metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: reference
dtype: string
- name: perturbation
dtype: string
splits:
- name: train
num_bytes: 159944865
num_examples: 1000
download_size: 149952129
dataset_size: 159944865
configs:
- config_name: default
data_files:
- split: train
path: data/train-*