Deep Learning


Now it's time to learn an ML framework and practice with some coding.

Python Engineer, our friend that helped us code linear and logistic regression from scratch, is a great resource here too.

Our subject matter knowledge extends up to about lesson 13 in this playlist (opens in a new tab), and that's where I stopped when I studied this.


You'll notice Python Engineer spends a lesson discussing PyTorch's autograd feature. This is the integral feature of a deep learning framework, in that it allows the computer to calculate the gradients of each parameter in an efficient, automatic way. As mentioned in the last section, it's good to let really difficult material wash over you from time to time, not understanding any of it but still letting it soak into your mind. In that spirit, this video from research scientist Ari Seff is a great resource on what's going on mathematically behind automatic differentiation (autograd's underlying math.)