Training on custom resolution 128x128? #127

Open
opened 2026-01-29 21:43:10 +00:00 by claunia · 2 comments
Owner

Originally created by @waleedrazakhan92 on GitHub (Dec 18, 2021).

Hi, I need to train the model with custom resolution 128x128. What changes do need to make in the code?
What I understand is I need to have a pretrained stylegan2 model with 128x128 shape. And also the UNET degradation model also needs to be of shape 128x128. I think that can be controlled in the ".yml" file in the options.
What else do I have to change for training such a model. Any instructions would be helpful,

Originally created by @waleedrazakhan92 on GitHub (Dec 18, 2021). Hi, I need to train the model with custom resolution 128x128. What changes do need to make in the code? What I understand is I need to have a pretrained stylegan2 model with 128x128 shape. And also the UNET degradation model also needs to be of shape 128x128. I think that can be controlled in the ".yml" file in the options. What else do I have to change for training such a model. Any instructions would be helpful,
Author
Owner

@kaikaizhu commented on GitHub (Dec 5, 2023):

I also want to use a custom resolution of 128x128 to train the model. Did you succeed? Does anyone know what to do?

@kaikaizhu commented on GitHub (Dec 5, 2023): I also want to use a custom resolution of 128x128 to train the model. Did you succeed? Does anyone know what to do?
Author
Owner

@waleedrazakhan92 commented on GitHub (Dec 7, 2023):

I also want to use a custom resolution of 128x128 to train the model. Did you succeed? Does anyone know what to do?

No i did not.

@waleedrazakhan92 commented on GitHub (Dec 7, 2023): > I also want to use a custom resolution of 128x128 to train the model. Did you succeed? Does anyone know what to do? No i did not.
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: TencentARC/GFPGAN#127