Build a binary with TensorFlow source and Zephyr as the target

[plonajakub]

Since any custom project for the Digilent board will require proper make process there is a need for exploring the process and customize it for our use.

The goal is to have valid binary suitable to be run in the Renode environment and on the real hardware - our Digilent board with Zephyr as the target. The binary must contain the TensorFlow source.

@plonajakub and @zawadzki-dawid will investigate the subject and run through the process.

Reference:

• Page on Wiki
• Zephyr documentation about building an app

[plonajakub]

I have successfully built _helloworld example mentioned on our wiki both with the west tool and manually by make/cmake (its pure Zephyr). I have also started exploring the building process of basic TFLite examples (_helloworld, _magicwand).

For now, it seems a bit unclear to me what must be done to be able to build the TFLite example on Zephyr. I will continue investigating the subject.

[plonajakub]

I've got familiar with tensorflow/lite/micro structure (base directory for TF Lite for microcontrollers project).

Documentation that I found useful:

• Developer guide on tensorflow.org
• Readme in the root of tensorflow/lite/micro structure

The next step is to find out how the sources are compiled with Zephyr as the target.

[plonajakub]

I have modified the build process of the micro_speech example - it is now working with the newest version of Zephyr. You can see my changes here.

[plonajakub]

Building Zephyr with TF sources (for micro_speech app) is now possible.