@iiiichimarU_03 - LoRA - ANIMA

LoRA trained on @iiiichimarU_03 artstyle for ANIMA

General usage Notes:

This style is obviously extremely stylized and leans heavily towards larger heads. If you're using it at full strength your mileage may vary with some concepts.

I've found strength to be best at:

• 1.0~0.8 for as close to the actual style as possible

• ~.8 or below when:

  • Mixing other resources (though this is to your discretion)

  • Trying to generate certain difficult concepts for this style (foreshortening, depth, detailed backgrounds)

Check the "About this version" section for more specific problems I've noticed.

General Workflow for Sample Images (on ComfyUI):

• Sampler from images is done with nothing special and the scheduler "normal"

• Initial resolution is selected from the SDXL whitepaper appendix I (https://arxiv.org/pdf/2307.01952) and is multiplied by 1.5x since ANIMA seems to be able to handle those sizes fine