Read the contents below before use.

ABOUT THE MODEL

Scarlet from Stellar Blade

20250805v1

Trained on IllustriousXL 2.0

All of the showcase images were generated using IllustriousXL v2.0 in combination with ADetailer.

Features

The output isn’t as good as the NAI version — there's a noticeable 3D stiffness in complex scenes. Because it uses E-predict, it doesn't handle lighting/shading as well as V-predict. Maybe downscaling the dataset to a semi-realistic style for retrain would help.（In fact, using a different checkpoint can already achieve this effect in image generation.）

However, its generalization and pose handling capabilities are slightly better.

• 1+ outfits.

Issues

• When the face becomes blurry, consider using ADetailer to fix it.

TRIGGER WORDS

All trigger words are based on “full body”, you can modify them as needed.

Example

You could find the example prompts from my showcase.

My prompts are basically composed in the order of [character traits] + [clothing] + [customize content] + [background], and you can delete or modify them as needed.

Default Outfit

sc213t, realistic, silver hair, long hair, bangs, sidelocks, large hat, gold hairpin, brown eyes, white leotard, shoulder armor, black shoulder belt, arm guards, armor skirt, white thigh boots ARCHIVES

USAGE GUIDE

Tips

I recommend using SDXL's original resolution first and then pairing it with ADetailer to fix the face. To achieve better visual results, I recommend setting Upscale value to at least 1.4x or higher.

Requirement

Checkpoints: IllustriousXL v2.0 or IllustriousXL base model

Recommended weight: 1~0.6

Hires.fix: Recommended Upscale value is around 1.2~2.0, denoising strength is 0.1~0.2.

Recommended size: 768*1344, 832*1216, 896*1156, 1024*1024, 1344*768, 1216*8

32, 1156*896

If blurriness appears on the face, please use ADetailer to fix it. The recommended denoising strength is 0.1~0.4

I’ve set the gallery moderation of this post to “R”, so explicit content may not appear on this page.