Latest updates:

(4/17/2025): illus01 v1.121.

• Rolled back to illustrious v0.1. illustrious v1.0 and above were trained with AI images deliberately (maybe 30% of its dataset or even more?). Which caused lots problems. Didn't notice until I read its paper.

• Lower character style effect. Back to v1.23 level.

• Other things just same as below (v1.113).

For advance users: if you only want to improve details, don't want to change character's style a single bit. Use "Touching Grass" LoRA. This LoRA has zero human in dataset, only objects and environments. Has much stronger effects.

Guys. I can't test everything before dropping a release. Remember to spam the comment section if weird things happen. And most importantly, remember to attach at least some images or settings. An image with metadata is worth than thousands words. Otherwise I have no idea what are you talking about...

(4/10/2025): illus11 v1.113.

• Trained on Illus v1.1.

• New dataset "Touching Grass" added. ~1K real world high quality photographs. Very diverse and creative. Same as previous, no human in dataset, just objects and environment. Old dataset was replaced.

• Full natural language captions from LLM.

• Bring back pixel level natural texture to overfitted models.

• Better lighting and depth of field effect. (Photography level)

• Better background structural stability. Less deformed background, like deformed rooms, buildings.

(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.

.........

See more in the update log section.

.........

(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.

Stabilizer

Just 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)

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.

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.

How to use

• Recommended strength: 0.4 for 2.5D/realistic models. 0.7 for pure flat colors anime models.

• No trigger words needed.

• You don't have to set the patch strength for text encoder. This LoRA does not patch it.

• Recommended -30% of your current CFG scale. This LoRA will increase contrast so you can lower CFG scale for more details.

Version prefix:

• illus01/11 = illustriousXL v0.1/v1.1

• nbep11 = NoobAI e-pred v1.1 (compatible with v-pred)

Older versions

New version == new stuffs and new attempt != better version for you base model.

You can check the "Update log" section to find old versions.

Dataset

Tldr: 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 (textures) 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 a sub dataset called "Touching Grass". Contains ~1K real world S-tier photographs. It contains nature, indoors, animals, buildings...many things, except humans. It adds more natural details and avoids overfitting. There is a "sub" LoRA that was only trained on this dataset. (Link above)

Full range colors

Implemented since 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 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 in illus v1.72.

(3/15/2025): illus v1.72

This 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.

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.

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 v0.4

Trained 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.

Above: Added more real world data. More natural texture and details. Can "fix" overfitted models.

Below: Pure anime style. Flat colors.

ani04 v0.1

Let'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 v1.23

nbep11 v0.138

Added some furry/non-human/other images to balance the dataset.

nbep11 v0.129

bad version, effect is too weak, just ignore it

nbep11 v0.114

Implemented "Full range colors". See the section above.

Added a little bit realistic data. More vivid details, illumination, less flat colors.

Above: High contrast. Wider color range.

Below: Normal anime contrast and colors.

illus v1.7

nbep11 v0.96

More 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.

Above: Has a weak default style.

Below: No default style, all data is equally weighted.

nbep11 v0.58

More images. Change the training parameters as close as to NoobAI base model.

illus v1.3

nbep11 v0.30

More 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.