Summary 

In this chapter, we explored various data structures and operations provided by PyTorch. We implemented several components, using the fundamental blocks of PyTorch. For our data preparation, we created the tensors used by our algorithm. Our network architecture was a model for learning to predict average hours spent by users on our Wondermovies platform. We used the loss function to check the standard of our model and used the optimize function to adjust the learnable parameters of our model to make it perform better.

We also looked at how PyTorch makes it easier to create data pipelines by abstracting away several complexities that would require us to parallelize and augment data. 

In the next chapter, we will dive deep into how neural networks and deep learning algorithms work. We will explore various PyTorch built-in modules for building network architectures, loss functions, and optimizations. We will also show how to use them on real-world datasets.