ResearchBenchmarksOffline Reinforcement Learning on D4RL Walker2d medium-expert v0Follow1,515ReturnGELATO-62.68346.91756.51,166.09Feb 22, 2021Evaluation ResultsMethodMethodLinksReturnGELATOdistance=metricdistance=metric2021.021,515MOPOvariant=bootstrapvariant=bootstrap2021.021,296GELATOdistance=l2distance=l22021.021,255Imitation2021.02329MBPO2021.02222SAC2021.02-2