Share your thoughts, 1 month free Claude Pro on usSee more

Offline-to-online Reinforcement Learning on D4RL Hopper expert discretized

47.1Online Normalised Score

DRIFT

Updated 21d ago

Evaluation Results

Method	Links
DRIFT 2026.05		47.1	0.4
PEX 2026.05		39.5	0.1
CQL 2026.05		29.3	8.8
Cal-QL 2026.05		25.7	8.8
DQN 2026.05		21.7	0.3
IQL 2026.05		17.2	0.1
AWAC 2026.05		11.2	11.5
PPO 2026.05		3.7	-
SPA 2026.05		0.3	0.3