ResearchBenchmarksContinuous control locomotion on MuJoCo Humanoid v3 (train)Continuous control locomotion on MuJoCo Humanoid v3 (train)Follow665Avg Performance (1M Steps)GePPO580.76602.63624.5646.37Oct 29, 2021Evaluation resultsMethodMethodLinksAvg Performance (1M Steps)Final PerformanceSteps (M) to Final PerformancePercent ImprovementGePPOclipping parameter=0.1...clipping parameter=0.1, value function estimator=V-trace2021.106651,3450.8514PPOclipping parameter=0.2...clipping parameter=0.2, batch size=20482021.105849721—