LLM training in simple, raw C/CUDA
- Recommended by
- 2
- GitHub stars
- 31k
- Licence
- MIT
- Language
- Cuda
- Last commit
- 26 Jun 2025
What people said
“Karpathy: Let's reproduce GPT-2 (1.6B): one 8XH100 node 24h $672 in llm.c (182 points)”
“Reproducing GPT-2 in llm.c (618 points)”
“Llm.c – LLM training in simple, pure C/CUDA (1050 points)”
Not software, abandoned or listed by mistake? Report this listing