LLM training in simple, raw C/CUDA https://github.com/karpathy/llm.c -- Let's build GPT / https://github.com/karpathy/nanoGPT Let's build the GPT Tokenizer