NXP Semiconductors is to support the Open Neural Network Exchange (ONNX) format within its edge intelligence environment (eIQ).
NXP’s eIQ is a comprehensive machine learning (ML) toolkit that helps original equipment manufacturers (OEMs) balance performance needs and system cost when deploying neural networks and their associated inference engines at the edge.
ONNX is an open standard for representing deep learning models that enables trained models to be transferred between existing Artificial Intelligent (AI) frameworks.
By importing models in the ONNX format, NXP’s eIQ enables models to be trained in one framework and transferred to another for inference.
ML developers can then deploy inference engines across NXP’s scalable portfolio of MCUs, high-performance i.MX RT crossover processors, and highly-integrated i.MX and Layerscape applications processors.
“When it comes to choosing from among the many machine learning frameworks, we want our customers to have maximum flexibility and freedom,” said Markus Levy, head of the Artificial Intelligence Technology Center at NXP.
“An interoperable ML ecosystem is key to driving innovation, where designers can have the freedom to develop what’s needed for their applications. We’re happy to bring the ONNX benefits to our customer community of ML developers.”
ONNX, a community project created by Facebook, AWS, and Microsoft, is an open ecosystem for interchangeable AI models that provides a common way to represent neural network models.
ONNX models are currently supported in Caffe2, Microsoft Cognitive Toolkit, MXNet, PaddlePaddle, and PyTorch, and there are connectors for many other common frameworks and libraries. More information on ONNX can be found at https://onnx.ai/.