Hey !
This Lora is an attempt to get a better Piutahayo style for Anima.
Keep in mind this lora was trained on the preview2 version of Anima. You can find more informations about this model, such as generating setting on their website : https://huggingface.co/circlestone-labs/Anima
I will let you have the pleasure of exploring the nsfw part ;).
Trained on Anima preview2 model.
Recommended settings
I recommend a weight of 1 for this lora and a CFG of 4-5 seems fine.
piutahayo_style,
Sometimes "flat colors" can be used too.
Putting "false text, happy birthday, ye-pop, deviantart" in the negative prompt seems to make it work better (yes happy birthday ???).
As for text, i found that simply writing your text in uppercase between "" works best.
For those of you wondering, for an obscure reason my other tests with the @ style tag didn't work well enough in comparaison to this one.
Feel free to give me your feedback ;D
Enjoy ~
1024x1024 or 896x1152
Steps 30
CFG 5
Sampler Euler a
Scheduler Simple