ResearchBenchmarksOffline Reinforcement Learning on D4RL Hopper-medium v0Follow1,676ReturnGELATO270.96635.731,000.51,365.27Feb 22, 2021Evaluation ResultsMethodMethodLinksReturnGELATOdistance=metricdistance=metric2021.021,676GELATOdistance=l2distance=l22021.021,320Imitation2021.021,234MOPOvariant=bootstrapvariant=bootstrap2021.021,202MBPO2021.02457SAC2021.02325