A LoRA that gives the model a better understanding of male chastity cages.

I've found a weight of 0.8 to work well.

Trained on the BigLove Photo4 Checkpoint with a 100 image dataset. May work on other checkpoints as well.

It seems to remain fairly flexible, but I have not tested all edge cases. Your comments are welcome so I can see where the weak points are.

Example Prompts:
xcage, metal, 1boy, sitting

xcage, pink, 1boy, sitting