README.md

metadata

dataset_info:
  features:
    - name: audio
      dtype:
        audio:
          sampling_rate: 16000
    - name: transcription
      dtype: string
    - name: channel_url
      dtype: string
    - name: video_url
      dtype: string
    - name: title
      dtype: string
    - name: attempt
      dtype: int64
    - name: video_index
      dtype: string
  splits:
    - name: train
      num_bytes: 133296409
      num_examples: 833
  download_size: 127709545
  dataset_size: 133296409
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*