ResearchBenchmarksOffline multitask Reinforcement Learning on D4RL Antmaze medium-diverseFollow631Episodic ReturnDiSPO220.2326.85433.5540.15Mar 10, 2024Evaluation ResultsMethodMethodLinksEpisodic ReturnDiSPO2024.03631COMBO2024.03418GC-IQL2024.03403USFA2024.03394FB2024.03294RaMP2024.03266MOPO2024.03236