Locomotion on D4RL HalfCheetah-medium-replay v2
91.65Online Normalized ReturnROID
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ROIDseeds=3, evaluation_rollouts=10, step_interval=50002026.05 | 91.65 | 43.9 | |
| BOORLevaluation_rollouts=10, step_interval=50002026.05 | 91.5 | 44.5 | |
| RLPDevaluation_rollouts=10, step_interval=50002026.05 | 87.6 | — | |
| Cal-QLevaluation_rollouts=10, step_interval=50002026.05 | 59.5 | 46.5 | |
| PEXevaluation_rollouts=10, step_interval=50002026.05 | 55.2 | 44.1 | |
| DUALTime steps=1M, Number of random seeds=52026.05 | 48.78 | — | |
| DACERTime steps=1M, Number of random seeds=52026.05 | 48.18 | — | |
| Diff-QLTime steps=1M, Number of random seeds=52026.05 | 46.91 | — | |
| EDISTime steps=1M, Number of random seeds=52026.05 | 46.65 | — | |
| DiffuserTime steps=1M, Number of random seeds=52026.05 | 45.82 | — | |
| IQLTime steps=1M, Number of random seeds=52026.05 | 45.77 | — | |
| ODTevaluation_rollouts=10, step_interval=50002026.05 | 40.4 | 39.9 |