metadata
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: codes
list: int32
- name: text
dtype: string
- name: speaker_id
dtype: string
- name: channel
dtype: string
- name: gender
dtype: string
- name: dialect
dtype: string
- name: tone
dtype: string
- name: duration_sec
dtype: float32
- name: audio_id
dtype: string
splits:
- name: train
num_bytes: 83658936047
num_examples: 440000
download_size: 81354614336
dataset_size: 83658936047