Type | Workflows |
Stats | 359 |
Reviews | (11) |
Published | Apr 17, 2024 |
Base Model | |
Hash | AutoV2 C040C2F5C7 |
Update: Fairly new to ComfyUI and made some mistakes when wiring this up. When using AI gens for the face reference it worked fine, however as pointed out in the comments if I used a photo the results didn't really look like them.
So went back to the drawing board, simplified some parts and it seems to be resulting in a resemblance to the face reference photo now. I've also added in an IPAdapter for style transfer. This is optional and there's a Group Bypasser, so you can switch that off if you don't want it. One thing to note is that the IPA is more aggressive with the Face Detailer group enabled. You can use the Fast Group Bypass to have both on, off or one on and the other off, always good to have extra options to play with :)
If you have any feedback please pop it in the comments so I can improve this :)
Txt2Img generator with InstantID for the face swapper (InstantID), and a Face detailer then a 2 stage upscaler taken from a workflow called "1minute 8K Upscale". The first stage uses SUPIR and the second stage uses Ultimate SD.
The Upscaler group has it's own checkpoint loader. Using ZavyChromaXL for the Txt2Img and faceswapper/detailer groups, but when I connected the upscaler group I was getting a lot of noise in the final image and couldn't get rid of it. So the upscaler uses JuggernautXL Lightning and that seems to have fixed the noise issue.
There's also a separate text prompt box connected to the Face Detailer node if the features such as eye colour etc. are changing too much.
Max res I've tried so far is 6944 x 5376 and I haven't run into any VRAM errors and I'm using a 2080Ti which has 11GB VRAM and I have 32GB of System RAM, not sure how well it will run on a 8GB card but 11 or higher will be fine.