Node classification - consider edge weights

Hi all,

I’d like to solve a node classification task. Right now, I don’t really have constraints on the model architecture (GAT / GCN / GraphSAGE would all work).

My only constraint is that I need to consider both node features and edge features. In my case edge features are actually scalars, so they’re actually edge weights.

I had a look at the code of the GAT, GCN and SAGE convolution classes:
GATConv / SAGEConv / GraphConv.

The edge weights/feature are never considered. Does anyone know how to address this task? Thanks.

Have you checked FAQ? Point 11 is related to your issue.

1 Like

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.