TensorFlow Lite for Raspberry Pi

Cross compiling

This has been tested on Ubuntu 16.04.3 64bit and Tensorflow devel docker image tensorflow/tensorflow:nightly-devel.

To cross compiling TensorFlow Lite. First you should install the toolchain and libs.

sudo apt-get update
sudo apt-get install crossbuild-essential-armhf

If you are using docker, you may not use sudo

Clone this Tensorflow repository, Run this script at the root of the repository to download all the dependencies:

The Tensorflow repository is in /tensorflow if you are using tensorflow/tensorflow:nightly-devel docker image, just try it.

./tensorflow/contrib/lite/download_dependencies.sh

Note than you only need to to this once.

You should then be able to compile:

./tensorflow/contrib/lite/build_rpi_lib.sh

This should compile a static library in: tensorflow/contrib/lite/gen/lib/rpi_armv7/libtensorflow-lite.a.

This has been tested on Raspberry Pi 3b, Raspbian GNU/Linux 9.1 (stretch), gcc version 6.3.0 20170516 (Raspbian 6.3.0-18+rpi1).

Log in to you RPI, install the toolchain.

sudo apt-get instal build-essential

First, clone this TensorFlow repository. Run this at the root of the repository:

./tensorflow/contrib/lite/download_dependencies.sh

Note than you only need to to this once.

You should then be able to compile:

./tensorflow/contrib/lite/build_rpi_lib.sh

This should compile a static library in: tensorflow/contrib/lite/gen/lib/rpi_armv7/libtensorflow-lite.a.