README.md

metadata

dataset_info:
  features:
    - name: audio
      dtype: audio
    - name: transcription_french
      dtype: string
    - name: transcription_creole
      dtype: string
    - name: prediction_small
      dtype: string
    - name: wer_small
      dtype: float64
    - name: prediction_large
      dtype: string
    - name: wer_large
      dtype: float64
  splits:
    - name: train
      num_bytes: 115513345.41
      num_examples: 1502
    - name: validation
      num_bytes: 81299189.507
      num_examples: 1311
  download_size: 142776148
  dataset_size: 196812534.917
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: validation
        path: data/validation-*