Ryzen AI Software

AMD Ryzen™ AI Software provides tools and runtime libraries designed to optimize and deploy AI inference on AI-enabled PCs. It allows developers to efficiently port pre-trained models from frameworks like PyTorch or TensorFlow to run on integrated GPUs (iGPU) or Neural Processing Units (NPU) in select Ryzen AI-powered laptops. The software utilizes ONNX Runtime for seamless inferencing across various hardware components.

Developers can quickly begin building AI applications with an easy installation process, access to a growing library of pre-trained models, and optimized large language models (LLMs) using ONNX runtime. The software supports a range of AI workloads, including large language models, image and video generation, recommendation systems, and computer vision tasks.

Distribute and Offload Models Strategically

Ryzen AI Software leverages APU architecture to strategically offload models to the NPU and iGPU based on computational intensity and resource needs. The NPU is optimized for AI tasks, enhancing machine learning workloads such as image recognition, language processing, and real-time audio transcription. iGPUs are suitable for parallel computing and machine learning tasks like image processing and generation.

AMD collaborates with over 150 AI partners to develop innovative AI applications. Notable collaborations include Amuse's use of Ryzen AI NPU technology for super resolution solutions, Topaz Lab's development of AI editing software using AMD GPUs and CPUs, and Reincubate's Camo Studio, which enhances streaming and conferencing with AMD XDNA2-based NPUs. Arkrunr also utilizes Ryzen AI NPU for automatic background removal in creator apps.

The software is compatible with AMD Ryzen AI Max, Ryzen AI 300, Ryzen 8000, and Ryzen 7000 series processors. Developers can access the Ryzen AI Developer Hub for documentation, installation steps, and additional resources. Stay informed on the latest product releases and updates through Ryzen AI News.