Machine Learning Resources for IoT and Embedded Developers
Get essential building blocks to streamline your ML workflows, and gain valuable insights from ML developers as they share their development journey and lessons learned. This curated list helps you build, train, test, and deploy ML at the edge.
Existing ML Projects
Explore AI and ML projects and case studies that solve real-life problems.
Object Detection with Grovety | 3 min | Case study sharing the architecture, components, neural networks, and tools chosen for AI-powered trail camera. |
Natural Language Processing with Sensory | 3 min | Case study showing how Sensory developed a voice assistant with natural language processing in a memory- and power-constrained consumer device. |
Neural Network Biometric Models with TinyMLOps | 25 min | Video showing how Fortifyedge rapidly prototyped with its Tensorflow models deployed on Google Android and Wear OS to run ML workloads on device. |
Setting a Wake Word Application | 4 min | Video on how to create a wake word application on the Raspberry Pi Pico. |
Accelerate Inference and Optimize Performance with Arcturus and NXP | 28 min | Video on accelerating AI IoT development, with a demo of how ML DevOps can simplify deployment. |
Building an IoT-enabled Artificial Nose Using TinyML | 32 min | Video about the journey to building an open source and open hardware DIY artificial nose, plus TinyML best practices. |
AI and ML Development Solutions
Gain valuable insights and explore hardware, software, tools, and frameworks to streamline your ML workflows.
End-to-End Neural Network Model Development Tutorial | 10 min | Video to learn how to train and deploy a neural network from a Jupyter notebook using Google Colab. |
Adventures in Debugging | 30 min | Video exploring troubleshooting and debugging tips, including GDB and python, signal probing techniques, and tracing. |
Securing Deployment of AI to Constrained Devices | 28 min | Video on using open source and open standards to secure AI models during deployment. |
Going Cloudless: AI at the Edge | 58 min | Video exploring the potential of edge AI and sharing technologies for cloudless edge AI solutions for vision, object detection, and vital signs prediction. |
Easy TinyML: Practical Examples to Get Started Today | 1 hr | Video exploring the potential of edge AI and sharing technologies for cloudless edge AI solutions for vision, object detection, and vital signs prediction. |
Do You Have the Right IP and Tools?
Access resources and step-by-step guidance to take your ML projects from inception to reality.
Arm Keil MDK
Development solution for Arm-based MCUs, including components needed to create, build, and debug embedded applications.
Vela Compiler
Open-source Python tool to optimize a neural network model that can run on an embedded system containing an Ethos-U NPU.
Machine Learning Inference Advisor
Optimize neural network models for efficient inference on Arm that provides performance analysis and actionable advice.
Usage Guides and Support
ML Integration Guide: Arm Ethos and Arm Cortex-M | 1 hr | Guide for and Cortex-M processors, tools to support software development, and the available evaluation paths. |
Bringing ML and DSP to Constrained Devices with Arm Helium Technology | 1 hr | Video on how Helium technology can bring ML and DSP capabilities to constrained devices with a demo of the tools and libraries to target Helium. |
Navigate Machine Learning with Arm Ethos-U NPUs | 10 min | Learning Path to help you get started with Cortex-M and Ethos-U ML application development. |
Build and Run Arm ML Evaluation Kit Examples | 30 min | Learning Path to build examples from machine learning evaluation kit and run the examples on Corstone-300 FVP or Arm Virtual Hardware (AVH) |
Build and Run a Letter Recognition NN Model using TensorFlow | 1 hr | Learning Path to build a letter recognition neural network model using TensorFlow and run the model on an STM32L4 board. |
Build and Run an Image Classification NN Model | 1 hr | Learning Path to build a convolution neural network model for image classification and run the model on an STM32L4 board. |
Deploy PaddlePaddle on Arm Cortex-M with Arm Virtual Hardware | 30 min | Learning Path showing how to export the Paddle inference model, compile the model with TVMC, and deploy on AVH. |
A Community to Build Your Future on Arm
Join the Arm Developer Program to build your future on Arm. Get fresh insights directly from Arm experts,
connect with like-minded peers for advice, or build on your expertise and become an Arm Ambassador.