ResearchBenchmarksOffline Reinforcement Learning on DMControl cartpole-swingup (expert)Follow20.43Normalized ScoreIQL10.8113.307515.80518.3025Sep 30, 2022Evaluation ResultsMethodMethodLinksNormalized ScoreIQLS2P Augmentation=falseS2P Augmentation=false2022.0920.43CQLS2P Augmentation=falseS2P Augmentation=false2022.0919.35CQL + S2PS2P Augmentation=trueS2P Augmentation=true2022.0918.54IQL + S2PS2P Augmentation=trueS2P Augmentation=true2022.0918.37SLAC-offS2P Augmentation=falseS2P Augmentation=false2022.0914.11SLAC-off + S2PS2P Augmentation=trueS2P Augmentation=true2022.0911.18