---
dataset_info:
  features:
  - name: conversations
    list:
    - name: from
      dtype: string
    - name: value
      dtype: string
  - name: image
    dtype: image
  - name: image_hash
    dtype: string
  - name: question
    dtype: string
  - name: options
    sequence: string
  - name: answer
    dtype: string
  splits:
  - name: train
    num_bytes: 3819138502.926
    num_examples: 2343
  - name: test
    num_bytes: 1332737153.0
    num_examples: 587
  download_size: 6811901310
  dataset_size: 5151875655.926001
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
---