VisionGraph
收藏资源简介:
VisionGraph是一个用于探索高级LMMs解决多模态图论问题的基准,包含从连通性到最短路径问题的八个图问题任务。通过引入描述-程序-推理(DPR)链,增强逻辑推理过程的准确性。
VisionGraph is a benchmark designed for exploring advanced LMMs (Large Multimodal Models) in solving multimodal graph theory problems, encompassing eight graph problem tasks ranging from connectivity to shortest path issues. By introducing the Description-Program-Reasoning (DPR) chain, it enhances the accuracy of the logical reasoning process.
数据集概述
名称: VisionGraph
目的: 探索高级大型多模态模型(LMMs)在解决视觉上下文中的图论问题方面的能力。
包含内容:
- 八个图问题任务,涵盖从连接性到最短路径问题。
- 描述-程序-推理(DPR)链,用于通过图形结构描述生成和算法感知的多步推理增强逻辑准确性。
数据集访问:
相关资源:
- 提示、数据集、检查点及评估方法均可在该仓库中获取。
论文引用:
@article{li2024visiongraph, title={VisionGraph: Leveraging Large Multimodal Models for Graph Theory Problems in Visual Context}, author={Yunxin Li and Baotian Hu and Haoyuan Shi and Wei Wang and Longyue Wang and Min Zhang}, journal={arXiv preprint arXiv:2405.04950}, year={2024}, }




