---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: id
    dtype: string
  - name: question
    dtype: string
  - name: long_answer_type
    dtype: string
  - name: url
    dtype: string
  - name: short_answer
    dtype: string
  - name: short_answer_type
    dtype: string
  - name: long_answer
    dtype: string
  - name: image_size
    sequence: int64
  - name: candidates
    sequence: string
  - name: pos_idx
    dtype: int64
  - name: bounding_box
    sequence: int64
  splits:
  - name: train
    num_bytes: 97077639271.282
    num_examples: 87031
  - name: test
    num_bytes: 3416932974.0
    num_examples: 3000
  download_size: 100338121313
  dataset_size: 100494572245.282
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
---