ls-gan loss

[PRigas96]

hello i was searching through the code for the lsgan loss and found: if gan_mode == 'lsgan': self.loss = nn.MSELoss() if self.gan_mode in ['lsgan', 'vanilla']: target_tensor = self.get_target_tensor(prediction, target_is_real) loss = self.loss(prediction, target_tensor)

i would be pleased if you could explain how is this equal to:

for example, where are a,b and c valued.

[junyanz]

b=1, a=0, and c=1. We use MSELoss to implement L2 loss.