Offline Reinforcement Learning on OGBench 50 tasks (offline)
89AL ScoreTRQAM
Evaluation Results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| TRQAMAlgorithm Category=OURS, Training Steps=1M, Number of Seeds=82026.05 | 89 | 41 | 84 | 36 | 79 | 100 | 99 | 81 | 50 | 19 | 68 | |
| QAM-EAlgorithm Category=ADJOINT MATCHING, Training Steps=1M, Number of Seeds=82026.05 | 86 | 6 | 60 | 4 | 63 | 89 | 54 | 71 | 11 | 9 | 45 | |
| QAMAlgorithm Category=ADJOINT MATCHING, Training Steps=1M, Number of Seeds=82026.05 | 62 | 29 | 64 | 4 | 64 | 15 | 1 | 71 | 19 | 18 | 35 | |
| DSRLAlgorithm Category=POST PROCESSING, Training Steps=1M, Number of Seeds=82026.05 | 53 | 1 | 53 | 1 | 80 | 100 | 61 | 72 | 34 | 9 | 46 | |
| CGQL-LAlgorithm Category=GUIDANCE, Training Steps=1M, Number of Seeds=82026.05 | 48 | 7 | 57 | 6 | 58 | 0 | 0 | 55 | 0 | 1 | 23 | |
| FQLAlgorithm Category=BACKPROP, Training Steps=1M, Number of Seeds=82026.05 | 38 | 2 | 74 | 2 | 70 | 25 | 9 | 44 | 7 | 9 | 28 | |
| IFQLAlgorithm Category=POST PROCESSING, Training Steps=1M, Number of Seeds=82026.05 | 29 | 12 | 93 | 30 | 36 | 64 | 42 | 9 | 24 | 6 | 35 |