Pony Version: https://civarchive.com/models/646743/dickless-penectomy-no-penis-unusual-concept-pony
Dickless / Penectomy / No Penis [Illustrious] [Unusual Concept]
Recommended Weights: 0.7-1.0
Triggerwords: d1ckl3ss, p33h0le, testicles,
Helpful Negatvies: (for male/femboys) penis, dick, cock | for futas/dickgirls: pussy, vagina, penis, dick, cock, cunt
Helpful:
testicle size easily adjustable via prompt. (small, Tiny, large, big, huge etc)
on girlish characters it can sometimes help to put futa/futanari into positive prompt, especially when pairing with males.
Optional: Adetailer/Face-Fix for Face and Hands + High-Res Fix/upscale
Known Issues: (will try to improve in the future, this is my first Lora for Illustrious)
might Need some more tries for multiple persons to get the Right composition depending on genders and genitals.
"proper cumshots" from that hole are difficult yet.
in rare cases and dependng on checkpoint it sometimes still renders original genitals. (atleast on Girls sometimes)
Description
FAQ
Comments (6)
Showcase Images are made with:
- Reforge
- Euler a
- 20 - 30 Steps
- CFG 5-7
- Adetailer for the Face & Hands
- High-Res Fix on 0.2-0.45 Denoise
- Around 10-12 steps on High-Res Fix
- 1.2x-1.5x upscale
WAI-NSFW-illustrious-SDXL (IL): https://civitai.com/models/827184/wai-nsfw-illustrious-sdxl
Fixed VAE: https://huggingface.co/madebyollin/sdxl-vae-fp16-fix/blob/main/sdxl.vae.safetensors
optional for local AI:
- 4x AnimeSharp https://openmodeldb.info/models/4x-AnimeSharp
or
- 4x-NMKD-YandereNeo Upscaler. https://nmkd.de/?esrgan
felt it just by looking :(
Oooh, this might be useful to use alongside a lora for chastity cage/flat nubs, to help get rid of any excess dicks that show up.
yeah true.
ill do flat chastity for IL soon
I use this in inpainting, and it's hard to make the peehole show up on some checkpoints, but I think that's really the checkpoint's fault more than this LoRA's fault. Great stuff though overall!!
thank you very much, appreciate it. ill update this in the near future, maybe itll work better then when trained on more recent models!