I'm new to deep learning, and I can't see the use of lnet. It seems that it just takes the output of the generator and the label and then output them. So why should we use it? How can loss_l help training? I'm quite confused, would you please tell me the reason or share your reference paper with me? Thanks a lot!
I'm new to deep learning, and I can't see the use of lnet. It seems that it just takes the output of the generator and the label and then output them. So why should we use it? How can loss_l help training? I'm quite confused, would you please tell me the reason or share your reference paper with me? Thanks a lot!