| Property | Value |
|---|---|
| License | Fair AI Public License 1.0-SD |
| Base Model | Animagine XL 3.1 |
| Model Type | Text-to-Image SDXL |
| Developer | Linaqruf |
What is Kivotos XL 2.0?
Kivotos XL 2.0 represents the latest evolution in anime-style image generation, specifically optimized for creating high-quality artwork in the Blue Archive aesthetic. Built upon the robust foundation of Animagine XL 3.1, this model has undergone extensive fine-tuning to deliver exceptional anime-style outputs with particular attention to character detail and composition.
Implementation Details
The model employs a sophisticated training approach utilizing H100 and A100 GPUs, with carefully tuned hyperparameters including a batch size of 48 for fine-tuning and specialized noise offset settings of 0.0357. It implements the StableDiffusionXLPipeline architecture with support for multiple resolution outputs and optimized inference settings.
- Advanced prompt structure support with character-specific optimization
- Multiple supported resolutions from 640x1536 to 1024x1024
- Optimized sampling parameters with recommended CFG scale of 5-7
- Specialized aesthetic tags inherited from Animagine XL 3.1
Core Capabilities
- High-quality anime character generation with Blue Archive style focus
- Flexible resolution support for various aspect ratios
- Enhanced aesthetic control through specialized tags
- Optimized performance with Euler Ancestral sampling
Frequently Asked Questions
Q: What makes this model unique?
This model specifically excels in generating Blue Archive-style artwork with enhanced attention to character detail and composition, leveraging specialized training on anime aesthetics and advanced SDXL architecture.
Q: What are the recommended use cases?
The model is ideal for generating high-quality anime-style character illustrations, particularly those following the Blue Archive aesthetic. It's optimized for character portraits, upper body shots, and detailed character artwork with specific style consistency.