*I've updated the User Guide once again. 11/16/24

NTR | illustrious-XL | Noob-XL can draw this model. It will be anime-style instead of realistic, but the results should be good.

DPM++ 3M SDE/Uniform/128 steps

*I've updated the User Guide. I now recommend using Zaxious_XL v4.0

as the checkpoint when using this LoRA. More information about this change can be found below.

Our public officials take the security of our detention facilities seriously. Detainees and inmates have been known to introduce contraband into these facilities by concealing the prohibited items on, or even inside, their person. This LoRA will help our law enforcement and corrections officers thoroughly search individuals to ensure their safety isn't compromised. The model will help you generate images of ???? fingering, ass focus, anuses, perineum, penises, ?????????, pants/underwear pulled down and guided penetration. It can also draw female subjects. It was created with 27 original stable diffusion images of my making at about 1.5k steps. All images in the data set were auto captioned and then personally pruned and amended. The captioning is tag-style using danbooru tags. That means many of the tags contain underscores, ie. - "anal_fingering" will probably get you further than prompting "???? fingering".

The trigger is "cavity_search" and I will also add "anal_fingering" as a supplemental trigger. This model will draw anatomically correct and aesthetically pleasing penises and erections of various sizes, granted that you are using a checkpoint that also knows how to draw them or are using an additional LoRA/ect to help guide the magic towards a human Earth-????, as opposed to the dystopian nightmare ????? that live inside the majority of models.

This model can create unusually expressive and realistic facial expressions. I didn't intentionally train the model with that goal in mind but it makes sense since I did extensively prompt and inpaint some of the sorce images to reflect intense emotional reactions. I've yet to find a checkpoint that can draw these expressions and draw male ??????? correctly in the same image. Below you will read about various ways I am trying to deal with that. My honest assessment is that this model has some unique and artistic potentials that I haven't quite figured out how to unlock. Maybe you will be the one to figure it out.

I will continue updating and improving this model as I've already invested a great deal of time in it and am motivated to see it fully realized.

*Update - LimitlessVisionXL v3.0_VAE

The checkpoint linked directly above in blue letters can be used for inpainting any fixes or changes to intimate ??????? that you might want to make along the way. Unfortunately, LimitlessVision v3.0_VAE dosn't perform well with this LoRA as the text to image base. I'm not sure why that seems to be the case, it works very well as the checkpoint for the LoRA I released under the title Prison Pipeline (gay).

I recommed using Zaxious_XL v4.0 to draw your scene and then inpainting ass/genitals/breasts/ect with LimitlessVisionXL v3.0_VAE when necessary until I can engineer a checkpoint that can do both things I need it to do.

Conversley, you can just stick to the Zaxious_XL v4.0 checkpoint and use a ????? LoRA, embedding,ect to generate male intimate anatomy.

I recommend something like Small ????? [Anime] Flaccid and Erect

for small to average male genitals or Xpenis

for average to large male genitalia.

Not specifying a racial description of your subject/s tends to draw Caucasians. "dark_skinned" or "dark_skinned_male" tends to draw African-Americans. "latino" will conjure Latinos.

Stable diffusion has many biases, as it can only know what it has aleady been taught. One of these biases is against pubic hair. The tag "excessive_pubic_hair" will usually get you neat and trim pubic hair. Not using it will usually result in a shaved/waxed situation. Perhaps the largest bias stable diffusion suffers from is its compulsion to draw women, as opposed to men. The nature of this particular LoRA can engage this bias at times. I will discuss negative prompt tags that are useful for this model below. Positive prompt tags that may be useful include - "yaoi", "male_focus" "male_pubic_hair", "bara", "clothed_male_nude_male", "fingering", "fingering_from_behind", "deep_penetration", "anal_object_insertion", "?????", "?????????", "perineum", "pants_pull", "male_underwear_pull", "??????????", "bent_over", "kneeling", "ass_focus", "squatting", "latex_gloves".

The theme of this LoRA is more focused on the visual and digital inspection of inmates and detainess rather than the physical setting of a detainment facility. The following list are prompts that can help increase the chances of generating secure facility themed backgrounds:

"prison, jail, inmate, prison_guard, police, policewoman, police_station, cell, prison_cell, cellmate, prison_clothes, exam_room, cuffs, shackles, handcuffs,"

The model was trained on images of various sizes that corresond to standard SDXL dimensions and should do well in either landscape or portrait.

The reference images I provided are all unaltered and come straight from text to image using this LoRA in prompts that did not exceed 75 tokens. No other LoRA or textual inversion, ect, were used in their creation. DPM++ 2M KARRAS, 38 STEPS

Negative Prompts are helpful for this model due to the subject matter triggering stable diffusions compulsion to draw female subjects. I find "?????" and "1girl" are helpful as negative prompts if you see mangina penetrations materializing. That should get things keistering properly. The subject matter might also be used to draw images that are focused on human hands. Despite the training images only featuring anatomically accurate hands, stable diffusion might still produce blurry, four or six fingered hands. You can use a "negative hands" type embedding in the negative prompt. That should be enough of a nudge to get great hands. If you still are having issues with hands, add a hands type LoRA to your prompt. They often won't affect the style or character appearance of your image. I recommend a LoRA that shows up as "detailed hands" in my UI, so you can try entering that into the searchbar of the models tab on this site.

Different lads respond differenly to this type of administrative procedure. How will your subject react? Does it make them blush? Embarrass them? Make them angry? Shout? Sweat? Become arroused? Try adding some emotion and_or reactions to your prompt.

Camera angles can help you get your image just right. Some camera prompts might be "close_up, wide_shot, from_above, from_side, from_behind, ect".

Please post some images on this page of any particularly spicy or interesting things you might create using this resource.

Enjoy.