Thanks for your help. I want to deploy a tensorflow neural network to a Raspberry Pi 3 B+ with the Raspbian Stretch OS, and I'd like to know the best way to install Tensorflow Lite (or Tensorflow Mobile) onto the Pi. I found a method to install the whole Tensorflow with https://github.com/samjabrahams/tensorflow-on-raspberry-pi/blob/master/GUIDE.md by Sam J Abrahams. I also found a method to use a cross-compiler https://medium.com/@haraldfernengel/compiling-tensorflow-lite-for-a-raspberry-pi-786b1b98e646.

But, I think there may be a simpler way to install Tensorflow Lite because of https://github.com/tensorflow/tensorflow/pull/16431. I'm not clear on the installation procedure and I appreciate your guidance.

  • 1
    The project is clear on how to install on the 'usual platforms' (mac OS, Windoze, etc), but the "Lite" version is just kind of hanging out in space somewhere with no clear instructions. The other alternative of course is to make up a Raspbian Jessie SD card, and use the Sam J Abrahams version. Have you tried to get in touch with the reviewers or assignees for the tensorflow site on GitHub? – Seamus May 3 '18 at 15:13

There is now a nightly build of tf, not lite, on ci.tensorflow.org but performance is pretty poor on a Pi 3B+. I’m {hop,expect}ing that lite will be significantly faster.

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Not the answer you're looking for? Browse other questions tagged or ask your own question.