>>2114
>>2111
Thanks, I'll do something along those lines next. I'm photoshopping some composition, don't wanna be too repetitive with the pussy removal

>>2121
Thanks, I saw your last gens, very clean and nice with the upscaling. Multi diffusers?

>Instead of a specific request for now I'd only like to ask more generally that you include pussy stabbings when and if it fits in your work, please.
I'll test that lora, haven't tried it yet but seems neat.

>I made the dataset for the pussystabbing Lora and I'll be working on expanding and improving the dataset further so that a v2.0 can be trained in the not to distant future. So any quality pussy stabbings would be of great help. So far I've got ordinary knives, medieval swords, sabres and katanas covered pretty well, but the dataset is especially lacking in rapiers (with either complex cage hilts or simple cuphilts). So just in case it fits in with your projects I'd love to see you give those a try. :)

Use anytest polished gens as expanded data? That is actually ingenious, you just blew my mind.
Although I wonder if odd unique swords such as rapiers wouldn't contaminate the generation process, being that guard construction is always so unique, I question whether it can be easily averaged. I don't know much about lora training but I know objects with a lot of shape variety are often prone to melting and odd shapes, such as guns: huge data sample, but still can't gen consistent firearms.