LoRA trained on publicly available images. Please be responsible with this model.

Might take a bit of coaxing, but in my opinion it's possible to get great results, weighting anywhere from 0.8-1.0.

Happy to get feedback on anything, leave a comment, this is my first LoRA ever.

E: After running it for a while I've noticed that every now and then it can put outputs in a jacket / coat without being prompted. Future training sets might address this. For now, if you negative prompt for (jacket, blazer, coat) it should fix it.