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
Settings I use:
DPM++ SDE or DPM++ 2M karras
30-40 Steps
CFG 5-6
I'm taking the series back to the typical 7GB model size; with the original text encoders.
FAQ
Comments (11)
Im happy to hear about his new model, trying it out, soon
It follows prompts well like the other versions. The one thing that stands out to me that it does well is backgrounds. People on the streets and builds in the background still look like part of the image and not turning to blurry blobs which is a big plus for me.
Glad to see, that you're alive @ironically_irascible even only seeing you in here and not at your other place.
Reality BoundXL is 98th on the list for featured models. Thank you to all that contributed yellow buzz! I will always use it for new and better projects.
A brief note on the versions. V4, V5, V6 are high-quality realistic models. V7 and especially V8 are flops with severe "sameface" issues and Pony-like anatomy. The V9 is borderline; it has less face variety than V4-V6 and more of a sameface. V10 is back to sameface/Pony face again, but without the anatomy issues of the V8.
Please don't take this review as an absolute; different faces can still be generated with careful prompting. It's just that the later versions of the model tend to stick to the same Pony face stronger than the earlier ones.
Same face is a big problem with most XL models that use merging primarily as the creation method. I try to add variety through lora manipulation during the merge process. Which means very specific keywords are usually needed to trigger a unique face. Which is why I always use ethnic keywords specifically, expression, and adjectives related to beauty or looks.
I find your assessment very valuable. Thank you for testing so many of my models, I know that is a long process. I will look into ways to get rid of "default face" in my future models; in a more targeted process. For each version I release, at least 30 previous models are created leading up to the final version I like (so many extra models saved on SSD for potential merging later on).
My top focus is creativity in prompt interpretation, diversity of ethnic variation, realistic anatomy, and understanding of less common concepts. I admit that I sometimes sacrifice 1 strength when I see an abnormally high strength in one of the other categories. I will take the challenge you indirectly gave me for the next release ;)
THANKS! if you had to choose just 1, wich one would you pick?
Love this model! Hope v10 is even better, but that will be difficult!
While v10 is quite nice, it seems to be just a bit overfitting for an*l sex.
It's a hard balance to strike while merging. With XL, it's either really hard to get a concept to work, or way too easy so it overshadows other concepts. Like 80% of my prompts for anal sex are ignored or very difficult to get right on every XL model I've tried. Either way, I hope v10 is useful, or at least fun, to use.
Getting used to it, thanks, lots of creative negative prompts. But I like it, very nice! Thank you for your efforts.