2Halfblood.safetensors

his is a highly specialized checkpoint merge, released as 2halfblood.safetensors, built on the NAI (v-pred) lineage. The model excels at producing high-quality Furry-Anime fusion artwork with a distinct, clean 2D to 2.5D aesthetic. The development priority was high prompt adherence and stylistic consistency across a wide range of subjects.
Key Capabilities & Style

• Fusion Style: Seamlessly blends the detailed rendering of Furry art with the clean lines and visual pop of modern Anime styling, optimal for Style Lora test.

• Versatility: Capable of generating complex, dynamic scenes with strong lighting (as demonstrated in the anthro character example) while remaining adaptable to simpler, cute, or chibi styles (as shown in the moth mascot example).

• Prompt Adherence: Shows robust capability in following complex and lengthy prompts without significant degradation of the base style.

• Color & Contrast: The v-prediction base allows for the creation of images with rich color depth and strong, defined contrast, particularly effective for dynamic lighting scenarios.

• Content Capability: The model is thoroughly tested and designed for robust generation of NSFW and SFW content.
  

Quick Start/Example Settings:
The following settings represent a balanced starting point for high-quality, stable generation:

• Steps: 28

• Sampler: Euler a

• Schedule type: Automatic

• CFG scale: 4

• Base Resolution: 1024x1024 (Hires. fix up-scaling is optional works great on native res)

• Clip skip: 2