TencentARC/GFPGAN
1
0
Fork 0
mirror of https://github.com/TencentARC/GFPGAN.git synced 2026-02-19 15:54:33 +00:00
Code Issues 476 Packages Projects Releases 8 Wiki Activity

Problem about finetuning GFPGAN v1.4 #503

New Issue
Open
opened 2026-01-29 21:48:20 +00:00 by claunia · 4 comments
claunia commented 2026-01-29 21:48:20 +00:00
Owner
Copy Link

Originally created by @leonsylarfang on GitHub (Aug 3, 2024).

Hello Xintao,

We found that the direct inferencing based on GFPGAN v1.4 performs pretty well on our own datasets, whilst GFPGAN v1 inferencing is not high-quality.

However, when we tried to finetune this model, we found that only the related files of training GFPGAN v1 was released.

Could you please share the .pth files, including discriminator, with us? So l could finetune the GFPGAN v1.4 with our custom data.

Thank you very much!

Best
Junheng

Originally created by @leonsylarfang on GitHub (Aug 3, 2024). Hello Xintao, We found that the direct inferencing based on GFPGAN v1.4 performs pretty well on our own datasets, whilst GFPGAN v1 inferencing is not high-quality. However, when we tried to finetune this model, we found that only the related files of training GFPGAN v1 was released. Could you please share the .pth files, including discriminator, with us? So l could finetune the GFPGAN v1.4 with our custom data. Thank you very much! Best Junheng
claunia commented 2026-01-29 21:48:20 +00:00
Author
Owner
Copy Link

@Afsalts commented on GitHub (Aug 4, 2024):

@leonsylarfang were u able to finetune

@Afsalts commented on GitHub (Aug 4, 2024): @leonsylarfang were u able to finetune
claunia commented 2026-01-29 21:48:20 +00:00
Author
Owner
Copy Link

@leonsylarfang commented on GitHub (Aug 5, 2024):

@leonsylarfang were u able to finetune
I have not got a reply yet... Finetuning GFPGAN v1 is ok but the data shape of the else cannot meet the provided states, as the value of channel_multiplier has been changed to 2.

@leonsylarfang commented on GitHub (Aug 5, 2024): > @leonsylarfang were u able to finetune I have not got a reply yet... Finetuning GFPGAN v1 is ok but the data shape of the else cannot meet the provided states, as the value of `channel_multiplier` has been changed to 2.
claunia commented 2026-01-29 21:48:21 +00:00
Author
Owner
Copy Link

@Afsalts commented on GitHub (Aug 9, 2024):

@leonsylarfang were u able to finetune
I have not got a reply yet... Finetuning GFPGAN v1 is ok but the data shape of the else cannot meet the provided states, as the value of channel_multiplier has been changed to 2.

is there any way i could talk to you . i have some doubts

@Afsalts commented on GitHub (Aug 9, 2024): > > @leonsylarfang were u able to finetune > > I have not got a reply yet... Finetuning GFPGAN v1 is ok but the data shape of the else cannot meet the provided states, as the value of `channel_multiplier` has been changed to 2. is there any way i could talk to you . i have some doubts
claunia commented 2026-01-29 21:48:21 +00:00
Author
Owner
Copy Link

@Tsai-chasel commented on GitHub (Aug 23, 2024):

能解决嘛？我看toRGB这层就有问题

@Tsai-chasel commented on GitHub (Aug 23, 2024): 能解决嘛？我看toRGB这层就有问题
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#503
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?