vscodeformer

[Selur]

Did some small tests with https://github.com/HolyWu/vs-codeformer.
https://imgsli.com/MTY2MTM5
Suzie: https://imgsli.com/MTY2MTQz
Suzie with weight=1: https://imgsli.com/MTY2MjEw
https://imgsli.com/MTY2MTQ0
https://imgsli.com/MTY2MTQ2
https://imgsli.com/MTY2MTQ3
https://imgsli.com/MTY2MTQ5
https://imgsli.com/MTY2MTUy
seems more like a gimmick filter than really usable atm., but its impressive.
=> not planning to include it in the torch package, atm.

Cu Selur

[Selur]

Did some further testing and if upscale is > 1 and the upscaling is later undone it basically can be used as a 'face detail enhancer' which might be interesting.
Also in combination with FeMaSR it can be used to fix the face distortions FeMaSR does.

=> I'll add it to the torch-addon.
SwinIRx4+vsCodeFormer: https://imgsli.com/MTY2NDA3, looks nice frame-by-frame, but isn't really temporally stable => adding 3xQTGMC_type3 seems more stable and all in all a nice result in my book.

Cu Selur

Ps.: got a nice example why, weight=0 usually shouldn't be used: https://imgsli.com/MTY2NDI2