Tag
Anyscale introduces a new Agent Skill for LLM post-training that automatically selects the optimal fine-tuning method (SFT, DPO, GRPO, etc.) and generates ready-to-launch configs, helping avoid wasted GPU runs.