NXP is hoping to improve its machine learning offerings after making a strategic investment in Au-Zone Technologies. The exclusive arrangement specifically concerns Au-Zone’s DeepView ML Tool Suite, which will be used to bolster NXP’s eIQ Machine Learning software development environment and lead to the creation of new Edge machine learning products.
In that regard, the DeepView Suite comes with a graphical user interface (GUI) and workflows that will make it easier to import datasets, and to train neural network models for Edge devices. DeepView’s run-time inference engine will give eIQ developers more insight into system memory usage, data movement, and other performance metrics in real time, which will in turn allow them to optimize their model before deploying it in a System-on-Chip (SoC) solution.
“This partnership will accelerate the deployment of embedded Machine Learning features,” said Au-Zone CEO Brad Scott. “This will serve as a catalyst to deliver more advanced Machine Learning technologies and turnkey solutions as [Original Equipment Manufacturers] continue to transition inferencing to the Edge.”
In other news, NXP also revealed that it will be integrating Arm’s Ethos-U65 microNPU (neural processing unit) into its own i.MX applications processors. The Ethos-U65 is comparable to the Ethos-U55 in terms of power efficiency, but extends its utility to Cortex-A SoCs. The microNPU is compatible with the Cortex-M core featured in NXP’s i.MX SoCs (including the i.MX 8M Plus), and will allow NXP to expand its Industrial and IoT Edge portfolio.
“NXP’s scalable applications processors deliver a broad ecosystem for our customers to quickly deliver innovative systems,” added NXP SVP and Edge Processing GM Ron Martino. “Through these partnerships, our goal is to increase the efficiency of our processors while simultaneously increasing our customers’ productivity and reducing their time to market.”
Both Au-Zone and Arm have collaborated with NXP on other projects in the past. In the meantime, NXP has announced that Facebook’s Glow Neural Network compiler is now available through its own eIQ development environment. It has also released i.MX RT106F and i.MX RT106L MCUs to support the development of applications with face and voice recognition.
Credit: Google News