Spatial-MM
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/FatemehShiri/Spatial-MM
下载链接
链接失效反馈官方服务：
资源简介：
该数据集是一个新颖的视觉问答（VQA）数据集，旨在研究大型多模态模型在空间理解和推理能力方面的表现。它包含了针对对象关系和多跳推理的分析，其研究发现强调了大型多模态模型在空间推理方面所面临的挑战。
This dataset is a novel visual question answering (VQA) dataset designed to investigate the performance of large multimodal models in terms of their spatial understanding and reasoning capabilities. It includes analyses targeting object relationships and multi-hop reasoning, and its research findings highlight the challenges that large multimodal models encounter in spatial reasoning.
