40 Algorithms Every Programmer Should Know

Presenting TensorFlow's Basic Concepts

Let's take a brief look at TensorFlow concepts such as scalars, vectors, and matrices. We know that a simple number, such as three or five, is called a scalar in traditional mathematics. Moreover, in physics, a vector is something with magnitude and direction. In terms of TensorFlow, we use a vector to mean one-dimensional arrays. Extending this concept, a two-dimensional array is a matrix. For a three-dimensional array, we use the term 3D tensor. We use the term rank to capture the dimensionality of a data structure. As such, a scalar is a rank 0 data structure, a vector is a rank 1 data structure, and a matrix is a rank 2 data structure. These multi-dimensional structures are known as tensors and are ...