README.md

metadata

dataset_info:
  features:
    - name: image
      dtype: image
    - name: filename
      dtype: string
    - name: caption
      sequence: string
  splits:
    - name: train
      num_bytes: 829880166
      num_examples: 6000
    - name: test
      num_bytes: 136929806
      num_examples: 1000
    - name: validation
      num_bytes: 138076299
      num_examples: 1000
  download_size: 1102895899
  dataset_size: 1104886271
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: test
        path: data/test-*
      - split: validation
        path: data/validation-*