mair-lab/vismin-bench
VisMin Dataset Overview The VisMin dataset is designed for evaluating models on minimal-change tasks involving image-caption pairs. It consists of four types of minimal changes: object, attribute, count, and spatial relation. The dataset is used to benchmark models' ability to predict the correct image-caption match given two images and one caption, or two captions and one image. Dataset Structure Total Samples: 2,084 Object Changes: 579 samples… See the full description on the dataset page: https://huggingface.co/datasets/mair-lab/vismin-bench.
| Author / organization | mair-lab |
|---|---|
| License (as declared) | cc-by-4.0 |
| Downloads | 723 (as reported by Hugging Face) |
| Likes | 0 |
| Created | 21 November 2024 |
| Last updated at source | 22 January 2025 |
Popularity, last 3 days
723 now (+112 since 2026-10-16). Counts as reported by the source, snapshotted daily by Silytics.
Recent changes
- 17 Oct 2026 Description: VisMin Dataset Overview The VisMin dataset is designed for evaluating models on minimal-change tasks involving image-caption pairs. It consists of four types of minimal changes: object, attribute, count, and spatial relation. The dataset is → VisMin Dataset Overview The VisMin dataset is designed for evaluating models on minimal-change tasks involving image-caption pairs. It consists of four types of minimal changes: object, attribute, count, and spatial relation. The dataset is
Changes detected by Silytics between refreshes of the source.
More from mair-lab
- mair-lab/visminDataset
- mair-lab/CulturalFramesDataset
- mair-lab/CulturalVQADataset
- mair-lab/escher-kitchen-actionDataset
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 18 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.