Pendulum
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Pendulum v1
1,378.78Mean Score
40
Pendulum
-145.49Avg Episode Reward
26
Pendulum
0.7MAE
21
Pendulum v1 (test)
-164.82Average Return
16
Pendulum High observation noise
8.881MSE
14
Pendulum High process noise
0.112MSE
14
Pendulum Default noise
0.084MSE
14
Pendulum (test)
0.003MSE
14
Pendulum v1
2.18Average AUC (z-scored)
13
Pendulum
0.72Robustness Gap
12
Pendulum
1.05Rollout MSE
12
Pendulum
5.6Median Samples
12
Pendulum v1
-150.8Average Cumulative Reward
11
Pendulum
3.32MSE
11
Pendulum 90cm
1.07Length (m)
9
Pendulum Nonmarkov v1 (test)
-556.9AUC@T
9
Pendulum
0.069MSE
9
Pendulum v0
21Median Samples
9
Pendulum (P) (test)
95Reward Ratio
8
Pendulum
0.02Regret
8
Pendulum discretized n = 10^5
0Failure Rate
8
Pendulum (test)
1MSE
8
Pendulum classical control (1M steps)
-133.42Return
8
Pendulum Numerical
99AUC
7
Pendulum v1
89.2Local Optima Escape Rate
7