taotaonice / FaceShifter

Try to reproduce FaceShifter
250 stars 62 forks source link

Line 110 in train_AEI.py #20

Closed ocastan closed 4 years ago

ocastan commented 4 years ago

Hello

Can you explain why you did not use the mean for L_adv for calculating G loss but use it for loss_true and loss_fake when calculating D loss ?

Thanks in advance

taotaonice commented 4 years ago

Not much difference