The edge_index of one node

[aa-oo]

Hello, I would like to ask, if there is only one node in my graph and I want to use gcn, how to define its edge_index? That's edge_index=torch.tensor([0],[0]],dtype=torch.long)? Thank you

[LukeLIN-web]

It is strange; how do you aggregate information from neighbors?

[aa-oo]

It doesn't have to consider its neighbors, in other words, the whole graph is equivalent to just one node, but I need to use gcn to convolve

[LukeLIN-web]

You can add self loop as edge_index=torch.tensor([0],[0]],dtype=torch.long)

[aa-oo]

ok,thank you!

…

------------------ 原始邮件 ------------------ 发件人: ***@***.***>; 发送时间: 2023年2月27日(星期一) 中午1:29 收件人: ***@***.***>; 抄送: ***@***.***>; ***@***.***>; 主题: Re: [pyg-team/pytorch_geometric] The edge_index of one node (Discussion #6798) You can add self loop as edge_index=torch.tensor([0],[0]],dtype=torch.long) — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: ***@***.***>