Makes tonguejobs more consistent!. Pony XL already supports the concept, but I had the training data so I figured I'd create an assistive LoRA. I've found I like the results better with this LoRA and don't have to bump the weight of tonguejob up as much to make it work.

A weight of 1.0 works well for initial generation, along with the trigger words "t0ngu3j0b, tonguejob, licking a cock". For upscaling, you can cut the weight to 0.5 since the base model already understands the concept.