wrong padding of AdaptiveFeatureGenerator

[htzheng]

Hi, I noticed that for AdaptiveFeatureGenerator, the encoder is implemented using padding of Conv,
self.layer1 = norm_layer(nn.Conv2d(opt.spade_ic, ndf, kw, stride=1, padding=pw))
However, the warping images have strong artifacts at boundry.

Instead, when replacing the padding with reflect padding, the result is significantly improved.

Because of this, it seems there is a difference between the pretrained model and the testing code. Also, when using reflect padding for training, I often find the training collapses with unknow error after several epochs. Did you notice the same issue?

[panzhang0212]

Could you show the error log?

[htzheng]

Hi, in the AdaptiveFeatureGenerator, I found feature backbone is extracted without reflect padding,
self.layer1 = norm_layer(nn.Conv2d(opt.spade_ic, ndf, kw, stride=1, padding=pw))
Because of this, the warped images have surrounding artifacts.

I wonder is there any ways to solve this artifact problem. Thanks.