README.md

metadata

dataset_info:
  features:
    - name: image
      dtype: image
    - name: text
      dtype: string
  splits:
    - name: train
      num_bytes: 495025255.1905698
      num_examples: 916
    - name: test
      num_bytes: 57052510.80943026
      num_examples: 102
  download_size: 547398664
  dataset_size: 552077766.0000001
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: test
        path: data/test-*
task_categories:
  - image-to-text
language:
  - ja