My first attempt at creating a futa concept which is focused more towards semi realistic 2.5D anime. It works great at lower weights and i would recommend to use it somewhere between 0.5-0.8. Going higher will introduce more artifacts and a change in style. Personally 0.6 was a good spot for me.

This model was trained on 105 images with focus on variety.

Works great with my DarkRevPikas checkpoint.

Recommended settings [Important]:

This model works best with a high step count. Using a lower count will usually results in multiple testicles or penises. My recommendations are as follows:

steps: 35-45

CFG: 5-7

Resolution: 512x768 or 768x512 (Afterwards send it to img2img and set resolution to 2x)

Instance token: ftnri

prompt:

<lyco:ftnri-FutanariConcept:0.6>, ftnri, futanari, penis, testicles,

You can also add more details like:

<lyco:ftnri-FutanariConcept:0.6>, ftnri, futanari, detailed penis, detailed testicles,

or

<lyco:ftnri-FutanariConcept:0.6>, ftnri, futanari, extremely detailed penis, extremely detailed testicles,

negative:

Use this in the negative to minimize a chance to get a extra penis or another pair of testicles:

(multiple testicles:1.4), (multiple penis:1.4),

Best way to use this model:

First do a txt2img. Then send the result to img2img with 2x resolution. Then send the result to inpainting and fix up the details like face, eyes, hands, penis/testicles etc..

If you are trying to replicate my images with same settings, you wont get the exact image because the seed has changed when i sent them to img2img. But if you follow my workflow, you will get the same or even better quality guaranteed.

Challenges with the model:

Something i'm trying to fix is when you send your image to img2img, you might notice on naked characters, they might get skin folds or fat rolls around the crotch and stomach area. I'm trying to figure out how to minimize this results. Something that worked great for me is to lower the weight even more when you get these results. Going down to 0.35-0.5 worked great.