Node regression using a dataset of graphs

[cap-jmk-real]

In contrast to the node classification example (https://colab.research.google.com/drive/14OvFnAXggxB8vM4e8vSURUp1TaKnovzX?usp=sharing), I wonder if it is possible to perform node regression given a dataset where I have a graph in which each node has multiple features and I try to predict multiple classes via regression. Has something like this been done before in PyTorch Geometric?

I guess it would play out similarly to this question: #3794

Curious about your thoughts! :)