Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Small noisy patches in output #339

Open
luser350 opened this issue May 11, 2024 · 5 comments
Open

Small noisy patches in output #339

luser350 opened this issue May 11, 2024 · 5 comments

Comments

@luser350
Copy link

luser350 commented May 11, 2024

Hi, I am training a face-retouching model using FFHQ and FFHQR datasets. Since I am training on 1024x1024 my batch size is 2. Here is the sample output
epoch008_synthesized_image

As you can see there is a small noisy patch on the forehead of the image. This noisy patch consistently shows up on every generated image.
How can this problem be rectified @aaronp24 @lukeyeager @javl @aflat ?
opt.txt

@luser350
Copy link
Author

3 weeks and still no answer.......??????????

@zhangxiaojuan66
Copy link

zhangxiaojuan66 commented May 30, 2024 via email

@luser350
Copy link
Author

luser350 commented May 30, 2024

Hello, I have received your email. I wish you all the best and a happy life!

Thanks for your reply, what might be wrong with the training? I have tried a higher batch size (6) but still I am getting these noisy patches.

@javl
Copy link
Contributor

javl commented May 30, 2024

@luser350 I haven't used pix2pix in a long time so I can't help you.

Keep in mind: you're using a free and open source tool, nobody is obligated to help and comments filled with question marks or sneers toward other users aren't going to convince anyone to spend time on this issue.
Maybe start by explaining what you've already tried yourself?

@luser350
Copy link
Author

luser350 commented May 30, 2024

@javl thanks for your reply
Firstly I felt mocked by @zhangxiaojuan66 reply. I have updated my previous comment and answered your question

what you've already tried yourself?

Here is my loss_log.txt

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants