Enable true edge intelligence with the MAX78000 and MAX78002 ultra-low-power AI microcontrollers. The MAX78000 and MAX78002 are Arm^® M4F microcontrollers with a RISC-V^® coprocessor and a hardware-based convolutional neural network (CNN) accelerator. The CNN accelerator executes AI inferences using µJ of energy, providing a path for small, battery powered applications to participate in the current AI revolution. MAX78000 and MAX78002 are perfect for audio, vision, and time-series applications.

Both ultra-low-power AI microcontrollers utilize the same development tools. The main difference is the MAX78002 has a larger memory than the MAX78000, but most edge AI models can successfully be implemented on either part.

Getting Started

• Order a MAX78000 or MAX78002 evaluation kit from the list below
• Get the code from GitHub

Evaluation Boards and Development Platforms

ADI offers multiple evaluation boards and development platforms to fit your needs.

MAX78000FTHR

1. Overview - Adafruit^TM feather form factor for rapid development of MAX78000
2. Start Coding - Get the code from GitHub
3. Additional Resources - Product details

MAX78000EVKIT

1. Overview

• Full evaluation board with MAX78000 microcontroller
• Touch-enabled color TFT display
• MAX32625PICO debugger
• Olimex ARM-JTAG 20-10 adapter
• Cables

2. Start Coding - Get the code from GitHub
3. Additional Resources - Product details

MAX78002EVKIT

1. Overview

• Full evaluation board with MAX78002 microcontroller
• Touch-enabled color TFT display
• MAX32625PICO debugger
• Olimex ARM-JTAG 20-10 adapter
• Cables

2. Start Coding - Get the code from GitHub
3. Additional Resources - Product details

MAXREFDES178

1. Overview

• Cube camera reference design using MAX78000 and MAX32666 microcontrollers
• Reference design comes pre-programmed with audio and vision-based demos such as keyword spotting and face detection

2. Start Coding - Get the code from GitHub
3. Additional Resources - Product details

Developing for MAX7800x

Start developing with the MAX78000 and MAX78002 by reviewing the AI microcontroller resources on GitHub, then follow the steps below. A typical development workflow begins with training, followed by synthesis, and ends with deployment. However, if you already have a model and have synthesized your data, you can proceed to deployment.

Typical Development Flow

TRAINING - Are you looking to train custom models?

Start here if your goal is to train custom models. See the AI Model Training tutorial.

SYNTHESIS - Do you want to utilize pretrained models?

Start here if you want begin using pretrained models. See the AI Model Synthesis tutorial

DEPLOYMENT - Interested in embedded development?

Start here if you are interested in embedded development. Check out the SDK and examples:

• MSDK Solution Center
• MAX78000 example code
• MAX78002 example code

Microcontrollers Software Development Kit (MSDK)

The MSDK contains necessary software and tools to develop firmware for the MAX7800X AI microcontroller. It includes register files, sample code, peripheral drivers, system startup files, documentation, utilities, third-party libraries, IDE support files, and a toolchain.

• Learn more about the MSDK

Additional Resources

Articles and Application Notes

• Data Loader Design for MAX78000 Model Training
• Developing Power-optimized Applications on the MAX78000
• The Dream of Edge AI

Documentation

• MAX78000 Datasheet and Documentation
• MAX78002 Datasheet and Documentation

Videos and Tutorials

• Edge AI in Vision Based Applications
• Understanding Artificial Intelligence Series

Support

• ADI Support