The most atomic way to train and inference a GPT in pure, dependency-free C

GPT in a QR Code ; The actual most atomic way to train and inference a GPT in pure, dependency-free JS/Python.

Interactive visualization of a minimal GPT implementation with autograd engine.

A complete GPT built from scratch in TypeScript with zero dependencies

The essence of text diffusion in ~150 lines of pure Python. Inspired by Karpathy's MicroGPT.

"Everything else is just for efficiency." — Karpathy's microgpt benchmarked across scalar autograd, NumPy, and PyTorch (RTX 5080)

Swift port of Karpathy's microgpt.py

Composable tiny intelligence from microgpt. Pure Python. Zero dependencies.

microGPT

A compact Transformer model for character-level generation with Mixture-of-Experts, Rotary Attention, and F-gram contextual augmentation.

Generating Colour Palettes Randomly

A Swift port of Karpathy’s microgpt. Runs natively on Apple Silicon GPUs via Metal Performance Shaders Graph.

Rust port of @karpathy's microgpt.py. The most atomic way to train and run inference for a GPT, now in pure Rust with zero ML framework dependencies.

Runnable microgpt_runnable.cpp (I will make it more elegant)

D port of microgpt of Andrej Karpathy

Minimal char-level GPT inspired by @karpathy's microGPT: multi-dataset runner for names, Pokémon, Shakespeare, Paul Graham essays and more. Educational, hackable, no fluff.

A conversion of Karpathy's MicroGPT to Typescript

MicroGPT rewriten in Golang by Justabot

microgpt in the fastest compiled languages running really fast!

Minimal GPT in 300 lines of pure Lua — a port of Karpathy's microgpt