# What are Tensors?

### Let's get **physical**

Sometimes, nothing beats holding a copy of a book in your hands. Writing in the margins, highlighting sentences, folding corners. So this book is also available from Amazon as a paperback.

Tensors are the building blocks of a machine learning model. They are nothing to be scared of; you’ve most likely been using them for a while but have never called them *Tensors*.

They are a set of numbers of a specific size and shape.

So far, we’ve explained how neural network models are just functions that manipulate numbers, a set of input numbers of a specific size, into a set of output numbers into another size and shape. Tensors are how we represent those numbers, the inputs, the outputs, and everything in between.

## Rank

The dimensionality of a Tensor we call it’s *rank*.

A 1D Tensor (or rank 1 Tensor) is just an array, like so:

A 2D Tensor (or rank 2 Tensor) is just a

A 3D Tensor (or rank 3 Tensor) is a cube. An array of arrays of arrays, like so:

Everything after 3D becomes harder to conceptualize, but let’s try. A 4D Tensor is an array of 3D Tensors, like so:

A 5D Tensor is an array of 4D Tensors, like so:

## Shape

The size of each dimension in a Tensor we call its shape. For example, a Tensor to represent a black and white image would have the shape [`width`

, `height`

, `colors`

].

For a 640 x 480 pixel black and white image, the shape would be [640,480, 1].

`colors`

in this case would be of size 1; in a black and white image, the only color is black; in a color image, the size might be 3 with a value for red, green, and blue.

The rank of the Tensor is 3; it’s a 3D Tensor.

## How to represent a data set of images?

For one type of machine learning, supervised learning^{[1]}, we provide large sets of training data for our machine learning models.

We represent a single black and white image as a 3D Tensor of shape [`width`

, `height`

, `color`

].

We represent a set of 1000 black and white image as a 4D Tensor of shape [** sample_size**,

`width`

, `height`

, `color`

].For example, a set of 1000, 640 x 480 pixel black and white images would be represented as 4D Tensor of Shape [1000, 640, 480, 1].

## Summary

Tensors are just buckets of numbers of a specific shape and a certain rank (dimensionality). Tensors are used in Machine Learning with TensorFlow to represent input data and output data (and everything in between) in Machine Learning models.

**Advanced** JavaScript

This unique course teaches you advanced JavaScript knowledge through a series of interview questions. Bring your JavaScript to the 2021's **today**.

[🌲,🌳,🌴].push(🌲)If you find my courses useful, please consider

**planting a tree on my behalf**to combat climate change. Just $4.50 will pay for 25 trees to be planted in my name. Plant a tree!