ResearchBenchmarksContinuous control locomotion on MuJoCo Swimmer v3 (train)Follow161Average PerformanceGePPO95.48112.49129.5146.51Oct 29, 2021Evaluation ResultsMethodMethodLinksAverage PerformanceFinal PerformanceSteps to Final PPO (M)Percent ImprovementGePPOclipping parameter=0.1...clipping parameter=0.1, value function estimator=V-trace2021.101611950.23—PPOclipping parameter=0.2...clipping parameter=0.2, batch size=20482021.10981361—