metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: transcription
dtype: string
- name: reference
dtype: string
- name: prediction_01
dtype: string
splits:
- name: train
num_bytes: 2837918903.875
num_examples: 10369
download_size: 2539257678
dataset_size: 2837918903.875
configs:
- config_name: default
data_files:
- split: train
path: data/train-*