Style model based on the style of REK1610 — If original artist request takedown then I will remove it from CivitAI.
Please support the original artist on their Patreon
Settings & Recommendation
Trigger Word: @rek1610
Strength: 1.0
Sampler: ER_SDE
Scheduler: Simple
If you want to get results closer to the example images, use any Turbo LoRA such as Anima Turbo LoRA with CFG 1 and 8 steps. Alternatively, you could use a 16-step distillation like 16-step Acc LoRA with CFG 1 and 16 steps.
I also recommend using a finetuner like RDBT b1 v2.1 base if you want more detail.
Experiment with reducing the strength of character tags if you want a character to have more or less of the style that was baked in Anima. [e.g., "(Princess Peach:0.8)" ].
Description
The artist does not have a wide repertoire of art so this is being trained on a medium-to-small dataset trained with natural language captioning





