Regression on vectors (list of predictions)

[Selda]

Hello,
Does DGL have a method for regression on vectors? I want to proceed with a GML workflow, which can predict a list of features/labels in node regression/link prediction.
I haven’t found any example yet. Thanks for your time!

[BarclayII]

DGL doesn’t have an example for node regression. However, adapting classification to regression merely involves changing the loss function from F.cross_entropy to loss functions for regression, such as F.mse_loss.

[Selda]

Thanks for your reply.
Do you have any suggestion for predicting a list of features generally? Any other method instead of DGL?

[czkkkkkk]

I think you just predict a feature vector and compare it with label vector by MES loss.