This post is an udate of the last post on how to make a standalone C++ project with tensorflow without using Bazel. This post aims at upgrading tensorflow to the latest stable version 1.8.0 and some tweaks that help tensorflow work nicely with opencv.

Updates:

• The configurations to compile for tensorflow 1.8.0 is described in this post. If you want tensorflow to work nicely with OpenCV, follow this post with only one change described in that post.

Before going into details of how an engineer shoud take the first step on Machine Learning (ML), I want to clarify that I am an embedded engineer who are amazed by the beauty of ML and have started learning it for just over a year. So everything I write in this blog is subjective to my point of view.