V2: This lora is trained on more close-ups than V1 with better quality images, it brings better flexibility and realism, it should even work good on men now! Use with standard .9+ strength, lowering too low will just affect the reliability of the results. There’s still a light bias towards large hips, but easy to control with things like narrow waist.

***TIP: If the bulge is too large, use without any trigger words (better realism).

V1: This lora is trained on a dataset with image of very large futanari bulge and various clothing; classy pants, yoga pants, jeans, bikini, panties, lingerie, latex and shorts, the body is mostly bimbo style. Flux seems to prefer lower strength settings on lora for better prompts, using a dataset with exaggerated features. My idea for this lors is that you can get the things you want with lower lora strength, if my theory is correct! Else, we’ll have huge bulges…