Stability AI and NVIDIA Accelerate Performance and Streamline Enterprise Deployment with Stable Diffusion 3.5 NIM

We’re thrilled to announce our collaboration with NVIDIA on the release of the Stable Diffusion 3.5 NIM microservice, designed to enhance image generation performance and streamline enterprise deployment. The SD3.5 NIM offers significant speed improvements, achieving 1.8x better performance over PyTorch, with NVIDIA H100 GPUs reaching optimized generation times of 3,700ms versus 6,800ms for standard versions. This microservice consolidates multiple models—including the SD3.5 Large, Depth, and Canny ControlNets—into a single, secure container, simplifying the deployment process for enterprises. Advanced users can now capitalize on faster iteration cycles, larger batch processing, and sophisticated multi-model workflows. The SD3.5 NIM is equipped to support Ada and Blackwell GPUs, making it an ideal solution for creative teams needing robust, efficient AI capabilities. This partnership unlocks new possibilities in enterprise image generation, ensuring rapid scaling in cloud deployments is more accessible than ever. Discover the future of AI-driven imagery with Stable Diffusion 3.5 NIM.

Source link