Please note, this is a test lora, I haven't found any style lora that's based from one of my favorite author, so I tried to train one. Feel free to provide feedbacks and criticism. Thanks!
Trigger word is danimaru style
Description
Retrained on new dataset from danbooru and other source like pixiv. Consists of 250 mixed images from colored to black and white manga pages and solo illustrations. Used Ars Divina as base model.
FAQ
Comments (4)
So, the problem I see is that even if you don't use the lora, using an illustrious model will give you pretty much the same result, at least something very similar.
Using the Lora more or less fixes the style that is already included, but if you trained it using "D4N1M4R00" instead of using a tag that already exists, I don't think it would work properly... So I would try to look a bit more that way, a style that doesn't rely on a tag that already exists because different mixes will give different results and the Lora could be better for cases like that.
tried that but it doesn't give the desired results with just based illustrious.
@NeverWinter13 Yeah, Illustrious have a hard time with Loras, I hope you get it right soon, I really like this style.
@Not_your_boss yeah, I like the style of danimaru. Hoping that I can properly do it justice when I retrain