WD 1.5 Beta 3
| Property | Value |
|---|---|
| License | Fair AI Public License 1.0-SD |
| Downloads | 1,243 |
| Framework | Diffusers |
| Task | Text-to-Image |
What is wd-1-5-beta3?
WD 1.5 Beta 3 is an advanced text-to-image generation model that comes in five distinct variants: Base, Radiance, Ink, Mofu, and Illusion. Each variant is optimized for different aesthetic styles in anime-style image generation. The model utilizes the StableDiffusionPipeline architecture and includes safetensors implementation.
Implementation Details
The model employs the same VAE as WD 1.4 (specifically kl-f8-anime2.ckpt) and is built upon the successful foundation of previous WD versions. The Base model serves as a foundation for training, while the aesthetic models (Radiance, Ink, Mofu, and Illusion) are optimized for specific generation tasks.
- Implements StableDiffusionPipeline architecture
- Uses safetensors for model weight storage
- Compatible with Inference Endpoints
- Includes specialized VAE from WD 1.4
Core Capabilities
- High-quality anime-style image generation
- Multiple aesthetic variants for different artistic styles
- Fine-tuning compatibility with Base model
- Specialized image generation through aesthetic models
Frequently Asked Questions
Q: What makes this model unique?
This model's uniqueness lies in its multiple specialized variants, each optimized for different aesthetic styles in anime art generation. The Base model serves as a foundation for custom training, while the aesthetic models offer ready-to-use solutions for specific styles.
Q: What are the recommended use cases?
The Base model is recommended for training custom fine-tunes and LoRA models. For direct image generation, users should utilize one of the aesthetic models (Radiance, Ink, Mofu, or Illusion) based on their desired style output.
