Trying out a more complex concept - experimental version. Inspired by kink's device bondage website photoshoots. This LoRA requires you to be precise in your prompting. Broken anatomy is expected, try different seeds and/or tweaking your prompts.

Strength of 1.0 works good, adjust if needed. Usually sweet spot is 1.0 to 1.2. Combines well with other LoRA's of mine to make it better understand bondage positions (such as hogtied, hands behind back etc.) and/or gags too.

Trigger phrase: "Woman in a device-bondage style scenario". You can use the following structure and headers too, this is how it was trained:

• Restraint Pose and Details

• Woman's Features

• Ancillary Details (gags, other persons, etc.)

• Environment & Lighting

• Style & Camera

Trained on natural language to support non traditional restraints like pipe, wood, straps, chains etc. It will produce watermark, simply type in "do not produce text watermark" at the end of the prompt to give clean images without it.

SEE EXAMPLES FOR PROMPTS. ALL IMAGES GENERATED ON DISTILLED WITH 6-8 STEPS.