---
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-*
---