pygcn
pygcn copied to clipboard
specifying modes train, validation and test
Hi guys!
I hope any of you happen to know why in line 88 of train.py, while it tries to specifies test mode, we see these line:
def test():
model.eval()
Dont you think it should be ### model.test() ### sine we are focusing on test set?
Thank you in advance guys.
@fansariadeh There's no model.test()
in pytorch.
When you train the model, call model.train()
will activate some layers like batch-norm, zero-out.
When you test the model, call model.eval()
will modify the operations through these layers.