NXP Semiconductors recently announced a collaboration with NVIDIA at GTC, enabling the deployment of NVIDIA-trained AI models across NXP's broad portfolio of edge processing devices using the eIQ® machine learning development environment.

This collaboration functionally integrates the NVIDIA TAO Toolkit into NXP's eIQ machine learning development environment, accelerating development amidst fierce competition in the AI industry. NXP is the first semiconductor company to directly integrate the NVIDIA TAO API into its AI-enabled products, enabling developers to more easily deploy trained AI models at the edge.

Simplifying the training and deployment of AI models is one of the biggest challenges facing AI developers today. To address this challenge, NXP collaborated with NVIDIA to integrate the NVIDIA Tao API directly into NXP's eIQ machine learning development environment. The NVIDIA Tao low-code AI framework makes it easier to leverage trained AI models and fine-tune them for specific use cases through transfer learning.

NXP's eIQ development environment facilitates the deployment of these models to the edge through a combination of software, an inference engine, a neural network compiler, and optimized libraries. This allows customers to accelerate AI development, access a library of pre-tested AI models, and deploy AI models across NXP's extensive portfolio of edge processors.

The NXP eIQ machine learning software development environment enables AI algorithms to be used across NXP's extensive portfolio of microcontrollers and microprocessors. It is fully integrated with NXP's MCUXpresso SDK and the Yocto Project Linux development environment, enabling developers to easily develop complete system-level applications.

The eIQ machine learning development environment directly integrates with the NVIDIA Tao API, providing a single platform for AI model training, optimization, and deployment. NVIDIA Tao provides an API-based workflow that allows users to build custom AI models, leveraging a collection of pre-trained models and transfer learning. NVIDIA Tao offers several optimizations, such as model pruning, that increase inference throughput.

This collaboration will accelerate development and enable NVIDIA's pre-trained models to run on the Neural Processing Unit (NPU) of NXP SoCs, such as the i.MX 93 applications processor.

“AI innovation will define the future of the smart, connected world,” said Charles Dachs, senior vice president and general manager of the IoT and industrial business unit at NXP. “NXP’s long-standing capabilities in industrial and IoT edge innovation combined with NVIDIA’s expertise in AI model training and testing creates synergies that will help customers get their AI models to market quickly and easily.”

“NVIDIA Tao dramatically simplifies the creation and deployment of AI models, including state-of-the-art generative AI models,” said Deepu Talla, vice president of Robotics and Edge Computing at NVIDIA. “Through this collaboration, we can seamlessly connect highly accurate, optimized AI models tuned by NVIDIA Tao to NXP’s eIQ development environment, accelerating edge AI deployments.”