Trained for use with H4N4 Semi-real, Versatile, V4. However, works on all pony models.

Alright, I've got a weakness for nerdy, geeky, or otherwise non-traditional women. They are my weakness. This Jane Doe is no different. Created from an amalgamation of source material, blended with img2img sampling to create a clean, robust data set. Some noise, some blur, re-injected to creat ambiguity, to allow your base model to fill in the small imperfections, and create an image that is full of soul, character, and life.

I'm tired of all the souless AI images that miss the distillation of life. This model comes with the slight imperfections that bring it to life.

I use an A-detailer and you should too. It allows a focused facial re-generation that allows faces with glasses to truly shine (and reflect). Also, because people argue about this everywhere, I use the clip skip AFTER the Lora loader.

Using at weights approaching 1 will cause facial distortion and blurring. Using below .6 will cause the model to begin to lose its unique imperfect style.

HAVE FUN! If you make something cool, please use the post button below, or at least show that you've used the model.