BEMBA_big_c / README.md
Beijuka's picture
Upload dataset
304930a verified
metadata
dataset_info:
  features:
    - name: transcription
      dtype: string
    - name: speaker_id
      dtype: int64
    - name: duration
      dtype: float64
    - name: audio
      dtype:
        audio:
          sampling_rate: 16000
  splits:
    - name: train
      num_bytes: 20255773612.548
      num_examples: 82371
    - name: valid
      num_bytes: 668655282.274
      num_examples: 2782
    - name: test
      num_bytes: 678701136.122
      num_examples: 2779
  download_size: 19796572013
  dataset_size: 21603130030.944
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: valid
        path: data/valid-*
      - split: test
        path: data/test-*