It attempts to do a 3D wireframe of the prompted asset. Use obvious trigger words. Highly WIP- but can produce nice results and works well with other loras.

If you find this is overfit- try reducing the strength to 0.7/0.75 and blending it with another lora.

It was purposely trained to be overfit, as this is a tough style to push.

If you support me, you can fuel my caffeine here: https://ko-fi.com/syntaxdiffusion/commissions

cheers.