Show HN: Minimal LLM Post-Training Experiments on an 8GB GPU (SFT, DPO, GRPO)

 (github.com)

21 points | by popopanda 2 days ago ago

No comments yet.