https://stability.ai/news/sd3-5-large-controlnets
Key Takeaways
Stability.ai has released three innovative ControlNets: Blur, Canny, and Depth for the Stable Diffusion 3.5 Large. Designed to enhance image precision and versatility, these models can be used across various applications like interior design and character creation. The Blur model supports high-resolution upscaling to 8K and 16K, whereas the Canny and Depth models use edge maps and depth maps, respectively, to structurally guide image generation. These models are available for both commercial and non-commercial applications under a permissive license, allowing free use for startups and small businesses with revenue under $1 million. Stability.ai emphasizes responsible AI use and is taking steps to ensure safe practices. More ControlNet variants and enhancements are expected in future releases.
Key Takeaways
- Stability.ai introduces new ControlNets: Blur, Canny, and Depth for Stable Diffusion 3.5 Large.
- These models are free for commercial and non-commercial use under the Stability AI Community License.
- Available for download from Hugging Face and code access via GitHub, with support from Comfy UI.
- ControlNets received the highest preference in a user study compared to similar models.
- Additional ControlNet models and features are planned for future release.
Stability.ai has released three innovative ControlNets: Blur, Canny, and Depth for the Stable Diffusion 3.5 Large. Designed to enhance image precision and versatility, these models can be used across various applications like interior design and character creation. The Blur model supports high-resolution upscaling to 8K and 16K, whereas the Canny and Depth models use edge maps and depth maps, respectively, to structurally guide image generation. These models are available for both commercial and non-commercial applications under a permissive license, allowing free use for startups and small businesses with revenue under $1 million. Stability.ai emphasizes responsible AI use and is taking steps to ensure safe practices. More ControlNet variants and enhancements are expected in future releases.
