Continuous Control on DeepMind Control Suite Cheetah Run
836RewardDream-MPC
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Dream-MPCIntegrate with=Dreamer2026.05 | 836 | — | |
| Dreamer2026.05 | 751 | — | |
| Dreamer V32024.10 | 737.7 | 0.0337 | |
| Dreamer V3 with MLE regularisationMLE regularisation=true2024.10 | 675.2 | 0.0283 | |
| CEM+policy2026.05 | 674 | — | |
| Policy+Grad-MPC2026.05 | 591 | — | |
| PlaNet2026.05 | 535 | — | |
| SAC+AE2026.05 | 495 | — | |
| Grad-MPC2026.05 | 438 | — | |
| RC-SACtraining steps=50 [k], seeds=3, w_target=0.52026.03 | 0.62 | — | |
| TD3training steps=50 [k], seeds=3, w_target=0.52026.03 | 0.46 | — | |
| RC-TD3training steps=50 [k], seeds=3, w_target=0.52026.03 | 0.46 | — | |
| SACtraining steps=50 [k], seeds=3, w_target=0.52026.03 | 0.42 | — |