Why Do we Normalize Inputs?
When the input is not normalized, the shape of the cost function can become distorted like the diagram on the left. This leads to instability when optimizing the model. The training speed decreases depending on where the weights and biases are initialized.
On the other hand, if you normalize the inputs, the shape of the cost function will become symmetric. This means, that the training speed will be relatively the same no matter where the weights and biases were initialized.
