---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: question
    dtype: string
  - name: answer
    dtype: string
  - name: ground_truth
    dtype: string
  splits:
  - name: train
    num_bytes: 1224460931.6
    num_examples: 3200
  - name: test
    num_bytes: 153123921.0
    num_examples: 400
  download_size: 1324055034
  dataset_size: 1377584852.6
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
---