metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: speaker_id
dtype: string
- name: transcript
dtype: string
- name: language
dtype: string
- name: gender
dtype: string
- name: age_bracket
dtype: string
- name: duration
dtype: float32
- name: snr
dtype: float32
- name: phase
dtype: string
splits:
- name: train
num_bytes: 530647936
num_examples: 5000
download_size: 498853600
dataset_size: 530647936
configs:
- config_name: default
data_files:
- split: train
path: data/train-*