Neural Network: Forward pass

Image by Alina Grubnyak on Unsplash

Table of Contents

1. Input, hidden, and loss layers
2. Forward pass
3. Backpropagation

Introduction

In the previous article, we laid out the basic building blocks of a feedforward neural network. Now that we have an understanding of the input, hidden and loss layers of a feedforward neural network, we will see how the data flows through the neural network and creates a prediction, also known as the forward pass. The neural network that we will be using as an example is below:

Figure 1.1: Three-layer neural network with a batch size of 4, and Cross-Entropy Loss

The Forward Pass (input layer):

Let’s go through the example in Figure 1.1, since we have done most of the hard work in the previous article, this part should be relatively straightforward.

The Forward Pass (1st…