The Building Blocks of CNNs
We’ll spend the next few pages browsing through the core components of convolutional neural networks. To be clear, you shouldn’t expect that you’ll be able to build a CNN after reading through these few pages. However, you will get an idea of how they work—enough of it to run a convolutional network on CIFAR-10.
Let’s start with the most important difference between a fully connected neural network and a convolutional neural network: how they look at data.
An Image Is an Image
The first time we built a network to process images (in Chapter 6, Getting Real), we noted a potentially surprising detail: the neural network doesn’t know that the examples are images. Instead, it sees them as flat sequences of bytes. We even ...