---
dataset_info:
  features:
  - name: id
    dtype: string
  - name: question
    dtype: string
  - name: question_speech
    dtype: audio
  - name: question_unit
    sequence: int64
  - name: reconstructed_speech
    dtype: audio
  splits:
  - name: train
    num_bytes: 4441760.0
    num_examples: 32
  download_size: 4356127
  dataset_size: 4441760.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---