ReGraP
收藏资源简介:
ReGraP数据集是由中国科学院深圳先进技术研究院、西湖大学工程学部、明尼苏达大学双城分校、多伦多大学的研究团队创建的,包含120组个性化知识,每组包含图像、知识图谱和问答对。该数据集旨在促进个性化多模态大语言模型的学习和推理能力,数据集的构建过程涉及知识图谱的构建和问答对的生成。ReGraP数据集被用于训练ReGraP-LLaVA模型,并在ReGraP基准测试中评估其性能。
The ReGraP dataset was developed by research teams from the Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, the School of Engineering at Westlake University, the University of Minnesota Twin Cities, and the University of Toronto. It encompasses 120 sets of personalized knowledge, with each set containing images, knowledge graphs, and question-answer pairs. This dataset is designed to advance the learning and reasoning capabilities of personalized multimodal large language models. The construction process of the ReGraP dataset involves the development of knowledge graphs and the generation of question-answer pairs. The ReGraP dataset has been utilized to train the ReGraP-LLaVA model and evaluate its performance on the ReGraP benchmark.
ReGraP数据集概述
基本信息
- 数据集名称:ReGraP
- 托管地址:https://github.com/xyfyyds/ReGraP
数据描述
(注:根据提供的README内容,该数据集未包含具体描述信息)
使用说明
(注:根据提供的README内容,该数据集未包含使用说明信息)
其他信息
(注:根据提供的README内容,该数据集未提供其他相关信息)




