Standard Loss

• Source: /losses/standard_loss.py

criterion = torch.nn.BCEWithLogitsLoss()
g_loss = criterion(d_fake, torch.ones_like(d_fake))
d_loss = criterion(d_real, torch.ones_like(d_real)) + criterion(d_fake, torch.zeros_like(d_fake))

examples

• Configurations: /losses/standard_loss/

{
  "class": "class:hypergan.losses.standard_loss.StandardLoss"
}