Get started

If you are new to Arm AI, get started with our guides for Cortex-A devices and Cortex-M devices.

Cortex-A devices are in everything from Android Phones to single-board computers like the Raspberry Pi. Cortex-M chips power trillions of embedded devices around the world and run on extremely low power.

Here are our popular guides for new developers on Arm. Each guide introduces its board and some basic ML concepts. If you want to build a full-fledged application running on a modern OS, select the Cortex-A guide. If you are interested in low-powered ML devices for embedded or IoT, select the Cortex-M guide.

Application developers

Cortex-A machine learning

Use TensorFlow to train a Raspberry Pi to burst into applause whenever you raise your hands in the air. All we use is a camera and the on-board Arm CPU of the Raspberry Pi. This activity works on the Pi Zero.

View the guide

Embedded developers

Cortex-M machine learning

Learn how to use your gestures to train a classifier in TensorFlow and then deploy to an Arm Cortex-M-based board running Mbed OS.

Access workshop

Application development

Automatic Trash Classification with Raspberry Pi and Arm NN

Learn how to build an Arm NN-based application for an IoT device that performs automatic trash sorting through image analysis.

Read article

Dog Mood Detector with Google Cloud and Arm-based Raspberry Pi

Use Google AI Cloud to create a model for categorizing different dog sounds, then run the model on an Arm-based Raspberry Pi to listen for dog sounds and identify them.

Read article

Build end-to-end ML workflows with Arm

Use Arm Pelion Device Management, data management, and Mbed OS to build ML frameworks for IoT and endpoint applications.

Access workshop

Privacy-focused voice AI in intelligent robotics

Use a privacy-focused voice assistant, Snips, with MATRIX devices, an edge, and IoT development platform, and powered by Arm microcontrollers and Xilinx FPGAs, to deploy a voice-enabled robot with sensors, feedback loops, and motor control.

Access workshop

Application development: Arm NN

Run and profile Arm NN on the HiKey 960

Use Streamline to profile the performance of Linux applications that run TensorFlow models on Arm devices

View the guide

Configure the Arm NN SDK build environment for Caffe

Download, set up, build, and test Arm NN and its dependencies for Caffe

View the guide

Configure the Arm NN SDK build environment for ONNX

Download, set up, build, and test Arm NN and its dependencies for ONNX

View the guide

Configure the Arm NN SDK build environment for TensorFlow

Download, set up, build, and test Arm NN and its dependencies for TensorFlow

View the guide

Configure the Arm NN SDK build environment for TensorFlow Lite

Download, set up, build, and test Arm NN and its dependencies for TensorFlow Lite

View the guide

Deploy a Caffe MNIST model using the Arm NN SDK

Import a Caffe model into Arm NN, optimize and load it onto a compute device

View the guide

Improve your ML workflow using the Arm NN SDK

Walk through a typical ML workflow with a 32-bit floating-point Convolutional Neural Network

View the guide

Quantize neural networks to 8-bit using TensorFlow

Improve the performance of neural network models on Arm compute devices using TensorFlow

View the guide

Deploy a quantized TensorFlow Lite MobileNet V1 model

Develop a lightweight image classification mobile application using the Arm NN SDK

View the guide

Deploy a TensorFlow MNIST model using the Arm NN SDK

Import a TensorFlow model into Arm NN, optimize and load it onto a compute device

View the guide

Build Arm NN custom backend plugins

Write a custom backend for Arm NN, with an example custom backend to illustrate the process

View the guide

Implement a neural style transfer on Android with Arm NN APIs

Learn how to build a style transfer Android application with Arm NN APIs

View the guide

Implement classical machine learning with Arm CMSIS-DSP libraries

Learn how to use the new Support Vector Machine and Naive Bayes Classifier libraries for classification

View the guide

Arm NN accelerated object detection with Autoware

Use Autoware.AI and Arm NN to build and run a real-time object detection system.

View the guide

Deep-learning AI on low-power microcontrollers: MNIST handwriting

Train a TensorFlow model with MNIST and convert your model to TensorFlow Lite. Create the embedded application to generate sample MNIST data for embedding and testing the MNIST images.

Read article

Application development: teach your Raspberry Pi

Teach your Raspberry Pi - Episode 1 Yeah, World

Train your Raspberry Pi to burst into applause when you raise your hands in celebration

View the guide

Teach your Raspberry Pi - Episode 2 Multi-gesture Recognition

Train your network to recognize gestures in many situations and learn how to use larger datasets

View the guide

Application development: Machine learning on Raspberry Pi

Cross-compile Arm NN and TensorFlow for the Raspberry Pi 

Cross-compile Arm NN on an x86_64 system to work around the limited memory of the Raspberry Pi

View the guide

Run AlexNet on Raspberry Pi with Arm Compute Library

Develop Convolutional Neural Networks using just the Compute Library and a Raspberry Pi

View the guide

Profile AlexNet on Raspberry Pi and HiKey 960

Use Streamline to profile the performance of an AlexNet sample application on two hardware platforms

View the guide

Create a text-to-speech engine with Google Tesseract and Arm NN on Raspberry Pi live

Learn how to choose and convert an existing TensorFlow model to work with Arm NN and explore best practices for model conversion and implementing Arm NN solutions.

Read article

Accelerating ML inference on Raspberry Pi with PyArmNN

Learn how to use Python APIs for Arm NN inference engine to classify images as Fire versus Non-Fire.

Read article

Embedded development: Cortex-M machine learning

Build an Arm Cortex-M voice assistant with Google TensorFlow Lite

Perform machine learning inference on an Arm Cortex-M microcontroller with TensorFlow Lite for Microcontrollers

View the guide

Convert a neural network for Arm Cortex-M with CMSIS-NN

Convert a neural network from any framework into an implementation on an Arm Cortex-M device

View the guide

Perform image recognition on Arm Cortex-M with CMSIS-NN

Improve the performance and energy efficiency of real-time image recognition on an Arm Cortex-M7 processor

View the guide

Deploy a Caffe model on OpenMV using CMSIS-NN

Run a smile detection program on an Arm Cortex-M7 processor using a Caffe model

View the guide

TinyML Application Development for Everyone

Learn how to use your gestures to train classifier in TensorFlow and then deploy to an Arm Cortex-M-based board running Mbed OS.

View the Guide

Build your own Harry Potter wand with TensorFlow Lite Micro

This is an experiential workshop that focuses on the use of TensorFlow Lite on a low-power microcontroller to perform machine learning. 

Download guide

Optimizing neural networks for mobile and embedded devices with TensorFlow

Prepare TensorFlow models for deployment on Android, Linux, and iOS

View the guide

Deploying cloud-based ML for speech transcription

Set up client-server speech transcription deployed as a service running on cloud-hosted Arm servers

View the guide

Get Support


Community Forums

Answered How to make Ethos-U NPU work on an ARM Cortex-A + Cortex-M processor? 0 votes 339 views 4 replies Latest 2 days ago by alisonw Answer this
Suggested answer Compiling Arm Compute Lib for QNX OS 0 votes 426 views 3 replies Latest 7 days ago by Ben Clark Answer this
Answered How to try TensorFlow Lite on Ethos-N NPU? 0 votes 567 views 2 replies Latest 29 days ago by alisonw Answer this
Not answered Arm Compute Library has now an option to be used with GStreamer
  • Deep Learning
  • Neural Network
  • Machine Learning (ML)
  • TensorFlow
  • Arm Compute Library (ACL)
0 votes 212 views 0 replies Started 29 days ago by jchaves Answer this
Answered How to make Ethos-U NPU work on an ARM Cortex-A + Cortex-M processor? Latest 2 days ago by alisonw 4 replies 339 views
Suggested answer Compiling Arm Compute Lib for QNX OS Latest 7 days ago by Ben Clark 3 replies 426 views
Answered How to try TensorFlow Lite on Ethos-N NPU? Latest 29 days ago by alisonw 2 replies 567 views
Not answered Arm Compute Library has now an option to be used with GStreamer Started 29 days ago by jchaves 0 replies 212 views