You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi !
I'm trying to re-set in a1111 the workflow I usually apply in ComfyUI : Applying Reactor and InstantID on segs.
Your extension is awesome !
But dealing with a1111 architecture i feel a little bit lost about how to "link" different extensions.
Do you think it could be possible via a script ?
In the case that isn't clear enough i precise i would like to :
Generate a picture in text2img
Get a segment mask of the face (let's talk about one only), let's say with face_yolov8m-seg_60.pt for exemple
Apply Reactor on this SEGS
Get a segment mask of the swap face, let's say with face_yolov8m-seg_60.pt for exemple
Apply InstantID on this SEGS
Get the final picture
My aim is to NOT have to do this things in a post-traitement (in img2img for example), but using the initial generated pic/latent/prompts/seed, etc
Big thanks by advance if you find so motivation to answer this question !
The text was updated successfully, but these errors were encountered:
Thanks for answer.
Yes, as you said, using the masks/segms of the detected results as an inputs of the Reactor/InstantID is exactly my aim, like i'm doing in ComfyUI.
Maybe this isn't really the aim of your extension, and if so i'm sorry to want to use it this way.
The fact is i'm a bit new to a1111, and this is in your extension that i better had the feeling of finding my bearings about segmenting.
Have you maybe some ideas about how i could achieve this aim (in a1111) ?
Hi !
I'm trying to re-set in a1111 the workflow I usually apply in ComfyUI : Applying Reactor and InstantID on segs.
Your extension is awesome !
But dealing with a1111 architecture i feel a little bit lost about how to "link" different extensions.
Do you think it could be possible via a script ?
In the case that isn't clear enough i precise i would like to :
My aim is to NOT have to do this things in a post-traitement (in img2img for example), but using the initial generated pic/latent/prompts/seed, etc
Big thanks by advance if you find so motivation to answer this question !
The text was updated successfully, but these errors were encountered: