M2M Day 372: On my way to Germany
This post is part of Month to Master, a 12-month accelerated learning project. For October, my goal is to defeat world champion Magnus Carlsen at a game of chess.
In a few hours, I get on a plane to Germany. I’ll be there until the 10th, and then will be in Denmark through the 12th.
In other words, I’ll mostly be traveling for the rest of this challenge.
I could of course stay in my hotel room and continue working on my chess code, but exploring these new countries is a much better use of my time.
So, let’s take stock of where I currently am…
Last night, I started running a new program that looked promising. However, almost immediately after publishing yesterday’s post, and after 4.5 hours of processing, the program failed: File sizes once again got too big, and this time, it seems like it had nothing to do with my infrastructure configuration, and instead, had to do with the limits of Python and Tensorflow.
Clearly, there’s some kind of optimization I’m overlooking.
This morning, I set up a bunch of other experiments, slicing and dicing the dataset in all sorts of ways.
I’m starting to get quite proficient at setting up new machine learning experiments (and also just conceptualizing new, useful experiments to run). However, even though I’ve made major progress on the software front, my hardware skills are still very raw, and I have a lot to learn about setting up the infrastructure necessary to successfully use large datasets and models.
I’m not sure I’ll have the time in Europe to address these hardware skill deficiencies.
Nevertheless, while my trip to Europe may slow down my chess code progress, it is still a continuation of this challenge’s narrative. I’ll explain soon…
Read the next post. Read the previous post.
