关于邻接矩阵的距离信息处理

[ReedGAOOO]

我在实际项目中需要用到真实距离的邻接矩阵，不是那种仅仅包含“0”或“1”的矩阵？这里的模型似乎都不能实现这一目标？大佬对此有什么好的建议？

[buffaloiron]

您好,你的邮件我已经收到,祝学业有成!工作顺利! --王晓栋I have recieved your e-mail,and I will reply you as soon as possible. Hope you have every successful on education and work! --xiaodong wang

[ReedGAOOO]

抱歉之前打扰了，试了一下输出SDCN中的邻接矩阵是带权重的，那应该可以解决了：
tensor(indices=tensor([[ 0, 0, 0, ..., 3024, 3024, 3024],
[ 0, 8, 20, ..., 2983, 2991, 3024]]),
values=tensor([0.0556, 0.0556, 0.0556, ..., 0.1111, 0.1111, 0.1111]),
device='cuda:0', size=(3025, 3025), nnz=29281, layout=torch.sparse_coo)

[buffaloiron]

您好,你的邮件我已经收到,祝学业有成!工作顺利! --王晓栋I have recieved your e-mail,and I will reply you as soon as possible. Hope you have every successful on education and work! --xiaodong wang

[yueliu1999]

真实距离的邻接矩阵也是可以使用的