Neural Networks and Tensorflow - 6 - Basics of TensorFlow - 1.png

Finally, after going through the much needed theory (though not to deep), we going to start looking at TensorFlow with a hands-on approach.

In this lesson we learn about the short history of TensorFlow (open sourced in 2015), it's huge adoption by academia and developers, as well as how it is used in deep learning.

One of the important points of focus is that a typical project in TF is composed of two phases, the creation of the computational graph, and the execution of the graph.

Ok, but until then, what the hell is a tensor, and why the name TensorFlow? Well, a tensor could simply be a multi-dimensional array, which we use to represent the data that we're working it. And the name TensorFlow can simply be understood as the data (as tensor) flows through the computational graph.

Obviously, if you watch the lesson in the video below, you'd better understand what I am talking about here because I also use some graphical illustrations while I explain...