This is the first experimental version—it turned out that compiling the dataset wasn't exactly easy 😅, since the goal was to create a realistic-looking bulge. It’s not perfectly stable, especially with looser-fitting clothing. I recommend using it with panties, leggings, dresses, and the like; the only thing I haven't tested it with is skirts. The LoRA performs best with Krea2_turbo_int8_convrot; it works significantly worse with models from Pornmaster.
I recommend starting the prompt like this: "A young futanari woman, the huge, long bulge visible through her ___"—filling in the blank with an item of clothing, such as leggings.
Specifications:
Trigger word - Futanari woman and huge, long, bulge;
Strength approximately 1 to 1.25 (1.25 is better);
Sampler/Scheduler euler/simple;
Steps ~8+4.
Description
FAQ
Comments (8)
great job!!this is amazing work!
Thanks. To be honest, I’m not really a fan of this LoRA; I used a small dataset, and it’s a bit unstable—basically, it needs some more work.
@_Chidori_ This is the best iteration I’ve used. Other LoRAs in this category tend to heavily modify outfits and are very unstable. I wish I could treat you to a coffee, but it isn’t possible where I live(cn). Anyway, your work is great.
@zss2088196014332 Thanks for the review. And does the abbreviation (cn) stand for China?
@_Chidori_ Yep, I’m really glad to meet a fellow creator who shares the same interests. I believe your work will keep getting better and better.
you are my hero
It's a bit too big.😂
A little bit 😸, but it looks good.






