🧬🍆🍒🧬

The most based version of the based mixes injected with the dominant genetic futanari factor from furry models!

Credits:

/d/ guys for the original AbyssFutaFufflyKemono Mix!

ninecats creator of futapocalypse 1.0 - the model is heavy carrying the dongs like fluffyrock carried the original futanari mix

Fake-Person and shupendupen repos as the source of futanari_v2_e3_s10000.ckpt

lodestones creator of the flufflyrock models

toynya creator of the crosskemono models - the model is heavy carrying dem balls

Merging formula:

This model is the result of merging various models using block weight merges, add difference and train difference for different steps and then added to stable bases that are widely know for high image quality and lora compatibility.

↑ family jewels tree ↑

MBW presets:

Hard lower body preset

0.5,0.6,0.4,0.8,0.3333333333,0.4444444444,0.6666666667,0.4,0.6666666667,0.6666666667,1,1,0.6666666667,0.5,1,1,0.3333333333,0.5,0.75,1,0.75,0.6,0.3333333333,1,0.6

Mid lower body preset

0.25,0.3,0.2,0.4,0.15,0.22,0.33,0.2,0.33,0.33,0.5,0.5,0.33,0.25,0.5,0.5,0.15,0.25,0.375,0.5,0.375,0.3,0.15,0.5,0.3

Configs:

VAEs:

Use it with kl-f8-anime2 or vae-ft-mse-840000-ema-pruned

Clipskip = 2

CFG Scale = 6~10

Sampling methods:

Euler a: 26 <- best ->

DPM++ SDE Karras: 35

Upscalers:

4x-AnimeSharp - 10 steps - 0.15~0.35

4x_foolhardy_Remacri - 10 steps - 0.15~0.35 <- best ->

Positives:

futanari, SIZE penis, SIZE testicles

<- replace "SIZE" with the actual penis/testicles size ->

Negatives:

furry, fur, animal ears

<-not really necessary, but use those if the furry genes get out of control->

Lora compatibility:

From my limited testing, the model should work with any kind of lora trained with clipskip 2. I tried my best to dilute the furry models without losing the dongs, balls and at the same time watching for compatibility issues. If you see something that should be working but looks fried, try using earlier epochs of the same model, reducing the weights or using the lora block weight extension.