cartpole
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
CartPole v0
200Mean Score
48
CartPole Pure
200Average Reward (2/0.5)
30
CartPole
1,000Average Reward
29
CartPole v1 (test)
500Total Reward
25
CartPole v1
495Mean Episodic Return
18
CartPole v1
354,122Return
16
Cartpole v1 (test)
100Optimality (%)
15
Cartpole
499.11Average Episode Reward
14
CartPole
0.1Wall-clock Training Time (min)
13
CartPole
500Max Return
13
Cartpole (test)
478.6Best Test Reward
11
Cartpole
1.63Median Samples
10
CartPole Pure-8-40
200Average Reward (EpLen=2, DF=0.5)
10
CartPole
6E1
8
CartPole Time-Varying (CP TV) (held-out trajectories)
0.001MSE
8
CartPole Time-Invariant (held-out trajectories)
2.28MSE
8
CartPole Setting C v0 (test)
236.4Performance (2/0.15)
8
CartPole Setting B v0 (test)
192.4Steps Survived (Config 0.15)
8
CartPole Setting A v0 (test)
98.2Performance Score (Config 2/0.85)
8
Cartpole (Ceval-in)
659AER
8
Cartpole (train)
658AER
8
Cartpole swing-up
111Median Samples
8
Safe CartPole
68.7Training Time (s)
7
Safe CartPole
200Episode Reward
7
CartPole 100k Gym
100Returns
6