May 2018
Beginner
490 pages
13h 16m
English
TensorFlow is a graph-driven solution based on graph theory, a branch of mathematics. Designing deep learning without graphs would prove quite difficult. My XOR FNN built from scratch in Chapter 4, Become an Unconventional Innovator, fits the need of the case study described. However, if thousands of nodes are required, TensorFlow, Keras (with TensorFlow backend), or other similar solutions will be needed. In this chapter, a neural network will be developed in TensorFlow.
The building blocks of graphs are nodes (or vertices, or points) and edges (or lines) that connect the nodes to one another.
In TensorFlow parlance, nodes are cleverly named ops for operations. In Chapter ...
Read now
Unlock full access