metadata
dataset_info:
features:
- name: mic
dtype:
audio:
sampling_rate: 16000
- name: ref
dtype:
audio:
sampling_rate: 16000
- name: clean
dtype:
audio:
sampling_rate: 16000
- name: vad_label
list: int64
splits:
- name: train
num_bytes: 3867751260
num_examples: 3995
download_size: 3194615552
dataset_size: 3867751260
configs:
- config_name: default
data_files:
- split: train
path: data/train-*