Continuous Control on DMControl 100k
986.38DMControl: Finger Spin ScoreSampled MuZero
Evaluation Results
| Method | Links | ||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Sampled MuZeroAction Repeat=1, Environment Steps=5M, Input=Pixels2021.04 | 986.38 | 868.87 | 982.26 | 914.39 | 975.46 | 977.38 | 417.52 | 984.86 | 998.14 | 846.91 | 972.53 | 963.07 | 528.24 | 926.5 | 837.76 | 923.54 | 933.77 | 971.53 | 931.06 | 987.79 | — | — | |
| ResWMseeds=52026.03 | 986 | 845 | 942 | 542 | 694 | 963 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 828.7 | — | |
| ResActseeds=52026.03 | 974 | 819 | 917 | 503 | 772 | 948 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 822.1 | — | |
| TDMPCsteps=100K2022.05 | 943 | 770 | 628 | 222 | 577 | 933 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PlayVirtual2022.01 | 915 | 816 | 785 | 474 | 600 | 926 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 729.3 | 800.5 | |
| PlayVirtualseeds=52026.03 | 915 | 816 | 785 | 474 | 460 | 929 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 729.8 | — | |
| MLRbase_agent=SAC, trajectory_length_K=16, mask_ratio=50%, masked_cube_size=8x10x102022.01 | 907 | 806 | 866 | 482 | 643 | 933 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 772.8 | 836 | |
| MLRseeds=52026.03 | 907 | 806 | 866 | 482 | 643 | 933 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 772.8 | — | |
| DrQ2022.01 | 901 | 759 | 601 | 344 | 612 | 913 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 688.3 | 685.5 | |
| DrQinput=pixel2022.05 | 901 | 759 | 601 | 344 | 612 | 913 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PSRLseeds=52026.03 | 882 | 849 | 621 | 398 | 595 | 922 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 711.1 | — | |
| TACOseeds=52026.03 | 876 | 782 | 821 | 402 | 601 | 902 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 730.7 | — | |
| SVEAseeds=52026.03 | 859 | 727 | 811 | 375 | 747 | 915 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 739 | — | |
| Baselinebase_agent=SAC, data_augmentation=random crop and random intensity2022.01 | 853 | 784 | 593 | 399 | 424 | 648 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 616.8 | 620.5 | |
| PEERbackbone=CURL, input=pixel2022.05 | 820 | 863 | 961 | 499 | 714 | 968 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| STATE SACInput Type=State2020.04 | 811 | 835 | 746 | 616 | 891 | 746 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| State SACinput=state2022.05 | 811 | 835 | 746 | 616 | 891 | 746 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MaDiseeds=52026.03 | 810 | 704 | 766 | 432 | 574 | 884 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 695 | — | |
| PEER + TDMPCsteps=100K, combination=TDMPC2022.05 | 772 | 848 | 841 | 636 | 876 | 937 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CURLInput Type=Pixels, Gradient updates per step=1, Seeds=102020.04 | 767 | 582 | 538 | 299 | 403 | 769 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CURL2022.01 | 767 | 582 | 538 | 299 | 403 | 769 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 559.7 | 560 | |
| CURLinput=pixel2022.05 | 767 | 582 | 538 | 299 | 403 | 769 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CURLseeds=52026.03 | 767 | 582 | 538 | 299 | 403 | 769 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 559.7 | — | |
| SAC+AEInput Type=Pixels2020.04 | 740 | 311 | 274 | 267 | 394 | 391 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SAC+AEinput=pixel2022.05 | 740 | 311 | 274 | 267 | 394 | 391 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SLACV1Input Type=Pixels, Gradient updates per step=12020.04 | 693 | — | — | 319 | 361 | 512 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SLAC2022.01 | 693 | — | — | 319 | 361 | 512 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 471.3 | 436.5 | |
| DreamerAction Repeat=2, Environment Steps=5M, Input=Pixels2021.04 | 498.88 | 833.66 | 935.08 | 894.56 | 961.67 | 962.48 | 365.26 | 979.56 | 941.84 | 812.22 | 825.86 | 891.38 | 368.97 | 923.72 | 833 | 888.39 | 931.61 | 817.05 | 824.67 | 977.99 | — | — | |
| Dreamer-V2steps=100K2022.05 | 414 | 697 | 633 | 501 | 705 | 693 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DREAMERInput Type=Pixels2020.04 | 341 | 326 | 314 | 235 | 277 | 246 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Dreamer2022.01 | 341 | 740 | 314 | 235 | 277 | 246 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 289.8 | 295.5 | |
| Dreamerinput=pixel2022.05 | 341 | 326 | 314 | 235 | 277 | 246 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SAC+AE2022.01 | 326 | 311 | 274 | 267 | 394 | 391 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 396.2 | 351 | |
| DrQ v2input=pixel2022.05 | 325 | 677 | 256 | 273 | 171 | 359 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PIXEL SACInput Type=Pixels2020.04 | 179 | 419 | 145 | 197 | 42 | 312 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PLANETInput Type=Pixels2020.04 | 136 | 297 | 20 | 138 | 224 | 0 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PlaNet2022.01 | 136 | 297 | 20 | 138 | 224 | 0 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 135.8 | 137 | |
| PlaNetinput=pixel2022.05 | 136 | 297 | 20 | 138 | 224 | 0 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — |