Join Discord
✨ Try the Model for Free! ✨
I’m hosting my model on SocialDiff.net—a project I’m developing in my spare time. ADetailer is fully integrated, making your experience even smoother.
💖 Show Your Support! 💖
If you enjoy the model, don’t forget to leave a like and a review—your feedback helps improve the project! Also, support the talented creators publishing their work with this model by giving them some well-deserved appreciation.
🌟 Digging my way to the Illustrious Checkpoint. 🌟
Why Illustrious?
• Exceptional Attention to Detail: Every aspect is thoughtfully refined.
• Beautifully Detailed Bodies: Impressive proportions and structure.
• Perfectly Crafted Hands: No awkward poses or inconsistencies.
• Illustrious Vision for Semi-Realistic Results: A balance between realism and artistic finesse.
Key Settings
• Sampler: Euler a
• Steps: 25
Tags to Highlight Quality
masterpiece, best quality, amazing quality, very aesthetic, detailed eyes, perfect eyes, realistic eyes
Additional Enhancements
• All images are polished using ADetailer (face_yolov8s) for flawless face results.
Description
A brighter version of ilustmix, high attention to prompt
FAQ
Comments (44)
Can't wait to try out this cinematic one.
Excellent LoRA. Thank You.
In my opinion, Ilustmix peaks at version 5.5. Up until that point, it has a one-of-a-kind style, and gets generally better with each update. I was happy to see where this checkpoint would evolve, and up to 5.5, it really felt like it was evolving into something never before seen in AI art. However, starting with version 6, it shifts its style and starts to look like a couple of other models out there and begins to lose its unique style and flavor. I consider 6 and 7 a downgrade in look/style/quality/uniqueness.
Thank you for your review, Version 5.5 will continue it's branch on a separate (more realistic) version, had complains from people disliking the realism in this semi-realistic version, so stay tuned ;)
@GZees Thank you very much. I look forward to try the more realistic versions. Your checkpoint models are fantastic.
1000% agree. personally, version 5.5 is the king of all illustrious models
A+, will spend all my buzz again..
v7 cinematic has great shadows and lighting gradient! Refreshing!
Interesting range: CFG 3-9 all work nice.
First time ever I can make the checkpoints actualy do (cfg:8) what i have prompted in my dynamic prompts (WIP; without making the image "bad/oversaturated")! :D
Of course, there's still variations/mismatches (to prmpt), but generaly speaking, its SOOOO MUCH cloers (and better) than many, many, MANY other checkpoints!
Yes, CFG:3 is a bit more bland/lame, while CFG:8 or 9 are very intense, but (many) other checkpoints already "break" (unwatned fragments, floating limbs, etc) at 8.
I'm using version 7 without any LoRA. At first, I thought this model would only perform well with NSFW prompts, but after several tries, I found that it also handles regular images quite well. It follows prompts pretty effectively. Sometimes I run into issues with fingers or with objects that aren’t the main focus of the image.
As for NSFW images, I find them nearly perfect, even without using any LoRA, the output is still solid. If you’re looking for a checkpoint that works well for this type of content, I’d say this is the best one I’ve used so far.
What I love most about this model is its handling of lighting.
I’ll definitely test it more, especially with LoRA next time, to see how well they work together.
good
Nice
This seriously feels like black magic. Awesome model!
Excellent. Having a blast and loving the results.
This model is great I haven't seen another really comparable to what this spits out!
Am totally in love with V7.0 - it is generally 'brighter' than V6.0 and with different prompts used, I was able to achieve a more semi-realistic look or something abit more anime like. I had been alternating btw V5.5 and V6 previously (for achieving different looks/style) but with this latest V7.0 version, it seems to do both really nicely (depending on my prompt content).
As usually, a big thank you for all the hard work and effort put into your wonderfully awesome Checkpt! :)
Amazing checkpoint; recognizes most known characters, knows what they wear, and even who they are associated with. Easy to prompt, even with my limited skill. Does what you hope it might do, and does so pretty realistically. Thank you.
The skin is filled with too much light after 2nd ksampler. It loses the details. Any solutions?
Can this model be used commercially? As per the license it seems like it.
Hey, yeah, you can use it the way you like
Simply excellent :)
With V7 the results seem to have Way Too much light, compared to the previous version, any idea why? O.o
I concur. My images are blown out on white balance and oversaturated and very shiny. I wish I could get the look of v5.5 with the stability of v7
añoña
Negative prompts?
Just use the same from any IL checkpoint. worst quality, bad anatomy, low quality, big hairy fannny, sweaty bumcheeks
ok thanks
compiled with stable-fast it just generates black. shame because it looked cool.
Good!
which Illustrious version is this based on ?
Pretty good
This model can put out some absurdly good looking images, but I find it extremely difficult to get any emotion at all. Almost every image I generate is an emotionless facial expression, with the only one that consistently works is a generic smile.
I have tried single-word prompts and long-form sentence descriptions. None of these produce expressive faces that capture any emotion (beyond happy/smiling).
Anything related to the mouth other than smiling does not really work. Cannot get a detailed image of the inside of a mouth. Cannot get the eyes to do what I want them to do. The eyebrows barely move or show expression at all with any prompt I try.
That said, I've generated a LOT of good images that I do like, but I just cannot get the range of expressions that I want and have to use other models if I want something very specific.
Hey, give a prompt example please, and what would be your result, also check the negative prompt, there shouldn't be any problems with emotions :)
maybe, try using it alongside a LoRA dedicated to emotions you want
excellent model, really like it alot. working well
One of the best stable models I have use here
Did I see somewhere you said you would do a fork based on 5.5?. It's the best model on the site imo. Capable of amazing semi real and near realistic images.
yeah, if everything goes well more of it next week
Yeah, I've been waiting as well. I've tested out the newer versions and they're fine, but they're similar to other models on the site that I already have. I like ilustmix because they do something the others do not. Since this model has moved away from its original tone, I've been shifting into Reijil's models - they seem to have their own styles as well. But once the 5.5 fork starts up, I'll definitely be trying it out. It's definitely one of the top models on the site, IMO. I have tested over 150 models and 5.5 is an absolute must-have.
Best model for what I do
i love this model , wow amazing , all i needed is to learn how to prompt !
good
Can you help me which vae is work for this model..
You don't need any VAE. Check for description )
There's one baked in OR you can use https://civitai.com/models/140686/fix-fp16-errors-sdxl-lower-memory-use-sdxl-vae-fp16-fix-by-madebyollin.
It's simple really.
Details
Files
Available On (4 platforms)
Same model published on other platforms. May have additional downloads or version variants.