Brianna one of Andava's OCs
You can check out his work here.
_____________
I tried to get the "braces" to work reliably, but failed despite all my efforts. I supplemented the 10- and 12-epoch dataset with new images of Brianna, and in the 10-epoch version, I also supplemented the dataset with images focusing on the brackets (which I specifically generated using Gemeni). However, the effect was too insignificant. While the 10-epoch version might yield some results, it's not worth it. Therefore, I excluded these images from the 12-epoch version. As the saying goes, "Better a fowl in hand than two flying."
Otherwise, both models are viable. I can't judge which is better, but if I had to, I'd prefer the 12-epoch version; I liked its color rendition.
Description
Brianna one of Andava's OCs
You can check out his work here.
Trigger: brianna, andava_style,
Secondry trigger: teen girl, blue eyes, eyeshadows, blonde hair, medium breasts, flexible girl
Best weight: 0.5 ~ 0.7
Trined on Ponydiffusion v6
P.S. I never managed to make satisfactory bracket generations, so I don't recommend using this tag.
P.P.S. Instead of “andava_style”, you can use other styles, works stable.