Keil Logo


Learning Platform

Machine learning (ML) technology is being increasingly adopted across the embedded sector in applications ranging from consumer to industrial IoT. All because of its unique potential for innovation. With the combination of popular neural network frameworks such as Caffe and TensorFlow, embedded-optimized neural network software stack for Cortex-M processors, and high performance cores such as the Arm Cortex-M7 processor, exciting new possibilities are opened for ML applications at the end node. This means that a wide range of neural network applications, like image and audio recognition, can now be applied to Cortex-M based processors with optimized performance and energy efficiency.

Jump start machine learning projects with CMSIS-NN on NXP i.MX RT


In this joint Arm and NXP webinar, you’ll learn how Arm NN and CMSIS-NN can help you develop efficient neural network applications for Cortex-M devices. Using a practical example, we show how the powerful i.MX RT processors can be used in conjunction with CMSIS-NN to run applications like keyword spotting. You will also learn how the on-chip Floating Point Unit (FPU) accelerates feature extraction from a live audio stream.

Blog: New CMSIS-NN Neural Network Kernels Boost Efficiency in Microcontrollers by ~5x

Neural Network Application Code diagram

Neural Networks are becoming increasingly popular in always-on IoT edge devices performing data analytics right at the source, reducing latency as well as energy consumption for data communication. CMSIS-NN is a collection of efficient neural network kernels developed to maximize the performance and minimize the memory footprint of neural networks on Arm Cortex-M processor cores targeted for intelligent IoT edge devices. Neural network inference based on CMSIS-NN kernels achieves 4.6X improvement in runtime/throughput and 4.9X improvement in energy efficiency.

Learn more >

Blog: How to Achieve High-Accuracy Keyword Spotting on Cortex-M Processors

It is possible to optimize neural network architectures to fit within the memory and compute constraints of microcontrollers – without sacrificing accuracy. We explain how, and explore the potential of depthwise separable convolutional neural networks for implementing keyword spotting on Cortex-M processors.

Keyword spotting neural network pipeline

Learn more >

Re-loadable CNN on the OpenMV Cam M7/H7 with CMSIS-NN


Soon, the OpenMV Cam M7/H7 supports loadable convolutional neural networks (CNN), based on CMSIS-NN. With loadable CNN support you'll be able to train a neural network on your PC using Caffe (or import the network from tensor flow into caffe) for things like robust presence detection.

  Arm logo
Important information

This site uses cookies to store information on your computer. By continuing to use our site, you consent to our cookies.

Change Settings

Privacy Policy Update

Arm’s Privacy Policy has been updated. By continuing to use our site, you consent to Arm’s Privacy Policy. Please review our Privacy Policy to learn more about our collection, use and transfers
of your data.