Offline Reinforcement Learning on Minari Expert
50Humanoid ScoreO2O
Evaluation results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| O2ONumber of seeds=102026.05 | 50 | 1 | 34.8 | 63.8 | 39.8 | 100 | 99.9 | 96.2 | 101.9 | 13.4 | 60.1 | |
| SACfDNumber of seeds=102026.05 | 19.1 | 72.9 | 43.4 | 42 | 43.5 | 98.7 | 99.3 | 99.1 | 102.6 | 97.2 | 71.8 | |
| RLPDNumber of seeds=102026.05 | 6.2 | 93 | 50.1 | 51.3 | 73.5 | 98.2 | 91.2 | 99.3 | 102.8 | 12.7 | 67.8 | |
| SPEQNumber of seeds=102026.05 | 4.7 | 95.4 | 72.1 | 41.6 | 55.5 | 100 | 100 | 99.6 | 102.7 | 97.5 | 76.9 | |
| SOPENumber of seeds=102026.05 | 4.7 | 100.1 | 76.8 | 50.5 | 66.1 | 100 | 99.3 | 99.9 | 102.2 | 98.1 | 79.8 |