DISAPPEARED13

norm[-1,1]for net='vgg'?

5

I found that you mention that we need to normalize the RGB images range into [-1, 1], but I saw that example worked at when ```net='alex'```, I wanna know that...

ImagePool with Pix2pix model

1

Hi, there! I noticed that Imagepool was used in cyclegan model, but not in pix2pix. As I known, Imagepool aims to stabilize the training progress for discriminators, and also strengthen...

About PatchGAN for CycleGAN

1

Hi, there. I notice that in the 3-layer discriminator definiation as PatchGAN, but I found that while training, just feed the **whole pictures** in D(), but PatchGAN will make a...

How to change num workers to make full use of GPU? (like Dataloader)

6

Hi, thanks for your code and it's really useful! I edit your source code for my task but I found that when running on GPU, I always got low Volatile...

I just use AttentionGAN-V1 combined with my GAN model, and my reg loss decrease to about 0,165 and then stay still, looking the `attention_mask` variable I found that after iterations...

Hi, there, I tried to adapt Strip Pooling as 3D and insert to my model, but after about 20-50 iterations in a epoch, I got nan loss. How you saw...

For 3D Networks?

3

Hello! I want to check my 3D CNN model's GPU using with torchstat, but it seems not support inputs with more than 3D... assertionerror occured.. any solutions to solve this...

Hi there, thanks for your work, it's really inspiring! I notice that in synthesis network, you use multiple MFB to mix feature maps, and in the MFB, there is a...

About two different loss while in training progress

1

Hi there, I download this code and adapted them for my semi-supervised segmentation (Pytorch version). And thanks for this genius code you provided! But I have a question is that...

About ablation study on alpha_0

4

Hi there, I notice that you ablate alpha_0 in Tab.7 of paper, showed that 20% is the best setting for your work. Firstly, the top-20% low confidence for negative samples...