TencentARC/GFPGAN
1
0
Fork 0
mirror of https://github.com/TencentARC/GFPGAN.git synced 2026-07-09 02:06:30 +00:00
Code Issues 476 Packages Projects Releases 8 Wiki Activity

Trying to fine-tune gfpgan #521

New Issue
Open
opened 2026-01-29 21:48:30 +00:00 by claunia · 1 comment
claunia commented 2026-01-29 21:48:30 +00:00
Owner
Copy Link

Originally created by @NetPranav on GitHub (Feb 14, 2025).

I wanted to increase the accuracy and train gfpgan on a particular lighting dataset to increase the details and lighting scenario of image precision
But I am new to this and got stuck as again and again the error occurs that Gfpganer has no parameter I tried to do the same with calling various like generator and discriminator but it just didn't work can any one tell me like how to actually FINETUNE a GFPGAN with new dataset

Originally created by @NetPranav on GitHub (Feb 14, 2025). I wanted to increase the accuracy and train gfpgan on a particular lighting dataset to increase the details and lighting scenario of image precision But I am new to this and got stuck as again and again the error occurs that Gfpganer has no parameter I tried to do the same with calling various like generator and discriminator but it just didn't work can any one tell me like how to actually FINETUNE a GFPGAN with new dataset
claunia commented 2026-01-29 21:48:30 +00:00
Author
Owner
Copy Link

@shubhamghimire commented on GitHub (Mar 5, 2025):

You can not finetune the pth file of Ver1.3 and Ver1.4 of GFPGAN. Because these pth files are converted from another pth files using the following script.
https://github.com/TencentARC/GFPGAN/blob/master/scripts/convert_gfpganv_to_clean.py

So you need to get the original pth file to finetune GFPGAN.
As far as I know the original pth files for Ver.1.3 and Ver.1.4 are not provided. But there are pth files for Ver.1.1 which you can finetune. You can download the generator weights and discriminator weights from below.

https://drive.google.com/drive/folders/17rLiFzcUMoQuhLnptDsKolegHWwJOnHu

@shubhamghimire commented on GitHub (Mar 5, 2025): You can not finetune the pth file of Ver1.3 and Ver1.4 of GFPGAN. Because these pth files are converted from another pth files using the following script. https://github.com/TencentARC/GFPGAN/blob/master/scripts/convert_gfpganv_to_clean.py So you need to get the original pth file to finetune GFPGAN. As far as I know the original pth files for Ver.1.3 and Ver.1.4 are not provided. But there are pth files for Ver.1.1 which you can finetune. You can download the generator weights and discriminator weights from below. https://drive.google.com/drive/folders/17rLiFzcUMoQuhLnptDsKolegHWwJOnHu
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
pull-request
Mirrored from GitHub Pull Request
No Label
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
claunia
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: TencentARC/GFPGAN#521
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?