Show HN: Minimal LLM Post-Training Experiments on an 8GB GPU (SFT, DPO, GRPO)(github.com)
7 pointsby popopanda5 hours ago1 comment