Edit for 4.0
More realistic, less bright and saturated though. Still very creative. Not necessarily better than the previous versions, but it offers something different.
Edit for V3.5
After several rounds of internal merges, tuning, clip editing, many almost V3 versions, I've ended up further than I thought with what I call version 3.5. It's a general upgrade that fixes many of the issues of the previous V2 and V1 model, and so far I'm not sure what are its flaws. It's real fun to play with, the model is very creative, you often have to get out of its way and let it do its thing, although it will follow prompts closely.
Quite good at low inference steps.
Minimal prompting, minimal lora use seem to work best.
Please share your creations! I love to see them.
Edit for V2:
V2 is an overall, upgrade. It's more balanced, bolder, more detailed. Subjectively, it's better than V1, but you might think differently and that's legitimate.
Original post (V1):
This is an NSFW-focused model based on ten rounds of merges. I'd love to give credits, but at this point I've forgotten the contributors, sorry.
It'll do SFW, but it'll occasionally generate sexually explicit material on its own, even without a prompt.
The goal with this model was creating an Illustrious variant which would make great semi-realistic demi-humans, without a complex prompt, without a complex Lora stack, just the model, and you're good to go.
I'm personally pleased with the results, but more than anything, I'm curious about what people think, and how they'll merge it to make something new. If you do use this model for merges, let me know so I can see what you're up to. Just out of curiosity.
All the examples images I selected were created with relatively simple prompts, no loras, and in a few second per image. No complex workflows here. I'm 100% sure people will know how to have better results than me with it.
One last thing, the model is aimed for and biased to generating mature/adult characters. I cannot advise you strongly enough to not generate minor characters. Don't do anything illegal or amoral with it.
Description
More balanced than V7 but still not quite right in my opinion. Still. It's easy to get great results. Simple prompts tend to give better results in my experience, but that may depend on multiple factors.
FAQ
Comments (18)
V7 looks amazing but definitely overtly sexual, keeps making my blonde lightning girl ride a dude but there's nothing in the prompt for that. Also lost some concepts, my succubus no longer wraps her tongue around it, best I can get is a long tongue licking. (See my posts)
Absolutely. I was hesitant to post it but it would be weird to go from six to eight, and changing the version number was a bit strange too. For this one you actually have to go against the sexual aspect.
@_Kav_ Having seen V8 now.... yea should've just skipped V7 and made V8 V7, we never would've known the difference, LOL. Faces more reminiscent of V4 (not a bad thing IMO), overall sharper and less artifacts. Still overtly sexual and missing a few of the concepts of previous versions just like V7 but definitely a step above V7 for sure.
V8
Just amazing. I really enjoy this checkpoint. By far my favorite
Glad you're enjoying it!
Awesome checkpoint, but have nasty issues. Like (bloody claws:1.3), trigger generation of weird steel wrench-like parts around paws/hands and not a drop of blood. Can't create claw-paw. Can't do any claw, actually.
Interesting. 1.3 is pretty crazy weight for illustrious though. I'll try to create claws, haven't tried yet.
@_Kav_ really? I thought it is common practice. Actually overall weight value can't be more then 1.0 float. Overall weight divided between all triggers. So (1:3) weight trigger may never appear in output in some cases. =) Trigger words, that have lowest weight, rarely appear at output,.. if they interfere with, related to, high-weight "stuff". The more scales is on your skin - the more reptile you are. "Standing - mean only for those, who have legs to stand."
Try this one prompt:
nsfw, very detailed, fine details, realistic photo, sharp focus, professional lighting, 8k, HDR.
naked, (anthro female insect:1.3) with (round spider-abdomen-head:1.5), (chitin-scale lips:1.1), (animal-paw-like spider-legs and animal-claw-like spider-hands:1.0), (gray spider-skin:1.3), (gray spider-fur:1.3), (multiple spider eyes:1.5), standing sexy pose, looking to viewer, pleased and horny.
(burgundy background:1.3), blue front-light, full-body focus, front view.
@vasilypodvornyak988 I haven't really tried spider anthros, not really my thing. That said, I tried to create claws, bloody claws and didn't see what you mentioned, I generated claw-paws and everything. Not sure why you had this issue.
With V9, I get "decent" results with "sexy anthro spider female, multiple spider eyes, explicit, beautiful, fur, breast, grey skin, claw"
I'm not exactly sure what you were going for with your full prompt since you haven't shared a result you consider good, but running it on my end just made a spider like thing, nothing anthro about it.
You'd probably need a lora with a more specific spider design to get better results or to refine the model to better handle insect-based anthros.
@_Kav_ here an example. It is a Pony model. Not specific one. I didn't made same for Chaos, but in future. Probably prompt just needed to balance weights for that specific model. Checkout. https://civitai.com/images/119727447
P.S. Here a model i've used to create one. It is nor even Merge. Trained from ground. And it is far away from any kind of "Furry" or "Anthro" specific models. Yet specific, but not of this kind. =)) https://civitai.com/models/675024?modelVersionId=1204588
@vasilypodvornyak988 That makes sense, that prompt was weighted for a pony model for sure. 99% of pony images feel like they were taken in a studio, they don't feel candid at all, but on the plus side, the model is very precise and can do complex prompts with high weights. I've tried a pony version of Chaos but it was... vastly inferior in its understanding and execution of certain concepts I care about. That said I also worked a fraction of the time on it compared to Chaos on Illu. Anyway, interesting concept.
@_Kav_ i've experimented a bit with Chroma. Here is a result. https://civitai.com/images/120125104
General benefit of Illustrious, as i see - is a very stable seed. Changing prompt affect composition much more less than Pony and SDXL. But prompt is much more weak in compassion to Pony. Weighting prompt is much harder. That one image prompt is short, but it is already reach limit. Pity. Probably it is better way, dual model load with merge. Pony, for CLIP and Illustrious for MODEL. But this definitely lead to destabilization of seed.
Really great semireal checkpoint. I wasn't even looking for demihumans, I just love the results! I think I prefer the lighting from v6 more though, it's a little too saturated and glossy now
Thanks! For the saturation and gloss, it could partially be from the prompt. I optimized for very minimal prompting, so if you push the model further into gloss or color, it could give funky results. That said, this model is a bit extreme for sure.
Honestly, v6 seems to be the best version. In v7 and v8, characters default to having their mouths open all the time (often with messy teeth), and the skin texture just isn't as aesthetic as v6. The pose control and facial details also feel like a downgrade compared to v6. It feels like the direction went a bit off-track after v6... or maybe the model is just chasing some 'artistic' goal that I don't understand (I'm not exactly an art expert) 😂
I agree. "Improving" a model is complicated. It's all very subjective. In some respects, V7 and V8 give great results, but they are also a bit extreme.
The goal of the model on my end remains to generate semi-realistic humanoid females in erotic or NSFW situations, with minimal prompting and great visual quality. It's going to be able to do other things accidentally, but the focus is there.
@_Kav_ You're absolutely right—you can't please everyone when it comes to aesthetics. Personally, I love the fur texture in this model,I absolutely love the ear fluff rendering; it's hands down the best I've found on Civitai. Thanks for all your hard work, and good luck with your goals!