ResearchBenchmarksOffline Reinforcement Learning on DMControl ball-in-cup-catch (mixed)Follow51.28Normalized ScoreCQL + S2P27.630433.770239.9146.0498Sep 30, 2022Evaluation ResultsMethodMethodLinksNormalized ScoreCQL + S2PS2P Augmentation=trueS2P Augmentation=true2022.0951.28IQLS2P Augmentation=falseS2P Augmentation=false2022.0941.94SLAC-off + S2PS2P Augmentation=trueS2P Augmentation=true2022.0940.41IQL + S2PS2P Augmentation=trueS2P Augmentation=true2022.0937.79CQLS2P Augmentation=falseS2P Augmentation=false2022.0930.82SLAC-offS2P Augmentation=falseS2P Augmentation=false2022.0928.54