Cycle gan g_loss and d_loss become nan after some epochs during training

[Ayesha-Rafiq]

Generator loss and Discriminator loss become nan after some epochs and cycle gan start to generate black images.

[Auth0rM0rgan]

Check your optimizer section, It may be related to set something wrong in optimizer.

[hala3]

i want to change the loss function by pixel wise loss , how i can do that please

[nagaswethar]

self.g_loss_a2b = self.criterionGAN(self.DB_fake, tf.ones_like(self.DB_fake))
+ self.L1_lambda * abs_criterion(self.real_A, self.fake_A_)
+ self.L1_lambda * abs_criterion(self.real_B, self.fake_B_)
+ self.Lg_lambda * gradloss_criterion(self.real_A, self.fake_B, self.weighted_seg_A)
+ self.Lg_lambda * gradloss_criterion(self.real_B, self.fake_A, self.weighted_seg_B) self.g_loss_b2a = self.criterionGAN(self.DA_fake, tf.ones_like(self.DA_fake))
+ self.L1_lambda * abs_criterion(self.real_A, self.fake_A_)
+ self.L1_lambda * abs_criterion(self.real_B, self.fake_B_)
+ self.Lg_lambda * gradloss_criterion(self.real_A, self.fake_B, self.weighted_seg_A)
+ self.Lg_lambda * gradloss_criterion(self.real_B, self.fake_A, self.weighted_seg_B)

can any one explain this loss function and what \ indicates in code