This series is focused on women and the female body, in every kind of setting or style. It can do fine art, social media style, surrealism, cinematic, and pornographic very well. The main goal is to have as much diversity as possible in an SDXL checkpoint. Diversity in ethnicity, body type, faces, posing, personality, and locations.
If you like it post your creations!
Donations to my Whop account will exclusively go towards hardware upgrades. Civitai Buzz will be used for site generation and promotions.
Upgrading my local hardware is my on-going goal so I can train locally.
Local training is imperative today. With Ai data centers there will be a vast drain on resources in the areas they operate. Since model & lora training will no doubt be hosted in the new centers, I want to be able to run locally only. The moment my Ai hobby causes actual harm, on any scale, is when I start collecting stamps instead.
General Prompt Guide
Weights in my prompts are primarily for testing the model's capabilities and trying to uncover limitations during test generation. However, I've created the tutorial below to help understand how model weights can be used more effectively.
Positive Attention
Weights between 0.5 & 2.0 are usually sufficient.
I sometimes use weights between 2.0 & 3.0 when I want to force attention to a specific ethnicity for testing.
Higher weights are also helpful when you want to magnify size or specific details. The difference between: large breasts, (large breasts:1.2), & (large breasts:1.8) is obvious.
Recognized weights are usually -5.0 to 5.0 with Dynamic Prompts.
Attention Equation
The parentheses ( ) on their own adds attention weight of 1.1. Unless you specify a weight.
(hyper-realistic) = 1.1
((hyper-realistic)) = 1.21
((hyper-realistic:1.3)) = 1.43
Brackets [ ] are negative attention weight of 0.91, not negative weight, it's just attention that starts under the 1.0 default of all words without weights. Stacking them decreases that number.
[deep shadows] = 0.91
[[deep shadows]] = 0.81
[[[deep shadows]]] = 0.72
Negative Attention
Weights over 3.5 are not usually effective or are a detriment during tokenization.
Negative weights do not have much effect in prompts.
Using too many high weights in the same prompt can cause distortions & deformities.
Using high weights will give specific keywords more attention, but that is often at the expense of other keywords.
Bonus
Prompt with high weights:
(HDR:2.2), cinematic, (natural sunlight:1.23), (hyper-realistic:1.45), (full body shot:0.6), masterpiece, incredible details, [feeling of solitude], mysterious, (deep shadows:0.72), Beautiful (30yo Malaysian woman:2.52), hazel hair, (mid_length grungy hairstyle :1.43), (natural large breasts:1.6), (wearing an unbuttoned stylish silk shirt), (brown pearl g-string, leather lace sandals:1.41), (sexy pose, reclining slightly on elbow:1.68) on a stunning designer wicker couch, at a rustic beach house rooftop with driftwood furniture, seashell decor, a view of the ocean, perfect for casual summer parties, sensual atmosphere, beautiful, guilty pleasure, shot on Canon EOS 5D, RAW.
Prompt without high weights:
(HDR:0.86), cinematic, (natural sunlight:0.8), (hyper-realistic), (full body shot:0.6), masterpiece, incredible details, feeling of solitude, mysterious, (deep shadows:0.72), Beautiful (30yo Malaysian woman:1.32), hazel hair, (mid_length grungy hairstyle), (natural large breasts), (wearing an unbuttoned stylish silk shirt), (brown pearl g-string, leather lace sandals), (sexy pose, reclining slightly on elbow:1.12) on a stunning designer wicker couch, at a rustic beach house rooftop with driftwood furniture, seashell decor, a view of the ocean, perfect for casual summer parties, sensual atmosphere, beautiful, guilty pleasure, shot on Canon EOS 5D, RAW.
To force clip & vae in FP32
Forge:
./webui.sh --clip-in-fp32 --vae-in-fp32 --no-half-vae(or add the commands to your webui-user.bat (Windows) or webui-user.sh (Linux) in
COMMANDLINE_ARGS= --clip-in-fp32 --vae-in-fp32 --no-half-vaeFor all commands (many are obsolete now since the developers have made them default)
Comfy:
python main.py --fp32-text-enc --fp32-vaeI don't use Comfy, so this is what I've seen on the web. Here is a link to all commands:
https://gist.github.com/ravetank/cedcd1ea89738b2e79b9f01e45a1a03d#file-comfyuihelp-txt
Description
Mixed precision
Settings I use:
DPM++ 2M SDE Huen Exponential
30-40 Steps
CFG 5-6
FAQ
Comments (14)
I'll have to try the V9 I have been having so much fun testing v8 faceSmash re-upload Hard to imagine anything better.
KissShot v9 - Amazing checkpoint! Great job. Quality, light, poses, atmosphere, details - everything looks very amazing.
I've tried the V9, it's good but went back to the V8 re-uploaded. I can't say enough for this model. It has actually taken away the desire to try and create my own. I love looking through images people post and use just the prompt from it and you can tell from my gallery the output. This model needs more recognition
If you don't mind. What makes v8 a better option for you? I develop new versions based on comments. Without feedback, I would create even more directionless monstrosities. Those are useful as seed models, but not official releases ;)
@ironically_irascible I find it follows prompts better as some prompts dont really capture the detail in there images, yours does. One example I have two Asian women in a pool of water, the original said two women but the image only displayed one woman. I come across a lot of that, plus the models used, if I keep regenerating from the same prompt I can the same person but with slight changes with back ground. It does animals really well. Im not good at creating my own prompts, in fact one word can destroy it. I dont like using Lora's, but two I do use which your models uses well are cameltoe which I cut down to 0.5 from 1, and puffy nipples. Things I cant get from it though is the word deepthroat. And some times when using a couple (man and woman) I add part of my own wording into the original prompt, it starts to add a penis to the woman. other than that, the women are beautiful, better that what I would compare to. I'm new to this and only learning as I go. I dont get tips from people and I used chatGPT to help me set stable diffusion up running offline (Webui) I dont know if ive helped much, I only started using SD 1.5 until I built my new system, now only use SDXL 1.0, to much work or understanding with all the others. I know you put a lot of hard work into this so I really mean it when I say thank you. I have tried many models and this one gives me everything I need for some time, plus I am learning with this model as well
Ok, good example here. Here is a prompt you used on your V9 model "expression immersive, (sexy 24 year old Bohemian girl), (rear view ass shot:1.4), medium teardrop-shaped breasts, a seductive boudoir, elegant room with a chaise lounge, a wooden desk, candles flickering on the walls, concupiscent atmosphere" I ran it on the V8 re-uploaded and posted for you to compare. They are both beautiful but because of the lighting and color, I feel mine looks more real. That is only my opinion but these are what I like about you V8 model compared to everyone else's models. More real looking
I use it with dmd2 lora and while v9 is awesome, it doesnt know clothes materials well(soft shine materials are vinyl look or simple matte).
One of the best I have ever used. Thank you!
This is perfect
I enjoy playing with your Re-bound model
So I decided to use your kiss model as a base model for something Im working on and so far Im really enjoying the results
I hope it serves you well. Considering the way I create my models, I wouldn't be surprised if many of them had latent unexpected surprises. :)
@ironically_irascible I found a tool that fixes clips in a model converter that actually works very well. But now you have this new surprise Thank you for going back
How did you come about your soulGrip model? I did a lot of crazy back and forth merging with your kiss model and images look close to your soulGrip model. I love how my model turned out. I'm going to try your soulGrip model tomorrow. I'm curious about your new model and what it can do, mine follows prompts crazy good. I asked about your model because I want to see if we have similar models merged. By looking at your images it looks like a strong model. Congrats on it by the way, I'm sure you're happy with it