Physical State Prediction on DeepMind Control Suite Acrobot Easy tasks (random policy)
0.0001MSEMoSim
Evaluation results
| Method | Links | |
|---|---|---|
| MoSimHorizon=16, Training data source=random policy data2025.04 | 0.0001 | |
| DreamerV3Horizon=16, Prediction steps=5-step, Training data source=random policy data2025.04 | 0.1015 | |
| MoSimHorizon=100, Training data source=random policy data2025.04 | 0.1043 | |
| DreamerV3Horizon=16, Prediction steps=5-step, Training data source=experience data2025.04 | 0.9356 | |
| DreamerV3Horizon=16, Prediction steps=1-step, Training data source=experience data2025.04 | 3.639 | |
| DreamerV3Horizon=16, Prediction steps=1-step, Training data source=random policy data2025.04 | 3.7423 | |
| DreamerV3Horizon=100, Prediction steps=5-step, Training data source=experience data2025.04 | 4.8957 | |
| DreamerV3Horizon=100, Prediction steps=5-step, Training data source=random policy data2025.04 | 4.9276 | |
| DreamerV3Horizon=100, Prediction steps=1-step, Training data source=experience data2025.04 | 9.7392 | |
| DreamerV3Horizon=100, Prediction steps=1-step, Training data source=random policy data2025.04 | 9.7547 |