Train your own ChatGPT on Apple Silicon — MLX port of nanochat

A minimal, hackable Vision-Language Model built on Karpathy’s nanochat — add image understanding and multimodal chat for under $200 in compute.

The best GPT that $100-$125 worth of pre-training and finetuning can buy

nanochat's inference engine re-vibed in C++ with GGML.

Ascend NPU fork of nanochat for LLM training with torch_npu/HCCL (experimental)

I built this repo to prove to my granny that I can implement GPT.

Implementing Karpathy's nanochat in rust.

Rust-first continuation of nanochat-rs, benchmarked against karpathy/nanochat for faster, reproducible LLM training, evaluation, and inference.

I see now that the path I choose through the maze makes me what I am.

AI agents running research on single-GPU nanochat training automatically

Run nanochat training efficiently on Huawei Ascend NPUs with minimal code changes, supporting tokenizer, pretraining, and evaluation workflows.

This script uses Modal to run Karpathy's nanochat training speedrun on cloud GPUs (8× H100s).

implementing a Language Model Chat from scratch