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:
Press enter or click to view image in full size
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.
