HNNewShowAskJobs Built with Analog

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

2 points | by popopanda an hour ago

No comments yet