The model I think is fairly self explanatory.
Description
FAQ
Comments (7)
So sexy. What checkpoint are you using? Mine aren't coming out nearly as refined and clean as yours.
Autismmix Confetti, keep in mind I'm additionally upscaling by 2x from 768x1024, that also may play a factor
@Malebolgia Im using the same one. Ill try that upscaling. What upscaler are you using?
@Corgblam I think for most things I do these days I either use SwinIR_4x or 4xLSDIRplusR, For these I believe I used 4xLSDIRplusR which you can find at https://openmodeldb.info/models/4x-LSDIRplusR
@Corgblam If you don't have an 'ESRGAN' folder in your Stable diffusion model folder where your checkpoint and lora folders are at, make one and put it in there to get it to integrate with your stable diffusion. Assuming you're on some automatic111 fork like reforge or forge.
I also tested it a little and you can upscale at lower upscaling ratio's than 2x if your PC takes too long for comfort. I'd say as low as 1.25x could be prefectly fine with good quality with the hires function at a denoising strength of .2-.3
@Malebolgia I just used 4xUltrasharp_4xUltrasharpV10 and got some good results