---
dataset_info:
  features:
  - name: audio
    dtype:
      audio:
        sampling_rate: 16000
  - name: logits
    sequence: float32
  splits:
  - name: train
    num_bytes: 660738566.3354514
    num_examples: 20610
  download_size: 575764354
  dataset_size: 660738566.3354514
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---