
data_files:
- split: train
path: data/train-*