Creates male selfie/mirror selfie images. Trained on 19 images as an experimental version.

I highly recommend using "mirror selfie" in your prompt, otherwise it does not have the full effect. "selfie" does not work entirely as intended.

An issue I ran into while testing was it would try and give the characters airpods a lot, due to semi-overrepresentation in the training data (my fault).

Refer to display images for keywords. I use very very simple prompts, but I believe detailed prompts should work as intended.

Rigid training, so mixing LoRA's may not produce great results.

*Any images generated with this LoRA do not depict real persons. Any resemblance is purely coincidental.