---
dataset_info:
  features:
  - name: audio_token
    sequence:
      sequence:
        sequence: int64
  - name: audio_duration
    dtype: float64
  - name: nemo
    dtype: string
  - name: cer
    dtype: float64
  - name: text
    dtype: string
  splits:
  - name: train
    num_bytes: 23162502823
    num_examples: 8282987
  download_size: 5158446109
  dataset_size: 23162502823
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---