This LoRA will give you the pose of woman kneeling down and licking someone's feet, which appears come from off-screen. It's quite similar to My LoRA -> Foot to Mouth, except this one involves foot licking while kneeling, while the other I mentioned involves foot sucking.
It seems the Z-Image Base version produces quite a bit of image distortion. Therefore, if you don't need the variety of images or features of the Z-Image Base version, I recommend using the Z-Image Turbo version instead.
However, using Base-trained LoRA with the Turbo version is excellent, and I highly recommend it.
V2 - Z-image {Base|Turbo}
This version is quite similar to LoRA - Foot to Mouth that I recently released. The only differences are the art style and camera angle.
Example prompt
ttai_kfl@v2, lezdom, foot domination, foot humiliation, foot licking, foot shoved to mouth, POV angle, from above, overhead shot.
realistic humiliation photo of submissive female.
she is opening her mouth and use her tongue for licking someone foot shoved into it.
knelling on floor, {cobalt heavy black armor, blue shawl|no outfit}.
{single|double}-foot shoved to mouth.
the scene is Romania castle.Some helpful tips are:
- sampling step 8 - 10
- res_multistep / simple
- cfg 1
- weight 0.45 - 1
V1 - Flux 1 Dev
Example prompt - No Outfit
Realistic photo of a below asian femal kneels down and looks up at other woman who put foot into her mouth and forces her to lick it. she has black hair, no outfit. She kneels and sees the soles of her feet. The scene is middle of the river.
kfl, from above, dogeza, foot come from out of screen and reach out into other woman mouth kneeling below her. deep-blue toe nails.Example prompt - Outfit
Realistic photo of a below asian femal kneels down and looks up at other woman who put foot into her mouth and forces her to lick it. she has blonde hair, cobalt heavy silver armor. She kneels and sees the soles of her feet. The scene is room inside old castle.
kfl, from above, foot come from out of screen and reach out into other woman mouth kneeling below her. green toe nails.* During the build, some weird images may appear mixed in with the complete image. I recommend choosing a batch_size of around 3-4 to help increase the chances of getting a complete image.
** However, if the resulting image isn't too weird, you can inpaint it to fix it, which is how I used it in this LoRA example image.
Some helpful tips are:
- sampling step 15 - 50
- euler / beta, simple, sgm_uniform, normal
- cfg 1
- weight 0.7 - 1