Kroma
The best version yet. It copies the style nearly perfectly and is not over-fit.
The LoRA strength can be pushed up to 1.5 to have stronger styling without too many side effects.
Kroma 0.3.1 OPD is most likely the best model I have ever touched for illustrations by a fair margin.
The LoRA works somewhat on Krea2, but you will get much better results on Kroma.
Anima
The Anima version of the LoRA does not work without the triggerword.
Anima already can do Muk's style decently. The anima only improves on it slightly if you don't use any style LoRA.
Though it fixes the white border issues that the base model has with the artist, as I cropped every image with white borders to fix it.
It also allows for LoRA pairing, as the model loses artist knowledge when using style LoRAs.
Description
Training Settings: batch size 4, 1024px, 108 images per epoch, optimizer is SinkSGD_adv
FAQ
Comments (2)
I was about to say "Anima already has Muk's style in it", but i think it's not that strong, this Lora will definitely help define it, thanks!
the LoRA doesn't do much if anything. at the very least, it does not provide this artist's style.
when I set the weight of the trigger word - in this case the artist tag - to zero, it still triggers the LoRA, but the result looks nothing like what it's supposed to look like.