Open-source multilingual code model trained from scratch by Carnegie Mellon — the first non-GPT code LLM.
