List your company
ResearchFormerly WizWand

Offline Reinforcement Learning on D4RL Gym hopper-medium-expert

112.7Normalized Avg ReturnRORL
-3.7826.4656.786.94Oct 1, 2021Jun 9, 2022Feb 15, 2023Oct 24, 2023Jul 1, 2024Mar 9, 2025Nov 15, 2025
Updated 5mo ago

Evaluation results

MethodLinks
RORLImplementation Source=Ours, Ensemble Size=102022.06112.7
SPQRBase algorithm=SAC-Min2024.01112.5
QQLDomain=Gym Locomotion, Hyperparameter Tuning=consistent2025.11112.5
GTP2025.10112.2
QIPO-Diff2025.10112.1
SPQRBase algorithm=EDAC2024.01111.7
ROMI-CQLbase algorithm=CQL2021.10111.4
XQLDomain=Gym Locomotion, Hyperparameter Tuning=individually tuned2025.11111.2
COMBOalgorithm class=model-based2021.10111.1
CQLImplementation=Original Paper2021.10111
CQL-Min2024.01111
SSCQL2025.10110.9
PBRL2022.06110.8
EDACImplementation=Ours2021.10110.7
EDACImplementation Source=Paper2022.06110.7
EDAC2024.01110.7
MXQLDomain=Gym Locomotion, Hyperparameter Tuning=individually tuned2025.11110.7
SAC-NImplementation=Ours2021.10110.1
SAC-Min2024.01110.1
QIPO-OT2025.10108
BCQImplementation=Reproduced2021.10106.9
CQLalgorithm class=model-free2021.10106.8
CQLDomain=Gym Locomotion, Hyperparameter Tuning=individually tuned2025.11105.4
FQL2025.10104.6
TD3+BCDomain=Gym Locomotion, Hyperparameter Tuning=individually tuned2025.1198
CQLImplementation=Reproduced2021.1096.9
CQL2022.0696.9
IQLDomain=Gym Locomotion, Hyperparameter Tuning=individually tuned2025.1191.5
BEARImplementation=Reproduced2021.1066.2
EDAC-10Implementation Source=Reproduced, Ensemble Size=102022.0658.1
BCImplementation=Reproduced2021.1053.9
BC2022.0653.9
BC2024.0153.9
BCDomain=Gym Locomotion, Hyperparameter Tuning=individually tuned2025.1152.5
UWACImplementation=Reproduced2021.1048.6
MORELImplementation=Reproduced2021.1046.2
BC2021.1036
BRACImplementation=Reproduced2021.1016.6
SAC-10Implementation Source=Reproduced, Ensemble Size=102022.066.1
REMImplementation=Reproduced2021.100.8
SACImplementation=Reproduced2021.100.7