README.md

metadata

dataset_info:
  features:
    - name: image
      dtype: image
    - name: image_filename
      dtype: string
    - name: captions
      list: string
    - name: query
      dtype: string
  splits:
    - name: train
      num_bytes: 1117905881.107
      num_examples: 8091
  download_size: 1115990442
  dataset_size: 1117905881.107
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*