I get a black background #300

Open
opened 2026-01-29 21:46:37 +00:00 by claunia · 0 comments
Owner

Originally created by @joe-eis on GitHub (Jan 25, 2023).

GFPGAN is a great piece of work. Many thanks for it. Unfortunately my local installation doesn't work like the one from the online demo website.

I get a black background on the images in the folder restored_imgs.
I only get the faces, they look wonderful.

What could be the cause of this? I suspect that the background and the faces are not composited as on the demo websites.

The application runs under Xubuntu 22.04 LTS, Anaconda-Env., Python 3.8.16, basicsr 1.4.2, realesrgan 0.3.0, torch 1.13.1.

I downgraded to opencv-python==4.5.5.64, somewhere I read that and also installed the version from:
[(https://github.com/Erickesau/GFPGAN)]
This is the same version, but with GUI!

Unfortunately without success regarding the background. The faces and the background do not want to marry. I don't see a directory where the model of RearESRGAN could get into ...

Originally created by @joe-eis on GitHub (Jan 25, 2023). GFPGAN is a great piece of work. Many thanks for it. Unfortunately my local installation doesn't work like the one from the online demo website. I get a black background on the images in the folder restored_imgs. I only get the faces, they look wonderful. What could be the cause of this? I suspect that the background and the faces are not composited as on the demo websites. The application runs under Xubuntu 22.04 LTS, Anaconda-Env., Python 3.8.16, basicsr 1.4.2, realesrgan 0.3.0, torch 1.13.1. I downgraded to opencv-python==4.5.5.64, somewhere I read that and also installed the version from: [(https://github.com/Erickesau/GFPGAN)] This is the same version, but with GUI! Unfortunately without success regarding the background. The faces and the background do not want to marry. I don't see a directory where the model of RearESRGAN could get into ...
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: TencentARC/GFPGAN#300