Datasets:

zzliang
/

GRIT

license: ms-pl
annotations_creators:
  - no-annotation
language:
  - en
multilinguality:
  - monolingual
pretty_name: GRIT
size_categories:
  - 100M<n<1B
source_datasets:
  - COYO-700M
tags:
  - image-text pairs
  - image-text-bounding-box pairs
task_categories:
  - text-to-image
  - image-to-text
  - zero-shot-classification
  - phrase grounding
  - referring expression comprehension
  - referring expression generation