README.md

metadata

dataset_info:
  features:
    - name: id
      dtype: int64
    - name: objects
      sequence: string
    - name: colors
      sequence: string
    - name: objects_included
      dtype: bool
    - name: file_name
      dtype: string
    - name: caption
      dtype: string
    - name: caption_source
      dtype: string
    - name: image
      dtype: image
  splits:
    - name: train
      num_bytes: 3609121093.4
      num_examples: 36200
  download_size: 3696158170
  dataset_size: 3609121093.4
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
license: cc-by-nc-4.0
task_categories:
  - text-to-image
  - image-to-text
language:
  - en
tags:
  - synthetic