A LoRA for characters getting swallowed by giant snakes. I didn't do this sooner because I could've sworn there were already a few out there for Illustrious, but I couldn't find any when I wanted one.
Plan is to include headfirst, feetfirst, and fully swallowed, but don't have them all made yet. Won't ever include digestion because of (a) personal preference and (b) CivitAI ToS.
Inside_IL_v1: Characters inside a snake. Doesn't include digestion.
Trigger is: snakevoreinside, snakevorexray, snake, oversized animal,
Some checkpoints (notably NTRmix v4) require "vore" as an additional trigger and potentially increased LoRA weight (try 1.25).
Can work with character LoRAs, but may require higher weights and/or lower weighting of the character LoRA (e.g. 0.8-0.9).
Feetfirst v1: Focused purely on feetfirst vore to give a more stable result. Also trained on-site vs IL_v0.1 instead of locally vs IL_v1.0
Main trigger: feetfirstsnakevore, oversized animal, snake, vore
Set depth with one of: kneesupout, waistupout, chestupout, faceonly
Not all depths are equally stable, "kneesupout" (knees up out) is rather finicky. Also my example images with it keep getting flagged for review.
Still has some stability and compatibility issues, particularly with character LoRAs. Some checkpoints just really don't like doing vore.
IL_a1: Didn't pick up all the concepts and is unstable when combined with lots of character details, so I'm calling this an alpha. Will have to collect more training data, clean out mixed concept images, and try again.
So far Hassaku XL v3 is the most reliable for headfirst, but WAI-NSFW-Illustrious v14 is better at feetfirst. Doesn't work with most other checkpoints. Leaving it up while I get all the v1 variants created.
Reliable:
Coiled: snakecoil, snake,
Semi-reliable:
Feetfirst: feetfirstsnakevore, oversized animal, snake, vore
Unreliable:
Headfirst: headfirstsnakevore, oversized animal, snake, vore,
Swallowed: snakevoreinside, snakevorexray, snake, oversized animal, (haven't gotten this to work yet)
Description
Version focused purely on feetfirst vore, trained on Illustrious v0.1. More training data on that specific concept than a1.
FAQ
Comments (2)
I would have loved to see this model work but unfortunately it seams that the training didn't really do anything, i use a more standard checkpoint (ntr4) and i can tell you that this lora doesn't add much as a vore model.
However I hope that you can improve this model and maybe make it one of the best vore model here
NTR mix is still a bit unstable with the new version, but it's better. There's only so much I can do, some checkpoints don't like some concepts.
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.