Illustrious Renaissance v6_NS Update
First things first: all the preview images you see here are pure Text-to-Image (t2i). No LoRAs, no Hires.fix, and no Face Detailer were used. What you see is exactly what the model can do out of the box.
Sorry for the silence! This update took a bit longer than I expected, but it’s because v6_NS is a total game-changer for the model's future.
I’ve been spending a ton of time in the "lab" experimenting and training text encoders from various models. The goal? To bake in enough flexibility so you can use your favorite LoRAs without headaches.
What does this mean for you? The model is now a bit of a hybrid. I’ve tuned it to be compatible with almost everything: SDXL, Illustrious, and Pony. I’ve also started adding support for Noob—some things already work, and full compatibility is next on my list.
It’s a work in progress: Is it 100% perfect? Not yet. But it’s finally at a point where it plays nice with a huge variety of LoRAs, and I wanted to get it into your hands as soon as possible.
I need your eyes on this! Try it out, break it, and please share your feedback and images. I really need to know how it’s performing for you so I can keep polishing it in the next updates.
Enjoy the new versatility! I hope you like it. —DeViLDoNia
⭐ Illustrious Renaissance
Quick Start: For v5 and earlier versions
Recommended Sampler: euler, euler_ancestral, dpm++ 2m
Steps: 30+
CFG: 3.5–7 (5 recommended for most cases)
Clip Skip: 2
Hires Fix / FaceDetailer: Strongly recommended
Resolution: 1152x1536 / 1024x1536 / 832x1216 / 896x1152
Recommended Upscaler: 4x_NomosWebPhoto_RealPLKSR.pth
Upscale Factor: ×1.5–×4 (depending on your hardware)
Trigger Words: For v5 and earlier versions
realistic_skin_texture, realistic, realistic_detail, detailed_background
Positive Prompt Example:
masterpiece, best quality, YOUR PROMPT, realistic_skin_texture, realistic_detail, detailed_background
Negative Prompt:
worst quality, low quality
Enjoy and give me all your love: like, follow, review, post, it helps a lot.
~ DeViLDoNia
📧 Business Contact:
For commercial inquiries, collaborations or licensing, please contact:
📩 email: [email protected] / Discord: hades5826
Description
🧪 Key Beta Additions
Enhanced prompt diversity and better handling of complex scenes
Integration of Realism from HaDeS v9 for even finer textures
Improved rendering of sci-fi, futuristic fashion, cybernetic elements, and dystopian themes
Greater depth, contrast, and a more cinematic edge
🚧 Known Issues
Some instability in multi-subject prompts
Occasional over-saturation or flickering in complex lighting
NSFW generations can still be too eager under certain prompts — this will be fine-tuned
Hand generation is still inconsistent — this will be addressed and improved in the final release
Still under development, so please expect quirks here and there
FAQ
Comments (8)
Speechless! Work of art.
no great so far. an easy prompt show a huge bias, got one male out of 8 seeds, and lighting is way off (too bright). Another Illustrious realistic checkpoint (Sophos) does better ( 5/8, and darker light). All anime checkpoint I've tried respect the prompt on every seeds.
living_room, (solo:1.2), (dark:0.85), 1man, (plump:0.9), mature_male, cowboy_shot, wrinkled_skin, (dynamic_pose:0.9), cityscape, bookshelf, sidelighting, rug, wall_lamp, pensive, potted_plant, night, book_stack, (from_above:0.8), looking_outside, looking_away, indoors, face focus, green_eyes, loose tshirt, boyshort_panties, glasses,Refined your prompt slightly and got 5/5 on target — not testing further.
Prompt:
1boy, male focus, boyshort_panties, bookshelf, book_stack, cityscape, cowboy_shot, dark, dynamic_pose, face_focus, glasses, green_eyes, indoors, living_room, loose_tshirt, looking_away, looking_outside, mature_male, night, pensive, plump, potted_plant, rug, sidelighting, solo, wall_lamp, wrinkled_skin, from_above, masterpiece, best quality, realistic, absurdres, highly_detailed, ray_tracing
Negative prompt:
worst quality, lowres, bad anatomy, text, watermark, 3d_render, cartoon
Also, regarding the lighting — the model doesn't "lock" it. You can easily prompt the lighting you want.
For example, Danbooru-style lighting tags that work well include:
sidelighting (already in your prompt)
rim lighting
cinematic lighting
dramatic lighting
soft lighting
low key lighting
backlighting
volumetric lighting
light rays
moody lighting
dark shadows
ambient lighting
If it's too bright for your taste, you can also lower the global brightness with a slight negative emphasis on bright, or add dark, low key lighting, moody lighting to steer it where you want.
The lighting is highly steerable, not baked in.
@DeViLDoNia thanks for the reply. don't like the fact that 1boy works dans 1man does not, but hey.
ran some other tests and struggled for small_breasts :
- small is debatable, but that's expected
- can't use that without the woman going topless....
on the positive sign, got some nice strange spacecrafts and fog.
@moutonrebelle Yeah, makes sense — thanks for testing more.
Just to clarify: models based on Illustrious, Pony, or similar merges are very different from standard SDXL models.
They're trained or fine-tuned on tag-based datasets (mostly Danbooru-style), so prompts like 1man, adult_male, etc., are usually not recognized properly — because those tags simply don’t exist or weren’t learned.
That's why 1boy, mature male, older, or even solo, male focus tend to work much better.
Same for a lot of other concepts — you often need to "speak the tag language" for better control.
That said, glad you're getting some cool spacecrafts and fog! If you want more control over surreal elements or lighting, let me know — happy to share more tricks.
@DeViLDoNia I ran the same prompt on 3 others anime illustrious checkpoint and they all worked fine. check my profile, I use almost only illustrious for the past 6 month, I think I know quite well how it behave :)
I am not saying your checkpoint is bad, just like all realistic illustrious checkpoint, its harder to prompt than anime one, and deviate in the way it renders lighting. You can get around with weights, negative prompts and loras, not sure I want to bother.
@moutonrebelle Yeah, it’s true — Illustrious-based realistic models are a bit harder to prompt at first. That’s because they’re retrained almost from scratch on realism, starting from anime foundations. So initially, they don’t “understand” as many tags as native anime models.
That said, I often see a lot of anime prompts using tags that technically don’t exist — but the models are so well-trained and overloaded with learned concepts that the results still look great anyway.
With Illustrious models, though, if you stick to proper Danbooru-style tags (which is how they’re actually trained), you usually get much better results — more consistent subjects, cleaner lighting, better anatomy, etc.
Over time, as these realistic models keep evolving, they’ll become just as easy and forgiving as the anime ones — it’s just a matter of dataset growth and refinement.