Usage Information:

• For nicer backgrounds for your victims, use with my Zelux for Flux set at 0.3 or 0.4

• Please use one of the FP16 LoRA settings in Diffusion in Low Bits

• Sampling steps: 30, otherwise it may forget whatever is supposed to be binding their wrists.

• Trigger: Doesn't need one, but you might make it extra reliable by using the training caption: "BDSMXPose, A woman is tied up. Her body is in an X pose. She is viewed from the front."

• LoRA strength: 0.7 or around there. Go much higher and you will struggle to get the backgrounds how you want them

Knowing bondage poses would be tough and probably create a lot of body horror, I decided to stick to a single pose, so that is all this does. If anyone is interested in more BDSM, let me know.

FACES

Custom faces on this one can only be achieved by inpainting. Inpainting is where you mark an area on the photograph and then tell the AI what to put there. It is a very powerful tool everyone should learn.

BODY

This should work with some breast mods. I know it works with mine. The only question is if the person who trained the mod left arms or legs or hands visible to any degree in the the training material. If that happens to be the case, it will cause problems.