Proprioceptive Control on DeepMind Control Suite
584Acrobot Swingup ReturnTD-MPC2
Evaluation results
| Method | Links | |||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| TD-MPC2Time steps=500k, Seeds=102026.02 | 584 | 984 | 996 | 1,000 | 875 | 845 | 917 | 265 | 506 | 407 | 486 | 986 | 979 | 947 | 659 | 425 | 952 | 181 | 658 | 754 | 846 | 942 | 963 | 983 | 960 | 854 | 991 | 981 | 783 | 896 | 868 | |
| ULDTime steps=500k, Seeds=102026.02 | 567 | 986 | 999 | 1,000 | 865 | 798 | 748 | 569 | 992 | 800 | 766 | 937 | 953 | 950 | 792 | 298 | 955 | 278 | 868 | 883 | 927 | 768 | 968 | 988 | 974 | 748 | 992 | 985 | 835 | 927 | 907 | |
| DreamerV3Time steps=500k, Seeds=102026.02 | 230 | 968 | 998 | 999 | 736 | 702 | 699 | 4 | 22 | 10 | 17 | 666 | 906 | 864 | 813 | 116 | 747 | 0 | 5 | 1 | 774 | 130 | 193 | 966 | 919 | 510 | 941 | 898 | 530 | 700 | 577 | |
| TD7Time steps=500k, Seeds=102026.02 | 58 | 983 | 999 | 1,000 | 869 | 573 | 821 | 69 | 582 | 21 | 52 | 335 | 912 | 470 | 86 | 87 | 670 | 57 | 317 | 176 | 500 | 645 | 949 | 970 | 898 | 804 | 983 | 977 | 566 | 613 | 612 | |
| PPOTime steps=500k, Seeds=102026.02 | 39 | 769 | 999 | 1,000 | 776 | 391 | 269 | 26 | 129 | 31 | 40 | 459 | 182 | 58 | 103 | 10 | 128 | 0 | 5 | 1 | 115 | 144 | 122 | 367 | 125 | 97 | 431 | 283 | 254 | 127 | 154 |