FuzzyHazel
| Property | Value |
|---|---|
| License | CreativeML OpenRAIL-M |
| Model Type | Text-to-Image Diffusion |
| Downloads | 4,191 |
What is FuzzyHazel?
FuzzyHazel is a sophisticated text-to-image generation model that combines multiple base models and LoRA adaptations to create a unique artistic style. It's built using a complex merging formula that includes HazyAbyss, OctaFuzz, MareAcernis, and other foundational models.
Implementation Details
The model uses a specific configuration with DPM++ 2M Karras sampler, operating at 28 steps with a CFG scale of 7. It supports high-resolution image generation with a 1.5x upscale factor and includes specialized negative prompting for quality control.
- Built using MBW (Multiple Base Weight) merging technique
- Incorporates multiple LoRA adaptations including Guardian Tales, Komowata Haruka, and Terada Tera styles
- Optimized for 768x512 base resolution with upscaling capabilities
Core Capabilities
- High-quality image generation with fuzzy artistic style
- Efficient 28-step sampling process
- Built-in upscaling support with Latent upscaler
- Compatible with EasyNegative and pastelmix-lora for enhanced results
Frequently Asked Questions
Q: What makes this model unique?
FuzzyHazel's uniqueness comes from its carefully crafted merger of multiple base models and LoRA adaptations, resulting in a distinctive fuzzy artistic style while maintaining high image quality.
Q: What are the recommended use cases?
The model excels at creating artistic images with a fuzzy, stylized appearance. It's particularly well-suited for generating anime-style artwork and illustrations with its integrated chibi and artistic style LoRAs.
