SchafferZhang

The loss with regard to the unlabeled data `l_unl` is defined as: `loss_unl = -0.5*T.mean(l_unl) + 0.5*T.mean(T.nnet.softplus(l_unl)) + 0.5*T.mean(T.nnet.softplus(l_gen))` But the first term contradicts with the second term, which obsess...

As I know, if the model is trained well, it could generate some realistic images to some extent.But the result is so blurry as not be able to recognize its...