CarRacing on CarRacing final 200 episodes (train)
88.45Episode LengthPPO + VC
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| PPO + VCVC Type=Positive VC (Approach), theta_thr=0.902026.05 | 88.45 | 1.36 | 46.52 | 47.88 | |
| PPO + VCVC Type=Positive VC (Approach), theta_thr=0.602026.05 | 83.2 | 2.05 | 38.89 | 40.94 | |
| PPO + VCVC Type=Positive VC (Approach), theta_thr=0.202026.05 | 81.33 | 1.65 | 46.52 | 48.17 | |
| PPO + VCVC Type=Composite VC (Positive + Negative), theta_thr=0.202026.05 | 78.67 | 1.09 | 46.39 | 47.48 | |
| PPO + VCVC Type=Composite VC (Positive + Negative), theta_thr=0.902026.05 | 61.82 | 1.51 | 21.21 | 22.72 | |
| PPO + VCVC Type=Composite VC (Positive + Negative), theta_thr=0.602026.05 | 57.46 | 1.82 | 22.28 | 24.1 | |
| Base PPOMethod Category=Baselines2026.05 | 33.27 | 4.66 | 0 | 4.66 | |
| PPO + VCVC Type=Negative VC (Avoidance), theta_thr=0.902026.05 | 24.79 | 2.83 | 4.1 | 1.27 | |
| PPO + StimuliMethod Category=Baselines2026.05 | 24.38 | 2.93 | 6.48 | 3.55 | |
| PPO + VCVC Type=Negative VC (Avoidance), theta_thr=0.202026.05 | 24.28 | 2.53 | 3.23 | 0.7 | |
| PPO + VCVC Type=Negative VC (Avoidance), theta_thr=0.602026.05 | 18.43 | 1.92 | 5.58 | 3.66 |