metadata
dataset_info:
features:
- name: prediction
dtype: string
- name: reference
dtype: string
- name: wer
dtype: float64
splits:
- name: train
num_bytes: 4577581.776598236
num_examples: 38431
- name: test
num_bytes: 239023
num_examples: 2023
download_size: 3265705
dataset_size: 4816604.776598236
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*