real-stanford/dice-rl

From Prior to Pro: Efficient Skill Mastery via Distribution Contractive RL Finetuning

语言：Python · GitHub 星标：95 · Fork：13

在 GitHub 上查看原仓库