Continuous Control on HumanoidBench without dexterous hands
893Pole ScoreFoG
Evaluation Results
| Method | Links | |||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| FoGEnvironment steps=1M, Action repeat=22026.05 | 893 | 674 | 466 | 81 | 616 | 770 | 828 | 331 | 971 | 114 | 2,434 | 749 | 671 | 866 | 0.846 | 0.794 | 0.802 | |
| DR.QEnvironment steps=1M, Action repeat=22026.05 | 887 | 355 | 401 | 92 | 205 | 843 | 931 | 354 | 973 | 344 | 8,101 | 820 | 856 | 850 | 0.864 | 0.823 | 0.825 | |
| SimbaV2Environment steps=1M, Action repeat=22026.05 | 791 | 487 | 493 | 143 | 723 | 679 | 875 | 313 | 946 | 202 | 3,850 | 415 | 814 | 845 | 0.799 | 0.781 | 0.776 | |
| TDMPC2Environment steps=1M, Action repeat=22026.05 | 744 | 334 | 378 | 31 | 42 | 723 | 790 | 244 | 962 | 387 | 2,654 | 778 | 798 | 814 | 0.734 | 0.696 | 0.71 | |
| SimbaEnvironment steps=1M, Action repeat=22026.05 | 716 | 277 | 269 | 75 | 337 | 512 | 833 | 354 | 923 | 175 | 3,874 | 232 | 772 | 550 | 0.521 | 0.598 | 0.606 | |
| MR.QEnvironment steps=1M, Action repeat=22026.05 | 578 | 303 | 235 | 69 | 135 | 553 | 850 | 344 | 932 | 131 | 4,902 | 278 | 800 | 716 | 0.519 | 0.602 | 0.604 | |
| TD7Environment steps=1M, Action repeat=22026.05 | 441 | 39 | 52 | 79 | 69 | 235 | 874 | 147 | 582 | 60 | 1,409 | 91 | 433 | 33 | 0.134 | 0.284 | 0.289 | |
| DreamerV3Environment steps=1M, Action repeat=22026.05 | 41 | 11 | 7 | 11 | 9 | 15 | 19 | 113 | 248 | 4 | 3,203 | 4 | 15 | 8 | 0.007 | 0.021 | 0.022 |