Offline Reinforcement Learning on D4RL maze2d-umaze (dense) benchmark leaderboard · SOTA2 Research

Offline Reinforcement Learning on D4RL maze2d-umaze (dense)

98.3Normalized ScoreROMI-BCQ

Evaluation Results

Method	Links
ROMI-BCQ2021.10		98.3
Repb-SDE2021.10		94.3
COMBO2021.10		86.2
BRAC-P2021.10		51.5
BCQ2021.10		48.4
BAIL2021.10		43.3
ROMI-CQL2021.10		39.4
MOREL2021.10		36.9
BEAR2021.10		32.6
MOPO2021.10		29.9
BRAC-v2021.10		17.6
CQL2021.10		14.4
BC2021.10		-6.9