yoso-normal-v1-5
| Property | Value |
|---|---|
| Author | Stable-X |
| License | Apache-2.0 |
| Pipeline | Image-to-Image (YOSONormalsPipeline) |
| Downloads | 595 |
What is yoso-normal-v1-5?
yoso-normal-v1-5 is a specialized diffusion model designed to generate high-quality normal maps from input images while reducing variance in the output. It's part of the StableNormal framework, offering sophisticated normal map generation with multiple processing modes for different scene types.
Implementation Details
The model is implemented using the Diffusers library and can be easily integrated using PyTorch. It supports various processing modes including object-specific processing with alpha channel masking, outdoor scene processing with automatic sky and plant masking via Mask2Former, and indoor scene processing.
- Supports multiple data types: object, outdoor, and indoor scenes
- Implements background masking capabilities
- Utilizes PyTorch for efficient processing
- Integrates with the Diffusers library
Core Capabilities
- Generates stable and sharp normal maps
- Automatic scene type handling
- Background masking for cleaner results
- Flexible integration options via PyTorch hub
Frequently Asked Questions
Q: What makes this model unique?
This model stands out for its specialized approach to normal map generation with reduced diffusion variance, offering multiple processing modes for different scene types and intelligent masking capabilities.
Q: What are the recommended use cases?
The model is ideal for generating normal maps in various scenarios including object rendering, outdoor scene processing, and indoor environment mapping. It's particularly useful when working with 3D graphics, game development, or architectural visualization where accurate surface normal information is crucial.
