Stabilizer illustriousXL/NoobAI

Latest updates:(4/10/2025): illus11 v1.113.Trained on Illus v1.1. Newer models should have already merged or based on illus v1.1 as it was downloadable many days ago.New dataset "Touching Grass" added. ~1K real world S-tier photographs, very diverse and creative.Also, now contains 90% (learning weight perspective) real world images. There is no real human in dataset. So 90% real world images will only add more texture details and "fix" overfitted 2D anime model.(4/10/2025): A short story.When I was testing v1.93 compatibility, I noticed the "overflowed" effect on one checkpoint by accident. I applied the LoRA with strength 0.4, but feel like it goes to > 1 and burned the output a little bit. Which indicates the LoRA was somehow applied twice. Yeap, it was merged in the checkpoint.The creator of the checkpoint does mention it is a "mix" model in the page, but other than that, no merge credits. So you never know which base models and LoRAs inside this "mix" model. Just something like "my model is so gooooood, please use it".I won't expose the usename of the creator I found, because I do not have the hard evident after all. The LoRA may be merged by other guy in other model and merged by this guy. Who knows.I just want to leave a message to the one who likes to grab others works and merge and wipe out all credits: F___ y__.(4/3/2025): New update plan:New added dataset/aspect will be trained as "sub" LoRAs. So if things go well, this dataset will join the Stabilizer. If you like raw and new things and like to balance weight yourself, make sure to check them out and leave feedback. I will add those new sub LoRAs to the collection so you can follow it to get notifications. FYI, you won't get notifications by just following this model page when I update the page info.(4/3/2025): New sub LoRA:"Touching Grass". This LoRA is trained on a new sub dataset contains ~1K real world S-tier photographs. It contains nature, indoors, animals, buildings...many things, except humans. Has much higher quality and diversity than what's currently in Stabilizer dataset. Also has natural captions from LLM. Depends on your usages. It can bring pixel level details. Can significantly improve background structural stability. Less deformed background, like deformed rooms, buildings. Or "fix" overtrained/overfitted models.Quick comparison on Wai v13. With/without.(3/30/2025): illus v1.93. v1.72 was trained too hard. So I reduced it overall strength. Should have better compatibility.(3/2/2025): You can find the REAL me at TensorArt now. I've reclaimed all my models that were duped and faked by other accounts.Can't believe the Creator program Civitai just came up with is ... Ponzi scheme, "pay me first, then you maybe earn something". Even I, who don't care much about monetizing, find this plan not only disappointing, but disgusting. Admire Civitai's confidence and attitude towards creators. Quoting another creator's comment: Wait a minute... So for all my contributions on the site... I still have to pay Civitai to get paid by Civitai?And, what is better than a pay-to-pay system? Pay-to-win!StabilizerJust a personal fun coding project. The name "stabilizer" is a bait now. There's no such definition of "stable". (The name used to mean "a better and stable default art style for vanilla NoobAI", back to NoobAI was released.)It can improve character's overall details, hands, backgrounds, natural lighting, brightness, contrast.....The advantage of doing this all-in-one thing is that you don't have to stack tons of different LoRAs for each of those aspects. So there is no conflict that might burn or blow up your output. (All conflicts in the dataset will be well averaged out to zero in one training)Also, this LoRA was trained in a way that will only modify final details. It will not dramatically change the base model output. So it's compatible with basically all models.Cover images might look mediocre or bad, because they are the direct outputs from the vanilla (not finetuned) base model in a1111-sd-webui with default settings, simplest prompts, no pipeline inpaint fixes, even no negative prompt. They demonstrate the effect of the LoRA, not clickbait.Recommended strength: 0.4~0.8.Version prefix:illus = illustriousXL v0.1nbep11 = NoobAI e-pred v1.1 (compatible with v-pred)Share merges using this LoRA is prohibited. FYI, there are hidden trigger words to print invisible watermark. It works well even if the merge strength is 0.05. I coded the watermark and detector myself. I don't want to use it, but I can.DatasetTldr: Only normal good looking things. No crazy art style. Not small (3k images, I won't brag it's big, there are many gigachads who like to finetune their models with millions of images).Every image is hand-picked by me. No AI images, which will cause overfit and wipe out all high frequency details instantly. No watermarks, links, etc...The main training dataset contains ~1k 2D images:Character-focus. (The characters take up most of the image. I also cropped and rotated many images if necessary.)Clear characters. (Clear character lines, hands, faces, eyes, no obstruction, no blur effect, etc.)Natural poses. Natural body proportions. (No exaggerated art, chibi, jojo pose, etc.)High quality and full of details. ~ Wallpaper level.And some sub datasets, ~2k auxiliary "real world no-human" images in total with a little bit everything, landscape, animals, buildings, rooms, weather, lighting... to balance the main dataset, add more natural details and avoid overfitting.Full range colorsImplemented after nbep11 v0.114.It will automatically balance the things towards "normal and good looking". Think of this as the "one-click photo auto enhance" button in most of photo editing tools.Pros: Big improvements in brightness, contrast and CFG stability. Now you can get high contrast and full color range (almost 0~255) at CFG 4~5 (v-pred is CFG 3~3.5, assuming you are using Euler and Normal Schedule) without oversaturation and deformed image.Cons: Normal == Not suitable for crazy bias effect. For example, you want 95% of the image is pure dark and only 5% is bright, instead of 50/50%.Update logNew version == new stuffs and new attempt != better version . A summary for big changes, if you are looking for older versions:now ~ illus v1.72 | nbep v0.160: More real world image in dataset. More details. Anti-overfitting/overtraining.~ illus v1.23 | nbep v0.114: Implemented "Full range colors". Better color, contrast.~ illus v1.7 | nbep v0.96: Finetuned again on ~100 wallpapers level images. Better and slightly more stable style.~ start: Pure anime dataset. Equal weight.Detail change log:(3/30/2025): illus v1.93.v1.72 was trained too hard. So I reduced it overall strength. Should have better compatibility.(3/22/2025): Noobai epred v0.160.Same stuffs below.(3/15/2025): illus v1.72This is a quite big update.Same new texture and lighting dataset as mentioned in ani40z v0.4 below. Brings RTX level natural lighting and super detailed texture to 2D characters. Tags that related to some kind of light sources, such as "lamp", "daylight", "soft lighting", may bring better lighting effect. Note: Those are not trained trigger words. You can try other things.A quick effect comparison on illus v0.1 / Wai v12 / Hassaku v2.1. Strength 0.8. Same sittings. Images below are directly from online generator, you can remix and test them at here: Wai and Hassaku.illus v0.1 w/wo | Wai v12 w/wo | Hassaku v2.1 w/woOk... images was compressed.. So here is a close-up comparison of Hassaku v2.1 w/wo.That's right, this is a Lora for the most realistic fur effect, beyond pixel level.Other small changes:Added a small ~100 images dataset for hand enhancement, focusing on hand(s) with different tasks, like holding a glass or cup or something. Can't say whether it's useful, as popular checkpoints already have similar enhancement. Anyways, the more, the better.Hand v1.72 vs v1.23 on good old illus v0.1:Removed all "simple background" images from dataset. -200 images.Switched training tool from kohya to onetrainer. Changed LoRA architecture to DoRA.(3/4/2025) ani40z v0.4Trained on ani40zero.The plan was to bring RTX natural lighting and super detailed texture to 2D characters.So I prepared a ~1k dataset focusing on natural dynamic lighting and real world texture.It was trained in a way that will only add small final details. So it won't blow up your output and create a realistic 3D character.ani04 v0.1Let's give Animagine 4.0 a try. Init version. In my test it fixed Animagine 4.0 brightness issues and brings much better contrast. Maybe too optimistic. Anyway this is a LoRA so you always can adjust the strength as you want. CFG 5 and LoRA strength 0.5 seems the best for most cases.illus v1.23nbep11 v0.138Added some furry/non-human/other images to balance the dataset.nbep11 v0.129: bad version, effect is too weak, just ignore itnbep11 v0.114Implemented "Full range colors". See the section above.Added a little bit realistic data. More vivid details, illumination, less flat colors.illus v1.7nbep11 v0.96More training images. Then finetuned again on a small "wallpaper" dataset (Real wallpapers, the highest quality I could find. Only ~100 images for now). More improvements in details (noticeable in skin, hair) and contrast.nbep11 v0.58More images. Change the training parameters as close as to NoobAI base model.illus v1.3nbep11 v0.30More images.nbep11 v0.11: Trained on NoobAI epsilon pred v1.1.Improved dataset tags. Improved LoRA structure and weight distribution. Should be more stable and have less impact on image composition.illus v1.1: Trained on illustriousXL v0.1.nbep10 v0.10: Trained on NoobAI epsilon pred v1.0.