Datasets:

Tarun
/

AbsVis

Tasks:

Visual Question Answering

Zero-Shot Image Classification

Modalities:

Formats:

Languages:

Size:

Tags:

abstract-concepts

vision-language-models

human-annotations

Libraries:

License:

Dataset card Data Studio Files Files and versions

AbsVis / data

89 MB

Ctrl+K

Ctrl+K

1 contributor

History: 1 commit

Tarun's picture

Initial release of AbsVis (EMNLP 2025): images with per-image CC attribution, annotations, nouns

9fd3f1a verified 4 days ago