---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: text
    dtype: string
  - name: qa
    struct:
    - name: answer
      sequence: string
    - name: choices
      sequence:
        sequence: string
    - name: question
      sequence: string
  splits:
  - name: train
    num_bytes: 1445004497.462
    num_examples: 4102
  download_size: 1428203094
  dataset_size: 1445004497.462
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---