README.md

metadata

dataset_info:
  features:
    - name: image
      dtype: image
    - name: image_filename
      dtype: string
    - name: query
      dtype: string
    - name: answer
      dtype: string
    - name: source
      dtype: string
  splits:
    - name: train
      num_bytes: 28217643498.125
      num_examples: 39463
  download_size: 7112536377
  dataset_size: 28217643498.125
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
task_categories:
  - document-question-answering
  - visual-document-retrieval