Both projects are available on tensor.art with the possibility of online generation.

WALUAN – Simplicity and Quality Without Complications

WALUAN is a model developed without upscalers or Lora, delivering pure and authentic images at a standard resolution of 896x1024.

The focus is on simplicity—no unnecessary complexity or embellishments. For those who prefer straightforward and effective prompts, WALUANv3 offers solid results without relying on additional optimizations.

Specifications:

Sampling Method: Euler a
Steps: 30
CFG Scale: 7
VAE: flatpiece.xl.vae
Upscaler: x4 Remacri (which I usually use)
Negative Prompt: lazyneg

Want to enhance image quality? You can apply an upscaler for further refinement. And if you're feeling creative, experiment with different prompts to see how far WALUAN can go!

Note: I don't usually upload models that I merge. I believe there are great models here, but I enjoy mixing styles to create one that I like. I think this one turned out really well after many attempts, so I'm sharing it. I hope you like it!

Description

Patch Note - WALUANv3 Updated Version - Refinement and Expansion

Description: WALUANv3 is a highly refined model created through the fusion of multiple visual approaches to deliver a unique and versatile aesthetic. Utilizing Anime Screenshot Merge, Pixelnova - Fresh, WAI x 3D CGI, and DarkMix Mojito Illustrious - 3D Semi-Realistic, this model merges the best aspects of different styles to achieve enhanced depth and detailing.

Additionally, custom-modified Loras, personally adjusted and merged into the model, ensure a more tailored and sophisticated visual outcome.

Key Enhancements:

Advanced Model Fusion: Precise combination of different visual styles for greater authenticity and quality.
Customized and Modified Loras: Fine-tuned adjustments for a uniquely harmonious result.
Lighting and Shadow Improvements: Optimized tones and contrast for more natural and immersive visuals.
Semi-Realistic Aesthetic: A perfect balance between realism and artistic expression.

This model is designed to appeal to a wide audience, from anime enthusiasts to people like me who simply love creating AI-generated images.