---
dataset_info:
  features:
  - name: audio
    dtype:
      audio:
        sampling_rate: 16000
  - name: reference
    dtype: string
  - name: perturbation
    dtype: string
  splits:
  - name: train
    num_bytes: 159944865.0
    num_examples: 1000
  download_size: 149952129
  dataset_size: 159944865.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---