i created something similar a while back. Inspired by micrograd for showing the connection between math/calculus and code, then building along the way to a full NumPy deep learning library that I pretrained GPT-2 124M model with it. One way to learn is to trace through the PRs merged to the repo in chronological order.

The philosophy is the same: "What I cannot create, I do not understand." https://github.com/workofart/ml-by-hand

wow this is nice work. i am diving into this world and was hoping to this exact thing. i am going to look to see how you did it. ty for sharing.