How can I judge the model is train or validation in the ‘forward’ function?
In fact, I have construct a network, and write the ‘forward’ function. I want to judge whether this forward process happened in the training process or in the validation process.
Hi @linyu,
You can know by checking the value of the nn.Module
attribute self.training
in forward
(set to True
in net.train()
and to False
in net.eval()
) .