Teach Yourself Transformers

This website is an annotated bibliography containing all of the free online resources I used to teach myself machine learning from July 2021-May 2023.

It starts with the absolute basics of machine learning (finding line of best fit to a scatterplot) and ends with coding a functional GPT-style text generation model in a Jupyter notebook (and being able to understand everything that's going on.)

I took notes on lectures in Obsidian (opens in a new tab), which were then converted into spaced repetition flashcards using Anki (opens in a new tab) so core concepts could be drilled. (There's a plugin (opens in a new tab) for it.) I highly recommend this approach if you're serious about retaining the material.

I cannot claim to be an expert in this field. However, if you're hoping to reach the level of "seasoned amateur," I think you could do worse than this sequence of videos.

Only prerequisite is Python. If you want to learn that, this Replit course (opens in a new tab) seemed pretty cool. FreeCodeCamp has a Python (opens in a new tab) section, too.

You can check out more of my work at joeholmes.dev (opens in a new tab).