I am looking to solve a toy problem. I have a numbers of graphs where each vertex has a few attributes e.g. (x, y, z) and I'm looking to train a model that can assign a label ('horizontal', 'diagonal', 'flat') to each edge in the graph.
Can this be done with pytorch geometric and is a code sample somewhere?
It seems possible in DLG: https://docs.dgl.ai/en/0.6.x/guide/training-edge.html not sure about Pytorch geometric though.