README.md

metadata

language:
  - ja
task_categories:
  - image-to-text
dataset_info:
  features:
    - name: image
      dtype: image
    - name: text
      dtype: string
  splits:
    - name: train
      num_bytes: 487524806.0863096
      num_examples: 907
    - name: test
      num_bytes: 55790355.913690485
      num_examples: 101
  download_size: 541073440
  dataset_size: 543315162.0000001
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: test
        path: data/test-*