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.
The VAE is baked in. No loras were used for the sample images. Only Hi-res fix and Adetailer, on some images.
If you like it post your creations!
If you like what I do, I now have a link for tips. You can also use the link below to send me a private message or model requests:
General Series Details
Indigo Bold V12
Create crisp clear images with the ability to adjust color saturation levels by increasing CFG. It is best at hyper-realism and can handle fantasy images equally as well as photorealism. Indigo Bold is temperamental with settings. Different steps, CFG & samplers will drastically change the images that can be generated. Try the CFG at 6.0 to see what I mean.
Diversity of ethnicities
Specify an ethnicity, facial expression, eye color for more unique people
Very good at hairstyles
NSFW & SFW fully clothed images
Great at female anatomy
Very nice textures & patterns with or without prompting
Very good prompt adherence
Diverse vocabulary, try new keywords and see what happens
Short prompts will generate intricate details, as will long prompts
Prompting recognizes tags & plain text
v10 soulGrip ***
Previous versions were in an experimental phase of testing checkpoints. I've developed a much more stable process that keeps all of the qualities I want, but without the bloated file size. Version 10 is the best release so far.
Better details overall
NSFW & anatomy
Unique art style combinations
Creative photo styles
Dynamic posing & rich lighting styles
Unique architecture & background designs
As usual, ethnic diversity
v9 kissShot
kissShot comes from an unreasonably long line of custom seed checkpoints. The focus of development was:
NSFW & anatomy
Amateur photo & Social media
Creative posing & lighting styles
Unique architecture & background designs
And as always, ethnic diversity
While it is within the Reality Bound series, it isn't necessarily an "improved" version of the last release. I make each version as unique as I can, while keeping the same themes. There will be a large diversity of people who prefer different versions.
This is one of my goals. I get bored easily, so I want this series, as a whole, to excel at as many different things as possible.
v8 faceSmash re-upload
This version has the exact same Unet and VAE as the fp16 clip model. However, I'm taking advantage of this situation to test my new clip model. Clip-G is Long Clip Distilled, same as before. But Clip-L is the beginning of a highly experimental clip series that I have been working on for the past month. There are 6 variations that may end up in my future releases as I continue to refine & test.
v8 faceSmash
Since the Long Clip series seems to be popular I have created 2 versions of v8 with both fp16 & fp32 clips swapped with the dull built-in SDXL clips.
fp16 version is using the Simulacrum text encoders
fp32 version is using the Long Distilled text encoders
Simulacrum has very well trained encoders that perform noticeably better than the standard.
Very responsive to quality tags:
8k uhd, high quality, masterpiece, vivid | deep | vibrant colors, sharp details, cinematic composition, soft glow, creative shadows | lightingMy tests:
Eular a for artistic gens
Eular a SGMUnifom for detailed soft skin (soft: obvious goosebumps)
DPM++ SDE for extra crisp details (lower steps/CFG) (crispy: intricate skin details)
DPM++ 2s a for dramatic skin texture and details (Hyper-realism)
Using a different scheduler for generation and a different scheduler for Hi-res fix can give beautiful results. Try a crispy start with a soft finish.
Pushing my models to reach "overwhelmingly positive" status one day.
https://civarchive.com/models/1044804/clip-l-and-clip-g-full-fp32-zer0int-and-simulacrum
v7 gutPunch
v7 is an entirely different beast than what came before.
Female anatomy is indistinguishable from real life
Groups & multiple people are easier to gen
Improved photography & lighting variation, Vivid colors
Added cinematic composition, decades, styles & genres
Great Diversity: people, body, ethnicities, pose (lewd)
Very responsive to quality tags:
8k uhd, high quality, masterpiece, vivid | deep | vibrant colors, sharp details, cinematic composition, soft glow, creative shadows | lightingNatural language works best in short prompts
Danbooru tags work best in Long prompts
Gen settings do have a sweet spot depending on the image type. Test out different schedulers and other settings for best results.
My tests:
Eular a for artistic gens
Eular a SGMUnifom for touchable soft skin
DPM++ SDE for extra crisp details (lower steps/CFG)
DPM++ 2s a for dramatic skin texture and details (Hyper-realism)
Using different schedulers for generation and Hi-res fix can give beautiful results. Try a crispy start with a soft finish.
Both versions of v7 have the HDR VAE baked in.
https://civarchive.com/models/1480240/hdr-effect-vae
V4.0 fullybound
I don't like making grand definitive statements on things that are subject to diverse opinions. So I won't tell you this version will knock your socks off. That would be presumptive.
Strong points include:
Creativity is through the roof
Background details are amazing
Realism & Hyper-realism is excellent, semi-realism is also easy to generate
Diversity is still very strong
It can handle many different styles and themes
Dynamic poses are especially dynamic
Clothing prompts work very well & very detailed
Expressions have really improved
Good skin details, eyes are better than the previous versions
Prompt adherence is good for simple prompts and sfw prompts, but may need a neg prompt
Bondage prompts...you'll see, it will not disappoint ;)
Weak points:
Some sex actions are hard to prompt for if given a lot of specifics
Fantasy creatures are ok, however fairies are very good
Hands are pretty good, but still has typical XL behavior
Multi-person images are better than most XL models, but the hands become harder to get right
Prompt adherence can go from excellent to "where did that come from?"
V3.0
Strong points include:
Impressive ethnic diversity
NSFW images are easy to prompt without using a lora in the prompt
Realistic anatomy (woman focused, but men too)
Beautiful women, handsome men, both with a very wide range of features
Prompt adherence is good for long prompts, and great for short prompts
Stronger ability to generate dynamic and artistic images
Photography: art, glamour, erotic, pornographic, social media, professional, amateur
Lighting: creative, dynamic, stylized
Poses: dynamic, sexually explicit, casual, unique, action
Better at handling Danbooru tags
Uncensored and very horny (post images responsibly, please!)
Weak points:
Some lewd acts are hard to prompt for but will then appear unprompted at times
Difficult to keep her clothes on...but it is possible with a detailed description of clothes
Distant faces need aDetailer
Eyes usually need to be described in the prompt or they default to dull or ugly
Hi-res fix is essential for complicated imagery (pretty standard, yes?)
ribaldVassalage V2.0
Strong points include:
Impressive ethnic diversity
NSFW images are easy to prompt without using a lora in the prompt
Realistic anatomy (woman focused, but men too)
Beautiful women, handsome men, both with a very wide range of features
Prompt adherence is good for long prompts, and great for short prompts
Stronger ability to generate social media-style selfies to epic cinematic images
Photography: art, glamour, erotic, pornographic, social media, professional, amateur
Lighting: creative, dynamic, stylized
Poses: dynamic, sexually explicit, casual, unique, action
Better at handling Danbooru tags
Uncensored and very horny (post images responsibly, please!)
Weak points:
Some lewd acts are hard to prompt for but will then appear unprompted at times
Difficult to keep her clothes on...but it is possible with a negative prompt and a detailed description of clothes in positive prompt
Cloned faces in a crowd
Distant faces need aDetailer
Eyes usually need to be described in the prompt or they default to dull or ugly
Hi-res fix is essential for complicated imagery (pretty standard, yes?)
V1.0 / V1.1
V1.1 is the reduced size of version 1.0. Re-merged with the original base model (16GB) using several methods. I managed to get it practically identical to V1.0. Only minor aesthetic variation in seed reproducibility.
Strong points include:
Impressive ethnic diversity
NSFW images are easy to prompt without a lora
Realistic anatomy (woman focused, but men too)
Beautiful women
Prompt adherence is great for long and short prompts
Ability to generate influencer-style selfies to epic cinematic images
Fine art and glamour/erotic/amateur photography
Uncensored (generate responsibly, please!)
Weak points:
Some lewd acts are hard to prompt for but will then appear unprompted at times
Cloned faces in a crowd
Eyes usually need to be described in the prompt or they default to dull or ugly
V1.0 InvokeAI reads as a lora network, rather than a base model (fixed in v1.1)
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
VAE is baked in for both versionsThese are the settings I use for this version.
faceSmash
Eular a SGM Uniform
30 steps
CFG: 5-6
Hi-res Fix
4x Foolhardy Remacri ExtraSmoother
8-15 steps
CFG: 5
Denoise: .4 to .5
V8 fp16
Internal encoders have been replaced with Simulacrum clip-L & clip-G and converted to fp16
https://civitai.com/models/1044804/clip-l-and-clip-g-full-fp32-zer0int-and-simulacrum
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-vaeComfy:
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
FAQ
Comments (8)
(Full Copy) Prompts do not reproduce images
I have my Forge set to produce seeds with CPU, not the default GPU. However, I've never seen the need to reproduce images exactly. So I don't keep a record of seeds used before Adetailer changes them (when I need to use it). Although, if the prompt itself gives wildly different results then that is something I need to investigate.
@ironically_irascible
That do not explain that the images do not even closely resemble the examples. Reproducing prompts is important to validate that ones setup works correctly. If you can't reproduce the prompt and the image quality is 1/10th of the example, then there is no way to figure out whats causing it as there are endless amounts of parameters.
But if you don't assure the prompt reproduces (doesnt need be pixelperfect but 98% to the example with same level of quality), you can't start making stuff with it as the quality is all over toe board then
@TScitai What version are you using? The uploaded models are the same versions I used to make the images. I use them for at least a week after uploading. So if I download the version you're talking about and get different images entirely, then I need to contact Civitai. I can also replace the model with the original.
Also, CPU seeds will produce very different images than GPU across different computers. From the Forge Settings:
"Random number generator source. (changes seeds drastically; use CPU to produce the same picture across different videocard vendors; use NV to produce same picture as on NVidia videocards)"
@ironically_irascible In general the generating UI should put all parameters into the image itself, inclusing CPU/GPU/NVidia GPU seeding, VAE etc. Im using different UIs, mostly Automatic1111 due to the number of working extensions for it, ComfyUI for some special workflows and sdnext.
Its sadly quite a common occurance on checkpoints that prompts do not reproduce the image, which is really a pain in the ass as it often results in greatly different image quality
@TScitai I agree with the last sentence, for sure. I ignore seeds for the most part. But the quality is required. Which is why I tested both versions of v8 and concluded that fp32 model wasn't what I thought I was uploading. I can't figure out what changed from the version I tested and the version uploaded.
I very much appreciate you pointing the discrepancy out to me. I never would have discovered it myself since it didn't occur to me that the model could deviate from the tested version. CPU vs GPU does change seeds "drastically", but it certainly can't explain why the preview images were so different. It was probably an error on my part, rather than something changing after uploading. So I am definitely going to be more attentive than I was before.
V2 has been the best realistic XL model on all of Civ for a while now, and I'd say V8 is better, but only slightly.
It's "sensitive", so you're going to want to keep the CFG a little lower, probably at 4 or less. It also tries to put pussy's on everythying. EVERYTHING. Like I've had pussy's show up on a dude's thigh...
Speaking of dudes, it can struggle with male anatomy in more complex scenes. It doesn't do gay scenes very consistently, will try to make everyone a futa or some abomination human centipede.
It doesn't do distance well (no models do really) but close ups of faces are amazingly well rendered.
It also skews young so, be careful...
Use terms like "adult woman" not "girl" or you'll look like a creep.
With all three editions of the V8, I am getting the infamous "sameface" issue. It really likes generating just one female face; there is simply no variety. Is there some Poiny or Illustrious in it?
To those experiencing the "cooked" or "burned" look: deducting DMD2 lora at 0.2 weight fixes that issue (but not the "sameface" issue).
Looks like we don't have an active mirror for this file right now.
CivArchive is a community-maintained index — we catalog mirrors that volunteers upload to HuggingFace, torrents, and other public hosts. Looks like no one has uploaded a copy of this file yet.
Some files do get recovered over time through contributions. If you're looking for this one, feel free to ask in Discord, or help preserve it if you have a copy.