I am working on generating LoRAs here on CivitAI, and these models are my first attempts. This LoRA features images generated with ComfyUI utilizing Everclear PNY. I think, on this attempt, I allowed the blank backgrounds which I used in creating the dataset to too heavily influence the LoRA, so heavy prompting in that regard is needed.

The subject is a fit, yet gawky, woman with an attempt at a Eurasian background. The key look was a consistent hairstyle, which I think I accomplished.