---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: question
    dtype: string
  - name: most_common_answer
    dtype: string
  - name: answers
    list: string
  - name: ori_width
    dtype: int32
  - name: ori_height
    dtype: int32
  - name: patch_gt
    list: int32
  splits:
  - name: train
    num_bytes: 8087515
    num_examples: 500
  download_size: 7614300
  dataset_size: 8087515
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---