YONG-XIANG LIN

@JiahuiYu Hi, thank your code and paper I have a question about the DeepFillv2 discriminator, Are all layer in discriminator apply like below `x = leaky_relu(spectral_norm(x))` Including the first and...

+1, Using light mode gets similar results.

@Serge-weihao I use the same way as GTA5-to-cityscapes to train. It means that I use 19 classes to train. I try to use 13 classes to train, it will get...

@Serge-weihao I will confirm it when I have time in recent days. I remember that I tried to use the GT dataset for a long time, but failed, so I...

@Serge-weihao I use the same dataset as your link, SYNTHIA_RAND_CITYSCAPES, about 20GB. I found that I use GT/LABELS dir instead of GT/COLOR. You can take a look below code. ...

@ETHRuiGong Nope, I haven't tried it yet.

Thanks, @Kppeppas explained it. I do some experiment to explain more detail about the patchnce loss In the intuition, we think the predictions shape is (B*S, 1(pos)+S(neg))， so we assume...