mair-lab/vismin-bench
收藏Hugging Face2025-01-22 更新2025-04-12 收录
下载链接：
https://hf-mirror.com/datasets/mair-lab/vismin-bench
下载链接
链接失效反馈官方服务：
资源简介：
VisMin数据集是一个用于评估模型在图像-标题配对的最小变化任务上的性能的基准数据集。它包含对象变化、属性变化、计数变化和空间关系变化四种类型的最小变化。总共有2084个样本，分别属于上述四种变化类型。每个样本包括两张图像和两个标题，以及四个相关问题。
The VisMin dataset is a benchmark 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 contains a total of 2,084 samples across these change types. Each sample includes two images, two captions, and four associated questions.
提供机构：
mair-lab
