context_encoder_pytorch parameter wtlD for Adversarial Discriminator loss is not used

some parameter is set as follow: parser.add_argument('--wtl2',type=float,default=0.998,help='0 means do not use else use with this weight') parser.add_argument('--wtlD',type=float,default=0.001,help='0 means do not use else use with this weight') but i find wtLD is nerver used, so what is it for?

Dec 23 '17 03:12 micklexqg

you can use wtlD instead of (1-wtl2) in this line

Dec 23 '17 03:12 BoyuanJiang

@BoyuanJiang , thanks for so quick reply! wtlD is 0.001 or 0.002?

Dec 23 '17 03:12 micklexqg

another question, I know the errG_l2 is for reconstruction loss, but what is the err_D and errG_D for respectively?

Dec 23 '17 03:12 micklexqg

err_D and errG_D are the loss of GAN, and sorry for first question,according to this paper,wtLD isn't used.

Dec 23 '17 03:12 BoyuanJiang

@BoyuanJiang , GAN? there are total two kind loss, reconstruction (or generation) loss and adversarial discriminator loss, but there three notations used. so which one is adversarial discriminator loss, the err_D or errG_D?

Dec 23 '17 05:12 micklexqg