Replies: 1 comment 4 replies
-
"weight" only applies to code former. It's a setting only available in that model. As for that modulation setting, I'm not really sure what that does. I'll have to look into what they're doing there to know if that can currently be done with chaiNNer. |
Beta Was this translation helpful? Give feedback.
4 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
I just discovered chaiNNer, in the search for the best workflow for in particular facial/portrait upscale of very low quality source pictures.
Am absolutely BLOWN AWAY by a model like "4xFaceUpDAT.pth". At times, a combination with GFPGan1.4 significantly improves the endresult even further.
I'm coming from AUTOMATIC1111 for Stable Diffusion, using the "extras" tab there to apply some of the upscalers. DAT models are not available yet, there. However, a feature that I DO use over there, is the ability to "modulate" the effect of GFPGan by tuning the "visibility" to something else than "completely ON" or"completely OFF". Most of the time, I find that setting "GFPGan visibility" to 0.75 or so, finds a better balance between "restoration" and "faithfulness".
Does chaiNNer offer the ability to do something similar?
The documentation of "Upscale Face" talks about "weight", but I don't have that, only "Output scale" (?)
Beta Was this translation helpful? Give feedback.
All reactions