标签:vit

中科院&华为Vision GNN:只使用图神经网络进行视觉任务

目前广泛应用的卷积神经网络和卷积神经转换器将图像视为网格或序列结构。本文提出将图像表示为图(Graph)结构,ViG)体系结构来提取视觉任务的图级(Graph-L...

NLPCC 2023丨Final Call for Papers

https,systems,standards;applications,◇ Applications/tools:◇ Theoretical:Papers currently under review in other conferences or journals are acceptable;