Something I like, was trying to do a "all in one" model with the style of BerryMix. Mix of 65% Real Berry (F222, NovelAI, Anything V3, R34) and 35% Smirking+BStaber (Smirking Face 50% / 50% BStaber). Do all I want, support sfw and nsfw. Support anime art, realistic art, do close up with detailed background. Best use : 20 step, 15 cfg, highres.fix (512x... , ...x512 , 512x512), Euler_A sampling work best for anime, when DPM++ 2M Karras sampling work best for realistic, after some testing. React to prompt keyword like : anime style, photorealistic, anime screencap, realistic Picture was done with basic prompt, I didn't saved them. Take in consideration i'm not a SD wizard, I don't really know what I'm doing, I was just trying merging models to create one that will fit my need, and wanted to share it with others. I hope you will like it anyway! Style art test : https://imgur.com/a/1oFfBiK Fellatio Embedding for this model : https://mega.nz/file/SQZAEISK#LMLxXnIwYd84DCFDsbZoyI9l9xjAciRW5hxqkUCyGBw (Put the .bin in ..\diffusion\stable-diffusion-webui\embeddings then use the prompt "fellatio") Edit 23/12 : Modified pictures showing, with prompts and all. Edit 24/12 : Added some pictures with christmas prompt. Merry christmas! |