metadata

dataset_info:
  features:
    - name: segment
      dtype: string
    - name: audio
      dtype: audio
    - name: transcription
      dtype: string
  splits:
    - name: train
      num_bytes: 56700949
      num_examples: 350
  download_size: 56343980
  dataset_size: 56700949
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
task_categories:
  - text-to-speech
language:
  - fr
  - zh
size_categories:
  - n<1K

French Speech Dataset from Ici Japon

This dataset contains approximately 30 minutes of French speech with transcriptions, divided into several segments. The audio is sourced from the Ici Japon channel.