This sample shows Edvice integration into the Microsoft Azure Sphere real-time core (cortex-m4) are able to run together on MT3620 Development Kit. This sample runs edvice inference model on Azure RTOS
To clone this repository, you can use this command.
git clone --recursive [email protected]:ENERZAi/ZaiOptimizer.gitPlease go to this page and follow the instructions to setup the Azure sphere SDK and device before running this sample Azure Sphere Quickstart
- Under the AzSphereEdvice repository, create lib_arm folder
- Place libModel.a file downloaded from edvice website under AzSphereEdvice/lib_arm directory (Note : Model is required to be trained in soft float-abi mode)
- Place Edvice.h file downloaded from edvice website under AzSphereEdvice/Includes directory (If it already exists, replace it)
- Open Visual Studio with Azure Sphere SDK installed and run CMake
- Build and run the example on your Azure Sphere device
Note: You need Azure Sphere SDK version >= 20.07 to build and run the sample.
BMI160 SCL <-> Azure Sphere SCL BMI160 SDA <-> Azure Sphere SDA BMI160 Vin (+) <-> Azure Sphere 5v BMI160 GND (-) <-> Azure Sphere GND
User LED on Azure Sphere will give you the inference result or indicate error state
- Blue -> Abnormal
- Green -> Normal
- Red -> Error (Most likely to be caused by sensor malfunction. Try reconnecting the sensor or reboot the Azure sphere)
To prepare your hardware to display output from the sample:
- Attach the USB-to-serial adapter to your PC.
- Open serial monitor (Putty in case of windows) and choose the right COM port that your device is conected to
- Serial monitor will display state to the user -> It will print inference result as "Normal" or "Abnormal"
Project Edvice belongs to Enerzai corporation (New tab (enerzai.com))
This project is based on this repository
([email protected]:Azure-Samples/Azure-RTOS-on-Azure-Sphere-Mediatek-MT3620.git)