Downstream model-based control on Walker2d OpenAI Gym (test)
2,134.6Accumulated RewardOurs
Evaluation results
| Method | Links | |
|---|---|---|
| OursPretrain=true2026.03 | 2,134.6 | |
| TrajworldPretrain=true2026.03 | 1,933.52 | |
| OursPretrain=false2026.03 | 707.61 | |
| TrajworldPretrain=false2026.03 | 395.23 | |
| TDMPretrain=true2026.03 | 207.61 | |
| MLPEnsemblePretrain=true2026.03 | 190.51 | |
| TDMPretrain=false2026.03 | 122.65 | |
| MLPEnsemblePretrain=false2026.03 | 119.18 |