---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: metadata
    struct:
    - name: file_name
      dtype: string
    - name: features
      struct:
      - name: article
        dtype: string
      - name: summary
        dtype: string
      - name: question
        dtype: string
      - name: abstractive_answer
        dtype: string
      - name: extractive_answer
        dtype: string
      - name: bbox
        list:
        - name: sentence
          dtype: string
        - name: bounding_box
          sequence: int64
      - name: image
        dtype: string
  splits:
  - name: train
    num_bytes: 221439345.403
    num_examples: 5407
  - name: validation
    num_bytes: 42724015.002
    num_examples: 1158
  - name: test
    num_bytes: 39131975.961
    num_examples: 1159
  download_size: 287632055
  dataset_size: 303295336.366
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
  - split: test
    path: data/test-*
---