metadata
dataset_info:
features:
- name: audio
dtype: audio
- name: transcript
dtype: string
- name: description
dtype: string
- name: wer
dtype: float64
- name: hsc
dtype: float32
splits:
- name: train
num_bytes: 167728905
num_examples: 100
download_size: 139563871
dataset_size: 167728905
configs:
- config_name: default
data_files:
- split: train
path: data/train-*