Model Advantages:

1. The color tones have been optimized for better aesthetics, eliminating the yellowish retro feel.

2. Character designs are better suited to Chinese aesthetic preferences, featuring more appealing character modeling.

3. Better responsiveness to traditional Chinese style (Guofeng) themes.

Prompt Usage Instructions:

Trigger words: Unrestrained publicity/3D style

Simply add the scene description after the trigger words, and avoid using other style words in the description as much as possible. When describing characters, try using 1 woman/1 man to lean towards adult characters, and 1 girl/1 boy for teenagers. Try to specify quantities, otherwise a random number of characters or objects may be generated.

Scene descriptions should not be overly complex.

Recommended WebUI Settings:

Base Model Recommendation: 基础算法_F.1 (Base Algorithm_F.1)

Sampler: Euler

Sampling Steps: 25-30 (Try increasing the step count if the image appears blurry)

Weight: 0.8

Image Size: Recommended 10241536 / 15361024

If you generate a composition you really like but some details are flawed (such as hands, legs, or facial features), you can try applying Highres. fix to that single image using the same parameters and seed.

Upscaler Algorithm:

4x-UltraSharp, Denoising strength: 0.3-0.5, Steps: 20-30.