Previously, we built several simple networks to solve regression and classification problems. These illustrated the basic code structure and concepts involved in building ANNs with PyTorch.
In this chapter, we will extend simple linear models by adding layers and using convolutional layers to solve nonlinear problems found in real-world examples. Specifically, we will cover the following topics:
- Hyper-parameters and multilayered networks
- Build a simple benchmarking function to train and test models
- Convolutional networks