ResearchBenchmarksOffline Reinforcement Learning on D4RL Hopper-medium-expert v0Follow3,907ReturnImitation440.681,340.592,240.53,140.41Feb 22, 2021Evaluation ResultsMethodMethodLinksReturnImitation2021.023,907MBPO2021.022,105SAC2021.021,850MOPOvariant=bootstrapvariant=bootstrap2021.021,063GELATOdistance=l2distance=l22021.02815GELATOdistance=metricdistance=metric2021.02574