03_Whisper_datasets / README.md
jindol's picture
Upload dataset
f5b0c2f verified
metadata
dataset_info:
  features:
    - name: audio
      dtype:
        audio:
          sampling_rate: 16000
    - name: transcripts
      dtype: string
  splits:
    - name: train
      num_bytes: 3215899.2
      num_examples: 24
    - name: test
      num_bytes: 372035.4
      num_examples: 3
    - name: valid
      num_bytes: 463315.4
      num_examples: 3
  download_size: 3560654
  dataset_size: 4051250
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: test
        path: data/test-*
      - split: valid
        path: data/valid-*