---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: id
    dtype: string
  - name: question
    dtype: string
  - name: short_answer
    dtype: string
  - name: short_answer_type
    dtype: string
  - name: candidates
    sequence: string
  - name: pos_idx
    dtype: int64
  - name: bounding_box
    sequence: int64
  splits:
  - name: train
    num_bytes: 39215022423.0
    num_examples: 49888
  download_size: 39127284960
  dataset_size: 39215022423.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---