Datasets:

google
/

docci

annotations_creators:
  - expert-generated
  - crowdsourced
language:
  - en
language_creators:
  - other
license:
  - cc-by-4.0
multilinguality:
  - monolingual
pretty_name: DOCCI
size_categories:
  - 10K<n<100K
source_datasets:
  - original
tags: []
task_categories:
  - text-to-image
  - image-to-text
task_ids:
  - image-captioning