Machine Learning Compute for the IoT and Embedded Market
This is the best thing, I read online today!
Main points about the article are:
- Machine learning (ML) technology is expanding rapidly to impact all markets, including the deeply embedded space.
- It is expected that AI-enabled IoT shipments will grow at a rapid rate in the coming years.
- Arm provides an unmatched solution to problems by accelerating ML workloads on existing Arm Cortex-M based systems
- Billions of embedded devices that use Arm Cortex-M processors already exist in the world and can easily accelerate ML workloads through the use of optimized libraries provided within CMSIS-NN.
- CMSIS-NN is open source and newly optimized libraries are being added at each quarterly release.
To increase the ML performance of these systems, Arm has announced new technologies:
- Cortex-M55 processor
- Ethos-U55 microNPU
A neural network can be efficiently accelerated in an extremely small area and power envelope using the following:
- TensorFlow Lite micro
- NN Optimizer Tool
- Ethos-U55
- Cortex-M55 and CMSIS-NN
Summary:
The Cortex-M55 processor, Ethos-U55 microNPU, and Arm’s industry-leading embedded ecosystem of software libraries and tools support will bring AI to the billions, removing barriers to ML adoption and deployment.
These processors will securely and efficiently increase ML and signal processing performance for the next generation of world-changing IoT and embedded devices.
They extend the performance of Arm’s AI platform for microcontroller-based endpoint devices, offering silicon providers a more diverse range of hardware choices and empowering developers to deliver this next revolution in computing.
For Further Information Please find below links:
