Can this framework be used for full supervision

[tanjia123456]

hello, I have found many framework is Semi-supervised or unsupervised？ For node classification, all my data has labels, so semi-supervised or unsupervised is not applicable. Do you know of a suitable framework for my data set?

[gyshgx868]

Sorry for my late reply.

If you want to apply this framework to fully supervised tasks, you can simply add a cross-entropy loss at the following line:

graph-ter/graph_ter_cls/runner/backbone_runner.py

Line 84 in 3b9bda5

loss = self.loss(t, t_hat) * self.args.lambda_mse

where you should also read the labels from the DataLoader.