A simple LoRA trained on the better images I have generated from Perchance through nearly a year and a half of prompt refinement, learning, perfecting, and a ton of trial and error. This model focuses primarily on anthro wolves and also a modest, non-muscular body types (If you are like myself and do not like the overly-toned bodies, then check out the "Muscle Slider" LoRA by Ostris, here on CivitAI. Set to negative values for a more natural looking body, or set to positive values for some humorous muscular characters).

If you enjoyed the LoRA, please leave a review and a thumbs up! The LoRA is mainly trained on anthro wolf images and may not handle other species well, I have not tested. Please comment if you have had any luck with other species!

Triggerwords for this LoRA are Rear View, Solo, Wolf, Anthro, and Furry

LoRA strength seems to work best from 0.4-0.7
Steps: 25-40 work best

VAE: vae-ft-mse-840000-ema-pruned (works fine with others)

Clip Skip: 1

Sampler: DPM++ 2m SDE (Karras) works best, Euler also does well.

Best results so far come from using the BB95 Furry Mix v. 14.0 and Indigo Furry mix v110_Realistic, both for SD1.5. The LoRA works fairly well with any other model, may need to place emphasis on Furry and Anthro tags.


Hope you enjoy and have fun! Feel free to reach out with questions or (friendly) comments. this is my first ever LoRA, so be nice!