metadata
dataset_info:
features:
- name: id
dtype: string
- name: text
dtype: string
- name: audio
dtype:
audio:
sampling_rate: 16000
splits:
- name: train
num_examples: 70043
configs:
- config_name: default
data_files:
- split: train
path: data/*/*