TRUALITY ENGINE (PRO)
THE BRAND NEW PRO VERSION IS A MAJOR ENHANCMENT OF V3
More Contrast, More Sharpness, More Depth and More Warmth and all your older and more current prompts from Version 1-3 will still work in the PRO VERSION. DON'T USE VAE
PRO VERSION ADDED JANUARY 6TH 2024
click here if you want to support me and donate to our common cause
SAMPLE IMAGES
Everything you see is coming straight out of the engine, i am using no Loras to create effects, cameras or moods. I am using women and faces from this checkpoint and of course my very own Vicious Girls Series
HOW TO PROMPT?
Prompt as you wish, depending on what you are going for it doesn't need many words.
If you want to create a particular scene with particular assets, put those in brackets or double brackets. If that doesn't do the job use the weighting method (+/-) from 0.5 to 1.5
The higher the weight, the lower the quality.
SETTINGS
Steps: 35
with this model, too high means it takes longer to load, test it for each scene individually, keep it at 35 if you don't mind waiting a few seconds longer.
CFG: 3.5 to 11.5
The higher the value the higher the contrast and the amount of colors. The engine will listen more subtly to your prompt up till CFG 7. Everything above could be described as over-exaggerating your prompt to adding new features. A Lens Flare can become something entirely else. Definitely interesting if you feel like experimenting.
When using TI's and Lora's you should not go higher as 4.5
Best Size 512 x 768
Best Samplers: DPM++ 2M SDE KARRAS
HIRES FIX
Upscale a 512x 768 by 2x
UPSCALER for V3 4x UltraSharp
Clean and less grainy results, yet very detailed
(Resize Steps 15 to 25, Denoise 0.23-0.33)
UPSCALER for V1 & V2 & Pro: 4x_NMKD-Superscale-SP178000_G.pth
(Resize Steps 10 to 15, Denoise 0.17-0.33)
This upscaler provides a special layer of depth for photo realistic images, slightly grainy.
LORAS
Most Loras will distort your results, you have to weigh them in, 0.7 seems to work best by default, each Lora is different and behaves differently. Many... if not the most will mess with your models face because they were trained with a different face, sometimes you can fix it with clip skip and if that doesn't work just use roop or face swap and write your girls face over the trained face. Most Loras are designed for 512x512, when bodies or objects get strethed or appear twice then lower the output size.
WEIRD FACES
Especially from side angles and long distances you will keep running into weird faces, Hires Fix will 9/10 times fix this.
Please share your favorite results, consider rating this model if you like it.
Get your own custom and exclusive Influencer---> contact me in chat
or see my Fiver Page
DISCLAIMER: All Models released under the Vicious Family are fictional, I can not know every person, every actor, every influencer in the world and resemblances are purely coincidental. While it is possible that I take inspiration from a look that I like, I have not, not will I ever use any of their original fotos to create a resemblance on porpuse. All my models are either entirely created by prompts and trained on the results or mixed with body parts, facial features of 2 female friends who have consented in writing for their images being used in my process. If anyone feels that one of my models resembles a celebrity, actress or influencer and feels concerned about it, please leave a comment, I will take it to the person, management or representative directly.
Description
If you are already using V1 or V2 as your main generation models, you should upgrade to V3. You will be able to achieve the exact same results as before but if you add more details into your prompt you will easily exceed your previous generations by a lot.
This is the final Version of the Truality Engine, I am certain that I have squeezed out every single pixel possible for this particular concept.
NEW in V3
1. More detailed shading
2. More light sources
3. More light effects
4. More contrast
5. More overall environments
6. Faster Image Generation
7. More consistency
8. More details on female bodies
Enjoy creating your visions, please share your results and if you are happy with this checkpoint please consider reviewing it.
Cheers :-D
VAE PRUNED - DON'T USE CUSTOM VAE
Please Read Main Thread for HighRes Fix changes for V3
FAQ
Comments (44)
Even minimal effort prompting can give some breathtaking results with this one! Wish I would have found this model sooner, but glad I jumped in at v3! Just beautiful!
Thank You so much :-D IN all fairness, as good as this one is listening, it sometimes it will ignore you when it has set it's "mind" to a scenery but 8/10 times you'll get what you ask for. :-D
Yeah, very nice and it plays well with most of my LoRA's.
@kmdcomp that's good to know, some Loras create a mess, environment loras and backdrops seem to work perfectly
It respects my prompts, so its win for me.
I am glad it works for you, I was trying to make it listen to a broad range of prompt styles, it works for many but not for all, i am happy it works for you :-D
does it need Adetailer???
NO :-D
I am not a friend of Ad Detailer or Controlnet, I was trying to put a checkpoint together where you can achieve these kind of effects with your prompt.
Ad Detailer would make sense if you want a consistent make-up with particular colors or a tattoo or freckles... something like that. I have not used any other tools for all of my model samples on this site.
That being said, I personally use and highly recommend HighRes Fix because it solves the issue of facial distortions when the Actor is standing with the back to the camera or sideways... or further away for that matter.
@Tower13Studios damn you're right, it works perfectly without adetailer, really great work!!
@hxra21 thank you, hope to see your results :-D
THANK YOU
since we cannot message people directly for their generous tips, support and reviews and it really bothers me that I cannot thank everyone personally, I want you to know that I see all your support and that I highly appreciate it. All I can do is follow someone when they've sent me "buzz" to acknowledge their support. We need messaging as a feature.
It's a great checkpoint, thank you. V3 preserved a bit of colourful artistry as well as realism. Would you be willing to share your process for your embeddings? You mention you use daz, would you be willing to write up a guide, or maybe just some quick pointers?
@AndromedaPerson Thank You :-D. You will find that the Pro Version has the artistic features enhanced, creates even better colors, high dynamic range a-like outputs and the generation process is faster.
A guide is overdue but i am using less and less DAZ and go straight into Blender or Unreal Engine since DAZ has such a horribly slow renderer.
@Tower13Studios Indeed, I'm just trying it out now! That's interesting for the embeddings, is this so you make a huge dataset? How many images do you use for them? What about captioning? And do you train them with A1111, kohya or something else?
@AndromedaPerson I train them in A1111, Kohya is the trend of the month, they come and go and are no delivering true consistency either. Yeah i use massive datasets, 50 to 100 images because if you only do the recommended 10 and mirror them to 20 then you have a consistent face but zero personality. I always add things in the background, mess with outfits, environments, cities etc... all of that becomes part of the embedding and the results become a character.
@Tower13Studios Thanks for the explanation, that is a bit different, nice technique!
Am i using it right if i just put the file into my SD Models folder, and add this special models prompt into the generator? Cause i hardly see an effect.
I feel there is a language barrier.
The file of the Truality engine goes into the folder.....i\models\Stable-diffusion
If by "special model" you are referring to one of the girls, these are embeddings which need to be downloaded individually into the folder Embeddings
@Tower13Studios Allright! That was my questions. Thank you for your amazing work btw. The problem is that i dont have that folder. May i just add it?
Second question: The 4xUltrasharp upscaler does modidify your models kind of comic like, and with loss of detail in skins. Do you have a solution for that? (i used your recommended upsacle settings)
@Crankerrr yes add the folder to the mainfolder on my PC it's called ....stable-diffusion-webui\embeddings (in Automatic 1111 of course)
The Ultrasharp shouldn't polish off the skin but I do know what you are referring to.
Try to put the CFG scale to 5.5 that brings out more skin details and contrast
Add the Highres Steps by 5 and reduce the denoising strength to a level where you are happy with the result. Highres Step shouldn't be done for simple upscaling, it should give strength to the facial structures too. If you are still unhappy, use another HighRes Model, I liked this one the most in terms of quality in context with speed. You can get better results but it takes three times longer. I hope that helped :-D
@Tower13Studios Thanx bro. Whcih upscaler do you reccomend? I mostly use latent, but at those low high rez steps it comes out blured, but with ultrasharo it works.
@Crankerrr I only use the Super Sharp right now but this one is great too 4x_NMKD-Superscale-SP178000_G.pth
@Tower13Studios Thanx bro, thats a nice one! Sometimes i just put a output into img2img and rezize it x2 with euler a and minimum values, that does not swallow so much details, at least i feel like it does not.
one other questiosn: Do you know a way to get different face expressions into 2 persons at the same time? no mater what i try, its always a "smile" or whatever, on every person in the output
@Crankerrr Highres Fix is not just an upscaler it gives stability to the faces. Often the faces look distorted when it is a full body shot, the Highres Fix will prevent that. For actual upscaling i use Topaz Gigapixel AI, costs nearly nothing and you can upscale 100 images in 30 seconds to 8 times the size with basically no losses.
Sorry i don't know how to get 2 different impressions with 2 girls in a prompt. You could do it with control net where you control the pose and the faces with open pose and DW Openpose Model
Bro i know its off-topic, but could you help me with two questions?
1. How can i find the used prompts of a jgp or png file? As far that i know its possible to save the prompts in a file. but how do find em?
2. Which is the best plattform to promote and sell AI Adult content, since Onlyfans does not allow AI and Patreon does allow nudity, but not sexual content?
@Crankerrr in Automatic1111 there is the tab PNG Info, drop the image with the prompt in there and send it to txt to image, otherwise there are extensions you can install where you can store prompts...
Only Fans does allow AI as long as it is labeled as AI otherwise you won't get through verification.
Don't know any other sites.
@Tower13Studios thanx! When i send it to txt to img it just keeps the prompts of the image that i generated before. And in the png info tab there are no infos. does that mean that the prompts was not saved inside of the image file?
@Crankerrr yes that's what it means, probably someone edited it already
Amazing work. Just the best merge out there.
Thank you very much :-D
a new update, I'm curious, and I will test it thoroughly. the "Vicious family"? What's that and why do I need to call my lawyer..
There wasn't an update since the Release of V3 , the site was completely bugging out yesterday, it wasn't showing comments and image posts or images, I thought if I unpublish it and re-publish it, the system would fix it. Apparently there is a bug that makes it appear to be updated which isn't cool.
No need for lawyers, the girls have pre-written prenups.
@Tower13Studios have you a Ko-Fi?
@Brekuji I actually just signed up for Ko-Fi the other day but didn't get around to do much yet as I am releasing my 100th embedding today (in 1 month) :-D https://ko-fi.com/tower13studios
What is new in this V3 Version appearing in the CIVITAI "latest releases list" Dec. 3 2023, compared to the previous released V3 Version from Nov 11 2023? Something must have changed otherwise i do not understand why is it listed as NEW having had its release weeks ago?
Thx in Advance
Nothing is new, the Website has a bug. Last night the display of the galleries was all messed up, so i though I'd unpublish it for a second in hopes it regenerates itself when I republish it, apparently their system is bugged and shows it as a new entry
@Tower13Studios Ah :D
Thx so much.
So i will remove this in 5 mins.
Actually it should be pinned to top alltime if it would be me.
Same to your embeddings.
:D
I was just confused if you changed the model because Yogi just did that without infos with his V1 Models.
@LVNDSCAPE please keep it, the more people know the better, I added the info into the description as well
Great, beautiful results easy to achieve
Thank you, glad to hear that. :-D
I really love this checkpoint and all the embeddings created for it. It works great without adetailer, and putting (hands:1.25) in the negative prompt alone yields surprising results, positive, mind you. Thank you for all your hard work!
Thank you and thanks for the recommendation with the hands, will try that.
@Tower13Studios Please do. I stumbled on the tip on reddit and it works better than all the negative embeddings I've ever tried. Beats ADetailer as well, not to mention it doesn't require additional processing time.
Take a look at the latest pictures I uploaded in the gallery.
@tvange365 i tried it, I didn't see much of a difference but my prompts are more complex and it might just not take it.