Pepper - Paladins : Champions of the Realm
Details
Download Files
Model description
Pepper from Paladins! A Vulpin female and primarily a support character on the team with an affinity for raves, dancing and thumping music! Requested by a fellow CivitAI user.
The Mol Keun version was trained specifically on the requesters favorite Checkpoint: /model/135477
Training data is identical for both versions. One was trained using Mol Keun as base, while V1 was trained on base Illustrious for wider compatibility. Please use the recommended Checkpoints below for the Original version.
Recommended Checkpoint: /model/827184/wai-nsfw-illustrious-sdxl
Or: /model/376130?modelVersionId=1957764
Trigger word: ChickenPepper
Additional tags: furry female, two-tone fur, blue fur, white fur, purple hair, aqua eyes, purple eyeshadow, fox ears, fox tail, animal nose, medium breasts, tied shirt, bra, denim shorts, thighhighs, sneakers, full body,
Tags for her accessories: headphones, eyewear, backpack, bead bracelet, fingerless gloves, ornate ring, tail ornament, test tube, cassette player,
Other tags used during training: from side, from behind, ass, crotch, upper body, lower body, from above, from below, cleavage,
Extra info: "Full body" was used on every full body shot of her during training, always use this when you intend for her entire body to be visible, it's important for the overall results. Use "shortstack" and "petite" to recover some of her unique proportions if you notice that she is deviating from her video game looks. For other situations, use "upper body" or "lower body" where applicable. Remove and add tags as necessary, the more tags you remove, the more flexible it gets. For example, the backpack straps that wrap over her shoulders like to be there even when the backpack is not tagged for. Change her clothes enough and this should eventually disappear. The model unfortunately isn't terribly good at her hands, many of the training images simply didn't have the correct angles to catch all 5 fingers and individual fingers were often partially obscured by other things. A V2 could potentially remedy this, if there is a demand for it.
All images generated locally using the recommended Checkpoint at resolution 1152x1528. Images generated on CivitAI at 832x1216 will look similar, but not identical. Downloading the showcase images and using "PNG Info" will not necessarily generate 100% identical images because I sometimes clean up some things in the prompt info, such as the name of the Lora. Use them as a springboard for your own ideas.











