SuperVee-02版本VeeXL-10 (ID: 389797)

I trained this model off of my photography profile from my years of being a photographer. It's leans more toward the model Vee that I have worked with for longer than any of the other women I've had the pleasure of photographing. It will produce NSFW results, as she is an Adult Model (as were the other models I worked with).

I used well over 500 HDR photographs from my portfolio, at 768 x 768. Initially I had focused on face only shots, but the results were too off-putting, so I let the AI crop the photos as it wanted. I only removed a few words from the auto-prompts (mole under eye, for example - running makeup seems to be translated as moles instead of makeup).

描述:

I ran the original dataset through SDXL training using PonyXL as a base model now that SDXL is easier to run with better VRAM management webui versions. I've been getting pretty decent results, still fine tuning my prompts, but all and all it's working out pretty well.

So far I've had good results using a variety of SDXL models that are based on PonyXL without much issue (aside from having to adjust face detail prompts on occasion). Autism-confetti is nice (and oddly relevant), as has been the original PonyXL. I've used Kim and Incursion too - I really don't have many qualms with the outcome even though it is all still a work in progress.

训练词语: