This is the documentation for the latest (master) development branch of Zephyr. If you are looking for the documentation of previous releases, use the drop-down menu on the left and select the desired version.

TensorFlow Hello World sample


This sample TensorFlow application replicates a sine wave and demonstrates the absolute basics of using TensorFlow Lite Micro.

The model included with the sample is trained to replicate a sine function and generates x values to print alongside the y values predicted by the model. The x values iterate from 0 to an approximation of 2π.

The sample also includes a full end-to-end workflow of training a model and converting it for use with TensorFlow Lite Micro for running inference on a microcontroller.


This README and sample have been modified from the TensorFlow Hello World sample for Zephyr.

Building and Running

This sample should work on most boards since it does not rely on any sensors.

This application can be built and executed on QEMU as follows:

west build -b qemu_x86 samples/tensorflow/hello_world
west build -t run

Exit QEMU by pressing CTRL+A x.

Sample Output


x_value: 1.0995567*2^1, y_value: 1.6951603*2^-1

x_value: 1.2566366*2^1, y_value: 1.1527088*2^-1

x_value: 1.4137159*2^1, y_value: 1.1527088*2^-2

x_value: 1.5707957*2^1, y_value: -1.0849024*2^-6

x_value: 1.7278753*2^1, y_value: -1.0509993*2^-2


The modified sample prints 50 generated-x-and-predicted-y pairs.

Modifying Sample for Your Own Project

It is recommended that you copy and modify one of the two TensorFlow samples when creating your own TensorFlow project. To build with TensorFlow, you must enable the below Kconfig options in your prj.conf and set a flag in your CMakeLists.txt.




set(CMAKE_CXX_FLAGS  "${CMAKE_CXX_FLAGS} -fno-threadsafe-statics")


Follow the instructions in the train/ directory to train your own model for use in the sample.