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

Offline Reinforcement Learning on D4RL maze2d-med dense

102.6Normalized ScoreROMI-BCQ

Evaluation Results

Method	Links
ROMI-BCQ2021.10		102.6
COMBO2021.10		84.2
BRAC-P2021.10		73
MOPO2021.10		70.5
BAIL2021.10		41.7
BCQ2021.10		41.1
ROMI-CQL2021.10		37
CQL2021.10		30.5
BRAC-v2021.10		24.4
Repb-SDE2021.10		22.7
BEAR2021.10		19.1
BC2021.10		2.7
MOREL2021.10		-1.9