This repo contains different ML models that can be built to run on the Alif Beta AI/ML AppKit development board. It is based on the ARM ML embedded evaluation Kit.
For instructions on setting up the build environment and creating loadable images, please refer to the ML Embedded Evaluation Kit document file in the root directory.
The instructions and build by default is for Generation 2 device AppKit. To build for Generation 1 devices please use the gen1 branch and make sure -DTARGET_REVISION=A.
By default the build is for AppKit; for any other board set the TARGET_BOARD apropriately.
The default camera for the build is ARX3A0. If you have MT9M114 camera, set the ENSEMBLE_CAMERA_MODULE accordingly (-DENSEMBLE_CAMERA_MODULE=MT9M114).
Supported camera modules:
- ARX3A0 This is the default camera
- MT9M114 This camera module has also been shipped with Alif Semiconductor kits
IMPORTANT NOTE: Due to syncing up with the latest from Upstream ARM ML Embedded Evaluation Kit, some dependencies have changed.
- The dependency on Python has been changed from 3.9 to Python 3.10. This may require you clean up existing setup and re-install version 3.10 again.
- The ARM Embedded Compiler version now needs to be version 6.19 or newer
The original README file content from the Arm ML embedded evaluation kit can be found in the root directory as ARM_Readme.md.