Offline Reinforcement Learning on D4RL maze2d-med sparse benchmark leaderboard · SOTA2 Research

Offline Reinforcement Learning on D4RL maze2d-med sparse

82.4Normalized ScoreROMI-BCQ

Evaluation Results

Method	Links
ROMI-BCQ2021.10		82.4
BRAC-v2021.10		70.6
BRAC-P2021.10		70
COMBO2021.10		68.5
MOPO2021.10		37.4
BEAR2021.10		25
BAIL2021.10		19.7
BCQ2021.10		17.1
CQL2021.10		14.6
ROMI-CQL2021.10		13.3
BC2021.10		-0.5
Repb-SDE2021.10		-4.7
MOREL2021.10		-4.8