---
dataset_info:
  features:
  - name: audio
    dtype:
      audio:
        sampling_rate: 16000
  - name: logits
    sequence: float32
  splits:
  - name: train
    num_bytes: 686706457.5207071
    num_examples: 21420
  download_size: 637385666
  dataset_size: 686706457.5207071
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---