Offline Reinforcement Learning on D4RL MuJoCo halfcheetah-medium-replay v2
70.7Normalized ScoreROMI
Evaluation results
| Method | Links | |
|---|---|---|
| ROMItraining steps=1M, seeds=5, version=v22026.03 | 70.7 | |
| COUNTtraining steps=1M, seeds=5, version=v22026.03 | 70.5 | |
| RAMBOtraining steps=1M, seeds=5, version=v22026.03 | 69.7 | |
| MOBILEtraining steps=1M, seeds=5, version=v22026.03 | 67.7 | |
| MOPOtraining steps=1M, seeds=5, version=v22026.03 | 52.1 | |
| CQLtraining steps=1M, seeds=5, version=v22026.03 | 45.3 | |
| IQLtraining steps=1M, seeds=5, version=v22026.03 | 44.3 |