✳flâneur — a map of the web's best reading

Chapter 0: Fundamentals - ARENA

learn.arena.education · 3,206 words · saved by 1 readers

Note - from this point on we'll start referring to the PyTorch documentation pages quite a lot. We will also include a lot of content within this material if we want to highlight it for you, however it's also an important skill to be able to use documentation pages to find answers to specific questions & assist you in debugging. One of the most basic parts of PyTorch that you will see over and over is the nn.Module class. All types of neural net components inherit from it, from the simplest nn.Relu to the most complex nn.Transformer. Often, a complex nn.Module will have sub-Modules which implement smaller pieces of its functionality. Other common Modules you'll see include The list goes on, including activation functions, normalizations, pooling, attention, and more. You can see all the Modules that PyTorch provides here. You can also create your own Modules, as we will do often! The Module class provides a lot of functionality, but we'll only cover a little bit of it here. In this sec

1️⃣ Making your own modules Learning Objectives Learn how to create your own modules in PyTorch, by inheriting from nn.Module Assemble the pieces together to create a simple fully-connected network, to classify MNIST digits Note - from this point on we'll start referring to the PyTorch documentation pages quite a lot. We will also include a lot of content within this material if we want to highlight it for you, however it's also an important skill to be able to use documentation pages to find answers to specific questions & assist you in debugging. Subclassing nn.Module One of the most basic p

Explore this link on the map →

related reading