@ashleyk Hi Ashley I tried your template 3.0.3 today (RTX A4000) and the "Reactor - Fast Face Swap"
@ashleyk Hi Ashley I tried your template 3.0.3 today (RTX A4000) and the "Reactor - Fast Face Swap" node doesn't work (attached image). However, another pod using template 3.0.1 (RTX 3090) is working fine with that node. I installed the node from ComfyUI Manager for both pods. Could you please tell me how to fix it? Thank you very much. (Also how do you usually do faceswap in ComfyUI?)
https://github.com/Gourieff/comfyui-reactor-node
GitHub
GitHub - Gourieff/comfyui-reactor-node: Fast and Simple Face Swap E...
Fast and Simple Face Swap Extension Node for ComfyUI - GitHub - Gourieff/comfyui-reactor-node: Fast and Simple Face Swap Extension Node for ComfyUI
13 Replies
Probably a bug in ComfyUI, log a bug on their Github
Its either a bug in ComfyUI or the node, nothing I can do about it
I wish AI developers were not so damn incompetent always breaking stuff
Ok thank you. Luckily I still have the old template
very good
gonna test on my sdxl dreambooth
Thanks! That means a lot coming from you 🙂 It took a long time and a lot of fiddling. I only managed to do 5 epochs for the DreamBooth (it took 24 hours to do 5!).
I would LOVE to see that!
wow
it must had got VRAM bottleneck
Oh, yes! I trained 768x768 on a 8GB RTX 3070.
You can add to the Github issue for this incompetent developer to fix it:
https://github.com/Gourieff/comfyui-reactor-node/issues/41
GitHub
Exception during processing - windows version (not portable) · Issu...
Dear developers, tinkerers and tech people Yesterday I was messing around with faceFusion, and im almost sure I broke something regarding to opencv or onnxruntime Not using venv :-( opencv-python==...
Ok thank you again
@Z3ph0d your prompt really complex 😄
I know! I'm struggling a bit. I understand the model and how the tokens affect the diffusion process (the basics at least), but finding the right tokens is really hard.
Wow! This are amazing!