DeblurGANv2

DeblurGANv2 copied to clipboard

I am a total Python beginner and want to run predict.py in anaconda/spider environment on Windows10. I have downloaded with firefox from "https://drive.google.com/uc?export=view&id=1UXcsRVW-6KF23_TNzxw-xC0SzaMfXOaR" and renamed it to inceptionresnetv2-520b38e4.pth, as indicated...

erichnewie

Hello, Thank you for your work. I would like to ask a naive question that what is the ideal value about the GAN loss and Discriminator loss when the model...

Awenjie10

Trained D model

1

Hi, have you save net_Ds' ckpts. Like full D and patch D.

TheLostIn

Perceptual Loss computation

9

Hi, thanks for sharing your code. I was reading the implementation of the PerceptualLoss and I am wondering why only the first image of the batch is normalized: https://github.com/VITA-Group/DeblurGANv2/blob/c5bbf780b152d5b877d9911b8323898d5f2e0a50/models/losses.py#L50-L51 Is...

rogertrullo

Please release the Discriminator model!

2

Please release the Discriminator model! Many Thanks!

ghost


zt-Github-1

Training dataset for pre-trained models.

4

Hi! First of all, congratulations for your work. I have a question regarding the pretrained weights. In the subsection is said "Dataset" gopro but in the paper it says: "We...

pvitoria

Can't convert to torchscript correctly

2

I just changed scale_factor and the torchscript is work, but the result on python and c++ are different. Do you tried that. Can you help with that?

SolicTous

I replace the backbone to another lightweight architecture, but when I train the model, the SSIM is less than 0.8 and the PSNR is less than 27. Looking for help!!!

oywstju

Unstable Content_Loss for validation

2

During the training process, I observed that my Val_G_Loss curve was quite abnormal using RAGAN-LS due to the unstable Val_Content_Loss. All metrics appear normal except that the curve of Validation_Content_Loss...

NoviceCrom