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

Train Llama 3 models from scratch. Any scale, any personality. By Arianna Method.

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

A legible and hackable port of Nanochat in JAX/Flax

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

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

Production-honest small language model training factory: data import, pretraining, SFT, eval gates, contamination checks, and GPU runbooks.

The best ChatGPT-style model that $100 of TPU time can buy.

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

llama.cpp fork adding nanochat architecture support, so Karpathy's nanochat checkpoints run as GGUF. See NANOCHAT.md

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

The best ChatGPT that $100 can buy ported to an Nvidia RTX 5090 which is NOT $100

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

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

Fork of karpathy/autoresearch: autonomous AI-agent experiments for single-GPU nanochat/LLM training research.

Implementing Karpathy's nanochat in rust.

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

Controlled benchmark of memory mechanisms for transformers, built on nanochat

The official implementation of Ringmaster LMO, an asynchronous distributed optimizer for neural network training under heterogeneous compute environments.

AI agents running research on single-GPU nanochat training automatically