Continuous Control on MuJoCo HalfCheetah
13,144Average RewardTOP-TD3
Evaluation Results
| Method | Links | |
|---|---|---|
| TOP-TD3Training steps=1M, Number of trials=102021.02 | 13,144 | |
| ND TOP-TD3Training steps=1M, Number of trials=102021.02 | 12,477 | |
| OACTraining steps=1M, Number of trials=102021.02 | 11,723 | |
| QAvatarTraining steps=500k steps2026.03 | 11,586 | |
| QR-TD3Training steps=1M, Number of trials=102021.02 | 11,170 | |
| SACTraining steps=500k steps2026.03 | 10,986 | |
| SACTraining steps=1M, Number of trials=102021.02 | 10,815 | |
| FTTraining steps=500k steps2026.03 | 10,756.8 | |
| TD3Training steps=1M, Number of trials=102021.02 | 9,566 | |
| CAT-SACTraining steps=500k steps2026.03 | 8,756.5 | |
| PARTraining steps=500k steps2026.03 | 8,097.4 | |
| IQ-LearnNumber of expert trajectories=102021.06 | 5,154 | |
| ExpertNumber of expert trajectories=102021.06 | 5,098 | |
| AGD-MBRLGuide=EAG2026.04 | 4,864.8 | |
| ValueDICENumber of expert trajectories=102021.06 | 4,840 | |
| AGD-MBRLGuide=SAG2026.04 | 4,642.25 | |
| GAILNumber of expert trajectories=102021.06 | 4,280 | |
| PolyGRADGuide=Diffuser Guide2026.04 | 4,188.43 | |
| PolyGRADGuide=None2026.04 | 3,879.16 | |
| BCNumber of expert trajectories=102021.06 | 2,701 | |
| PPO2026.04 | 2,408.2 | |
| TRPO2026.04 | 2,014.91 | |
| CATTraining steps=500k steps2026.03 | 46.1 | |
| CMDTraining steps=500k steps2026.03 | -253.1 | |
| RandomNumber of expert trajectories=102021.06 | -282 |