Lora in the style of Doug Sneyd. Kinda vintage and sketchy style. Unique faces. Looks best in Anima Base. For closest results donz use tages like "score_9, score_8_up, score_7_up, high quality", just "normal quality" is enough. Styles like that are imperfect. and those tags make everything smooth and nice. But thats not the goal here.
Added v2: Had to edit 5 images, cuz of signature. But u only notice it when using Turbo versions of Anima. Cuz at cfg 1 it has no Neg prompt.
Anima :
v1 and v2:
Trained on Anima Base 1.0
Weight 1.2 - 1
Trigger : dougsneyd
Would be awesome if u leave a like and post a pic ;)
Description
FAQ
Comments (4)
Looks like you really caught the original visual style! I'm always happy to see attention going to Western visual styles, offering variation from the massive dominance of anime in AI image generation. =)
If you're looking for more examples, check out Erich Sokol. (Dean Yeagle will always be my fave, tho. =D)
hey, glad u like it! i ve made Dean Yeagle for pony and illu, but it was already made for anima, so i skipped it. but i like it very much too, always used it for example images. i will check out erich sokol!
@BloodySunkist Your Dean Yeagle for IL/NAI is one I've used a lot. =) I know there's several ones already for Anima, but so far I haven't managed anything as close to Yeagle's original style as I can with NAI and your LoRA.
@Otto_Mattiq ha! nice! i think i still have the dataset. if time allows it i train my dean yeagle version then. I looked up erich sokol. it is a very nice cartoon style, but every face is so unique, i doubt i would make a good style. i understand ai training like it boils the dataset down to one "face version" in the end. with a little variation. so i think the uniqe features of most dataset images would get lost. doug sneyd and dean yeagle faces are all in one style, so its a nobrainer it works good as a style. but perhaps ill try it. anima learns much better then illu and pony.