---
dataset_info:
  features:
  - name: instruction
    dtype: string
  - name: input
    dtype: string
  - name: audio_id
    dtype: string
  - name: dataset
    dtype: string
  - name: task
    dtype: string
  - name: output
    dtype: string
  - name: audio
    struct:
    - name: array
      sequence: float64
    - name: path
      dtype: string
    - name: sampling_rate
      dtype: int64
  - name: question
    dtype: string
  - name: answer
    dtype: string
  - name: thinking
    dtype: string
  splits:
  - name: test
    num_bytes: 2663175112
    num_examples: 2101
  - name: train
    num_bytes: 16645783420
    num_examples: 13164
  download_size: 16570516834
  dataset_size: 19308958532
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
---