I've given this a brief but detailed testing, and found that even with copying over prompts, sampler, CFG and other listed settings I can't quite get it to work the way I see in the examples with AbyssOrangeMix2 (perhaps it has to do with the negative prompt embeds that are used in said examples? I am not an expert.)

With that being said however, I changed my model used to Anything4.5 and made some slight alterations to my prompting with "flat color" and "lineart" being essential to help guide it towards the end result. I also added "realistic" to the negative prompt to help catch any outlying style influences from the model.

I set the LoRA strength originally at 0.8 but found that it was overfitting certain poses/faces, and bringing it down to 0.6 helped.

Included are example images of the LoRA along with comparison examples without the LoRA active. The sole prompt change is exclusion of the <Lora> tag, so as to give a direct comparison--even the "skistyle" tags are still in there.

Overall, this is a very useful LoRA and I'm very glad to have it in my toolset now. Thanks Nerf!

Usage:

Model: Anythingv4.5

LoRA Strength: 0.6-0.8

Prompt helpers: "flat color", "lineart" Neg prompt: "realistic"