Continuous Control on DMC Suite Hard v1 (test)
721Dog Run ReturnDR.Q
Evaluation results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| DR.Qsteps=500K, action repeat=22026.05 | 721 | 972 | 925 | 950 | 465 | 938 | 925 | 0.917 | 0.844 | 0.842 | |
| FoGsteps=500K, action repeat=22026.05 | 613 | 976 | 901 | 921 | 292 | 931 | 878 | 0.88 | 0.788 | 0.787 | |
| MR.Qsteps=500K, action repeat=22026.05 | 569 | 967 | 877 | 916 | 200 | 868 | 662 | 0.796 | 0.722 | 0.723 | |
| SimbaV2steps=500K, action repeat=22026.05 | 562 | 981 | 861 | 935 | 194 | 916 | 651 | 0.808 | 0.729 | 0.729 | |
| Simbasteps=500K, action repeat=22026.05 | 544 | 960 | 824 | 916 | 181 | 846 | 668 | 0.773 | 0.706 | 0.706 | |
| MAD-TDsteps=500K, action repeat=22026.05 | 437 | 967 | 867 | 924 | 200 | 870 | 684 | 0.787 | 0.707 | 0.708 | |
| iQRLsteps=500K, action repeat=22026.05 | 380 | 926 | 713 | 866 | 188 | 727 | 688 | 0.694 | 0.64 | 0.642 | |
| BROsteps=500K, action repeat=22026.05 | 374 | 966 | 783 | 931 | 204 | 920 | 672 | 0.772 | 0.694 | 0.693 | |
| TDMPC2steps=500K, action repeat=22026.05 | 265 | 506 | 407 | 486 | 181 | 658 | 754 | 0.464 | 0.486 | 0.465 | |
| TD7steps=500K, action repeat=22026.05 | 69 | 582 | 21 | 52 | 57 | 317 | 176 | 0.134 | 0.069 | 0.182 | |
| PPOsteps=500K, action repeat=22026.05 | 26 | 129 | 31 | 40 | 0 | 5 | 1 | 0.021 | 0.026 | 0.033 | |
| DreamerV3steps=500K, action repeat=22026.05 | 4 | 22 | 10 | 17 | 0 | 5 | 1 | 0.008 | 0.005 | 0.009 |