So, It's been a while since I had time to make a LoRa, and I decided to go with Lopunny. I saw an image a long time ago that I thought was effectively the 'perfect' Lopunny where it was just anthrofied enough to not look 100% model accurate, while still maintaining a large part of the spirit of it.

And as nice as the other Lopunny Lora's are, I wanted one that was closer to the correct canon height.

A common misconception is that Pokemon are large. When in reality 60-70% of them are smaller than the average height of a woman. A few are between that and the average height of a man, and then the rest are slightly... to massively larger - and I wanted to be a bit more accurate.

Anyways... onto the parameters.

It only seems to work 'well' and 'frequently' on a FEW models.

Yiffymix (which is old, but still works quite well) - https://civarchive.com/models/3671/yiffymix?modelVersionId=1623867

Wai (choose your preferred version) - https://civarchive.com/models/827184/wai-illustrious-sdxl?modelVersionId=2514310

Furstein - https://civarchive.com/models/2413819/furstein

Tags: 

alop, petifeet, (lopunny:0.5),  colored sclera, black sclera, pokemon (creature), brown fur, furry, standing, full body, furry female, body fur, two-tone fur, brown rabbit tail, rabbit girl, eyelashes, (shortstack:0.8),

For even closer to on-model (if slightly stylized) - Add:

toon \(style\), 

For even CLOSER to on model, add:

chibi body,

((Petifeet helps keep her feet from becoming 'humanlike'))

when facing directly at the viewer, remember to put things like 'tail' or 'rabbit tail' in negatives.

Sampling Method & Schedule Type: Choose whatever one you feel is best from the above image, or use your own preferred!

Hires Fix is highly suggested to use at a 896 x 1152 ratio, with 1.3 upscale-by. And I used R-ESRGAN 4x+ Anime6B

Denoise: 4-7

Sampling steps: 35

Hi-res sampling steps: 15

CFG Scale 3-5

It's also HIGHLY suggested to use Lazyneg and Lazyhand

The rest is up to you!