Show HN: Minimal LLM Post-Training Experiments on an 8GB GPU (SFT, DPO, GRPO)github.com21 points·popopanda··0 commentsOpen articleSaveView on HN