In Chapter 3, Diving Deep into Neural Networks, we built an image classifier using a popular Convolutional Neural Network (CNN) architecture called ResNet, but we used this model as a black box. In this chapter, we will cover the important building blocks of convolutional networks. Some of the important topics that we will be covering in this chapter are:

• Introduction to neural networks
• Building a CNN model from scratch
• Creating and exploring a VGG16 model 
• Calculating pre-convoluted features
• Understanding what a CNN model learns
• Visualizing weights of the CNN layer

We will explore how we can build an architecture from scratch for solving image classification problems, which are the most common use cases. We will also learn how to use transfer learning, which will help us in building image classifiers using a very small dataset. 

Apart from learning how to use CNNs, we will also explore what these convolutional networks learn.