Introduction to scalars, vectors and matrices

Introduction

Before we start talking about TensorFlow, let's try to understand what a Tensor is.

A tensor is basically an N-dimensional array with a uniform data type (called dtype). You can think of tensors as similar to NumPy arrays(np.array).

Note: Tensors are immutable. You can never update the content of a tensor but you can always create new ones.

Some of the basic tensors include:

• Scalars - tensors with rank-0 or no axes/dimension
• Vectors - tensors with rank-1 or 1 axis
• Matrix - tensors with rank-2 or 2 axes

And then you can have as many axes as you want.

Here's a diagram to help you visualize this:

Let's start off by creating some basic tensors.

Import the libraries

The first step is to import the tensorflow & numpy library, so go ahead and add the code snippet below in the cell and run it by pressing shift + Enter on your keyboard: 

import numpy as np
import tensorflow as tf

Check the version of tensorflow using:

print(tf.__version__)

You should have some version of TensorFlow 2.x. If not you can use ! pip install tensorflow==2.4 to install.

Here's what this should look in the colab notebook:

With the tools imported for us to use, let's create some tensors!