Skip to content

Instantly share code, notes, and snippets.

View ashokvarmamatta's full-sized avatar

Matta Ashok Varma ashokvarmamatta

4 followers · 8 following
View GitHub Profile
@ashokvarmamatta
ashokvarmamatta / ml-from-zero-ch3-makemore-bigram.md
Last active June 30, 2026 11:21
makemore — Your First Tiny Language Model From Scratch: a bigram neural net (one-hot → @W → softmax → NLL) that generates name-like text, trained with the same backward pass from Chapter 2. Chapter 3 of ml-from-zero. For ECE/CS grads new to ML.

makemore — Your First Tiny Language Model From Scratch. A bigram neural net that generates name-like text: one-hot → @W → softmax → negative-log-likelihood, trained with the same backward pass from Chapter 2. Chapter 3 of ml-from-zero. For ECE/CS grads new to ML.

🔤 makemore — Your First Tiny Language Model

From a Pile of Names to a Net That Invents New Ones — One Character at a Time

Typing SVG