• Hello Guest!

    We have recently updated our Site Policies regarding the use of Non Commercial content within Paid Content posts. Please read the new policy here.

    An offical announcement about this new policy can be read on our Discord.

    ~The VaMHub Moderation Team
  • Hello Guest!

    We posted an announcment regarding upcoming changes to Paid Content submissions.

    Please see this thread for more information.

Lets imagine and create hot girls together - Your prompts become VaM models!

You get a hidden +1 for that of course, ill have to do that to vote myself aswell. The thing is, i cant just let the other girl get the votes because her face is so different.

I could offer that we just take her initial +2 since you and me voted, every other +1 needs to be at the new post.
Fair enough; but I think the original two posts showing the image should be deleted then, because otherwise people will just drop their like on it without knowing it's another picture way down the thread they're supposed to be voting on... I mean, my "voting post" now basically has two decoys 😄

If that's not possible, I'll delete my new version and I'll stick with yours. Less complicated that way
 
Fair enough; but I think the original two posts showing the image should be deleted then, because otherwise people will just drop their like on it without knowing it's another picture way down the thread they're supposed to be voting on... I mean, my "voting post" now basically has two decoys 😄

If that's not possible, I'll delete my new version and I'll stick with yours. Less complicated that way

Alright, lets take the original one to make it easy. In any case, im happy that i could motivate someone to work around with stable diffusion themselves. Maybe try around a bit more and create your perfect girl - then easily win with her in phase 2? :)

I'm not sure it matters who wins in the end... It's all just good fun. But I do hope to meet...

Miss LOVEBOTTOM:
https://hub.virtamate.com/threads/l...r-prompts-become-vam-models.38522/post-109174

:love:

Im doing some more Gens of her right now, youll get a PM soon. :p
 
Alright, lets take the original one to make it easy. In any case, im happy that i could motivate someone to work around with stable diffusion themselves. Maybe try around a bit more and create your perfect girl - then easily win with her in phase 2? :)



Im doing some more Gens of her right now, youll get a PM soon. :p

You are a Gentleman!
 
I'd be more interested in seeing transforming VAM screenshots into depictions of real girls with IMG2IMG. Getting the settings right is what's currently eluding me
 
I might be late, but ill give it a try also
my prompt;
female, 5 feet 2 inches tall, green eyes, oval face, ( jet black hair), big lips, small eyebrows, fine nose, high cheekbones, fair skin,
sri lankan, long hair, straight hair, hair parted in the middle, slender, (slight overbite), miniskirt, oversize T-shirt, color image, no makeup
 

Attachments

  • my girl stability.png
    my girl stability.png
    435.5 KB · Views: 0
I wanna keep this thread going, now that i'm an addict to SD, it's a great 'research' tool for new looks and clothing
 
The point of the thread is that you get one picture that you nominate, and then eventually the one with the most votes wins. Unfortunately, the OP seems to have abandoned the contest.
 
I'm going ahead and making a Vam version of "Linda Smallberries" I nominated with a prompt. I'm using my own workflow through Daz and Headshop. It still takes a lot of dial spinning, but the AI image is a good start.
 
If someone wants one done lemme know too
two trailer park girls go round the outside, Nevermind :) Easyvam is back, back again...
 
Last edited:
This is an impressive way of creating models.

Do you think it would understand a term like mugshot, or be trained on arrest/booking photo's as a dataset? It would be handy to get the front and side profile images of the same model as reference for making new unique skins.
 
This is an impressive way of creating models.

Do you think it would understand a term like mugshot, or be trained on arrest/booking photo's as a dataset? It would be handy to get the front and side profile images of the same model as reference for making new unique skins.

there are two main ways to use it, txt2img and img2img

With txt2img you can force it to go that way by adding terms like "side by side, front profile, mugshot" but the results will be hit & miss, mostly miss.

With img2img, besides the text you also give it an input image. The AI will then do the same as above, but will also use the input image as a guideline.
So you can take any image from google (e.g."front profile pic") and force the AI to give you a result that follows the colors and style.

There's also an extra thing that does more magic, ControlNet, that works on both methods above. With that you can give it an extra image and the AI will follow just the pose or the outlines. Here's an old pic about controlnet to get a better sense of what it is:

Un!titled.jpg



Bonus some old stuff but my entry:
----i00009.jpg
i00011.jpg
 
Last edited by a moderator:
Well if this contest is going to be hijacked and not taken seriously anymore, then I might as well start posting random AI art as well.
 

Attachments

  • 00004.png
    00004.png
    13.1 MB · Views: 0
Last edited:
Sorry, for whatever reason the Hub didnt give me anymore notifications for this Topic so i thought there was nothing new. I got alot on my plate currenty and barely have time to get stuff done for VaM. Worktrouble will cool off sometime soon though - i still plan to create the winners as looks and do 5 phases.

Im really happy seeing so much people contribute and having fun here. A special thank you @AWWalker for offering to help out and keep the Thread going <3

Ill make sure to keep checking this regularly now as the Hub obviously doesnt like sending me notifications for this Thread.

Phase1 lasts until 22 of July and ill create all requests until then. After that, ill let the voting go a bit and finally create the winning girl as a VaM look. :)
 
I might be late, but ill give it a try also
my prompt;
female, 5 feet 2 inches tall, green eyes, oval face, ( jet black hair), big lips, small eyebrows, fine nose, high cheekbones, fair skin,
sri lankan, long hair, straight hair, hair parted in the middle, slender, (slight overbite), miniskirt, oversize T-shirt, color image, no makeup

You aint too late my guy, heres your Girl!

00011-1464107262.png
00012-2547769100.png
 
Last edited:
Think about what the Girl of your dreams would look like and let AI create it! Pairing facescan software with Stable-Diffusion 2.0 and proper NSFW checkpoints, we now live in a time where we can just Imagine our favourite girl and let her come to Virtual Life inside VaM. I thought it'd be cool to make a community thingy out of it, since not everyone has access to hard&software needed for this process.


Heres what you do: You prompt me. Imagine your Girl, try to grasp Details of her in your inner image and post it as a prompt. Ill let stable diffusion create her and answer with the generated image to your Post.

If youre already having fun with Stable-Diffusion yourself, feel free to post your Girls here and let them take part in the competition! (if you want her to be created as a VaM-model, her face needs to be unobstructed, clearly visible and without facial expression, otherwise FaceGen wont work properly)

Ill try my best to re-create the winning girl as a VaM model and release her as CC-BY (Free) content, the final goal will be to have 5 hotties created from the community's heads, smashing together a Community-Look apearance preset pack.

Ruleset:
-Favourably, the most liked girl will be determined the winner, incase we get draws or too little people take part in liking other's looks, ill make a Poll. Worst case ill just pick the one best suited for FaceGen.
-If tons of people want to take part this, remember that im unfortunately not an AI yet and cant create tens or hundreds of images per week. If it comes to this, ill click randomly through pages and pick the posts with the most detailed prompts.
-If you want some aspects of her hightlighted, put it into brackets. (X)=important | ((X))= more important | (((X))) = most important
-If you want to see genitals in the picture, please dont hide (((visible pussy))) in the middle of your prompt, it makes it fiddly for me to cut it out. Put it at the start or the end.
-Your prompts need to look somewhat like mine below, preferably each prompt seperated by a Comma.
-Only one prompt per person, per phase. There will be 5 different phases, resulting in 5 VaM models for the appearance pack.


Phase: 1 (Ends July 22)


This is an example of how those prompts look like and what came out of it. The more Details you prompt, the closer your Girl will be to your imagination. Now lets find out how the Girl in your head looks like!

Describing prompts:

female, brown eyes, round face, bright red hair, big lips, small eyebrows, fine nose, angled eyebrows, (pink lipstick), earrings round,
western europe, beach in background, bigger eyes, long hair, (round medium-sized boobs), skinny, small nipples, small aerola

Result:
View attachment 240002
This is the best one so far, but all these AI generated girls are so good, Ill be happy any of them the winners
 
This is the best one so far, but all these AI generated girls are so good, Ill be happy any of them the winners

Very glad you like her. Yea, with the right checkpoints almost all of the AI girls turn out gorgeous!
Maybe a couple of your prompts could make an even better one? Why not try creating one yourself? :)
 
Last edited:
The lateste
This is an impressive way of creating models.

Do you think it would understand a term like mugshot, or be trained on arrest/booking photo's as a dataset? It would be handy to get the front and side profile images of the same model as reference for making new unique skins.

Even if it doesn't understand, you can grab a picture of mug shot, use it in IMG2IMG, add your prompt, let it diffuse it down to replace the face, ad ADetailer with the face touch up and you can get a good one.

And on top of that, you can use the front and profile picture with a prompt like a picture of the exact same woman, from front and side, and it'll draw it good enough for face software.

I didn't put much effort into generating this, but can see the idea. Two methods
1. Image to Image -- get a mug shot image, use img2img console on SD Automatic1111
Untitled-1.jpg

Untitled-3.png

2. Use OpenPose Editor -- not going in to detail on this, but it's not hard to figure out.

Untitled-4.png
Untitled-4.png

Control points are eyes, ears, nose, neck, shoulders... hope that helps you... attached is the json ( mug-shot.json ) if you have stable diffusion you can import in to openpose editor, and below is the settings i used to produce the response from control net.
1689643088799.png
 

Attachments

  • mug-shot.json
    2 KB · Views: 0
windswept hair
Very glad you like her. Yea, with the right checkpoints almost all of the AI girls turn out gorgeous!
Maybe a couple of your prompts could make an even better one? Why not try creating one yourself? :)

Ok I will thanks,

Female, age 22, Dark blue eyes, bigger eyes, small eyebrows, angled eyebrows, round face, auban red hair, long wavy hair, big lips, fine nose, European, Tanned skin, Round medium-sized boobs, small perky nipples, medium textured aerola, athletic, sports oval in background, ((small gap between front top teeth)), (Tartan skirt), pleasured expression, rings on fingers
 
Last edited:
Back
Top Bottom