| Property | Value |
|---|---|
| License | CreativeML OpenRAIL-M |
| Type | Text-to-Image Model |
| Framework | Stable Diffusion |
| Format | Safetensors |
What is lametta?
Lametta is a specialized Stable Diffusion model designed for generating anime-style characters with distinct characteristics. The model excels at creating female characters with round, large eyes and shorter body proportions, with particular attention paid to eye highlights and finger details. It's optimized for producing consistent, high-quality anime-style illustrations with a unique artistic style.
Implementation Details
The model is built through careful merging of multiple base models and has gone through several iterations (v13xx to v20xx series). It uses various VAE implementations, with recommendations for ClearVAE series or Anything's VAE for optimal results. The latest version (v2012) includes sophisticated merging techniques incorporating multiple model bases to achieve its distinctive style.
- Optimized for DPM++ 2M Karras sampler
- Supports various aspect ratios with best results in portrait orientation
- Includes built-in VAE with option for external VAE usage
- Specialized in female character generation
Core Capabilities
- Exceptional eye detail rendering with preserved highlights
- Consistent anime-style character generation
- Strong performance with shorter prompts
- Specialized in female character illustrations
- Effective handling of various lighting and background scenarios
Frequently Asked Questions
Q: What makes this model unique?
Lametta's unique strength lies in its ability to consistently generate round-eyed anime characters with detailed eye highlights and specific body proportions. It's particularly effective with minimal prompting and maintains quality across various scene types.
Q: What are the recommended use cases?
The model is best suited for generating anime-style female characters, particularly those with large, round eyes. It excels in both simple character portraits and more complex scenes, though it's specifically optimized for female character generation and may not perform as well with other subjects.