Intro

This model is intended for anime/manga-style image generation with "pants-type diapers" (and often called "pullups" by us westerners). That being said, this version was trained on a greater variety of art styles, including more realistic ones. Unlike earlier versions, only white pullups were included in the training set. Print pullups might still work, but I intend to train a couple variants of this LoRA for specific designs (namely a couple Moony Super Big XXXL pullup patterns).

Usage Information

No trigger word is required for this LoRA. Add "diaper" for a dry pullup, and add "wet diaper" to the prompt for a wet diaper. For a messy pullup, including "soiled diaper" in the prompt may yield results depending on how high a guidance value you give to the LoRA.

This LoRA was trained on Illustrious 0.1, but the images posted alongside this model were generated with Nova Anime XL. That said, I tested a "beta" version of sorts with checkpoints that use more realistic lighting / shading, such as Nova Unreal, with pretty good results. For best results, a low CFG scale (like 2) is advised.