ResearchBenchmarksContinuous Control on InvertedDoublePendulum v5Continuous Control on InvertedDoublePendulum v5Follow9,349.2Average Episodic RewardPDA6,679.9367,372.9188,065.98,758.882Mar 10, 2026Evaluation resultsMethodMethodLinksAverage Episodic RewardPDAEnvironment steps=1M,...Environment steps=1M, Seeds=10, Tests per epoch=102026.039,349.2PPOEnvironment steps=1M,...Environment steps=1M, Seeds=10, Tests per epoch=102026.036,782.6