G github.com Show HN: Minimal LLM Post-Training Experiments on an 8GB GPU (SFT, DPO, GRPO) 21 points popopanda 3 months ago 0 comments
