Over the last couple of days I finished Fast AI's lesson four and anyone watching would have seen me on the edge of my seat grinning like an idiot at what probably looked like an Excel tutorial (two separate people did come up and make comments to that effect, but that's another story).
The important thing is that we were able to cover THREE neural network-related concepts in spreadsheet form (my favorite form): convolutions, gradient descent, and collaborative filtering. I was probably most excited by the demonstration of collaborative filtering (because it's kind of dope), and gradient descent second (because I've been having a lot of trouble with it), but we're going to go in order because it does kind of make the most sense in order.
I'm so excited to do this I don't want to go for (dance) class right now but at the same time it's important to me that I go for class so as soon these epochs finish running I should go.
I'm back! I was gone for like nine hours. Let's get into it. This one definitely works better on a larger screen so... Maybe go do that.