Tools FlashFace - ComfyUI

chipmoss7

Lurker
Joined
Apr 15, 2023
Posts
16
Reaction score
0
https://github.com/cold-hand/ComfyUI-FlashFace
https://github.com/ali-vilab/FlashFace
With only a few pictures you can generate high quality images of a person.
Example raw outputs:

Unfortunately the inpainting model was not released by the authors of the paper.
What could be done is trying to rewrite the denoising process to condition it on a latent image.
This would unlock an img2img workflow and being able to use it for faceswaps.
But this requires some coding skill and AI diffusion knowledge.
 
Does anyone have a quick guide as to how to actually run something from github that looks like this? Can't find anything that's suitable on youtube..
 
ComfyUI is confusing af, StableDiffusion feels way more natural to me lol
 
Anyone else having an issue with the nodes? None of them show up in my node list after the install.
 
How much gb of ram do u need? I have a rtx 4070ti super but my i only have 16 gb of memory ram and everytime i ran the pokimane flux model it crashes hard
 
You shouldn't have a problem with those specs running flux, I am able to create my own images using the exact same graphics card and vram you have.



You could be using the full model that requires 24 gb
Try using Flux FP8 model instead.
I followed this link when installing flux https://stable-diffusion-art.com/flux-comfyui/
 
It seems like the installation goes wrong at my side, it recognizes the folder FlashFace in the custom_nodes folder but it fails to import. Anyone else?
 
impressive stuff, unfortunately i don't have the required setup to run flux, wouldn't mind seeing more of your work tho, are you sharing them somewhere or just individual topics?
 
Back
Top