Make some fine-ass moms. Well, try to - this is an early model.
This is an early LoRA that adds saggy breasts, some fupa/tummy depth, and c-section scars. Still working on training and variety, results may vary based on the core model you use this with!
Use in a 0.5 - 2.0 scale - I tend to find 1.2 to 1.5 gets the best results. V1 is narrow, V2 has more variety, but possible for more concept slip.
Pretty Good / OK (Needs Work) / doesn't always work well
Version 2 Details (Flux Only for now)
Trained on these concepts, pair them like "small saggy breasts"
breasts: tiny, small, saggy, empty, perky,
nipples: small, medium, oval, round, dark, light
pubic hair: shaved, groomed, stubble, strip, triangle
fupa: small, medium, (pair this with "soft build" or "fatty" desciptors - doesn't work well on skinny)
stretch marks: on stomach, on thighs
hips: wide
build: skinny, soft
cellulite: on legs, on butt
venus belly - some base models treat this as pregnant?
dimples of venus
c-section scar
soft pubic area
Version 1 Details
Some things to try:
mombod - will help influence the woman to have some/all of these things below:
tiny/small/medium/large saggy breasts (use a size, or not!)
empty breasts
small/medium/large nipples
round butt / saggy butt
wide hips
cellulite
small/medium/large fupa (or just fupa)
soft pubic area (for a fatty/soft looking vulva)
c-section scar
Small dataset, so it doesn't always play nice to specify these all individually. Breasts/fupa tend to follow the weight of the woman (larger women will have more pronounce fupa and slightly heavier breasts) - again more to train!
Used the Civ Trainer, so YMMV with this LoRA. Couldn't figure out how to put it in Early Access / request buzz - so enjoy. Tips are nice.
Flux 2 Klein
Use this Lora at 0.5 up to 2.0, outside of that, results can go wild.
SDXL
I don't run these models, you just got a LoRA because I accidentally generated that one...
Description
FAQ
Comments (1)
If you find a good balance, please post. I don't use SDXL and don't have my prompts/LoRA set up to do clean tests.
I think this will need to be a 0.6-0.8 for more natural results.


