List your company
ResearchFormerly WizWand

Offline Reinforcement Learning on D4RL Gym walker2d medium-expert

121.2Normalized Average ReturnRORL
-5.05627.72260.593.278Jun 3, 2021Mar 1, 2022Nov 27, 2022Aug 25, 2023May 22, 2024Feb 17, 2025Nov 15, 2025
Updated 5mo ago

Evaluation results

MethodLinks
RORLImplementation Source=Ours, Ensemble Size=102022.06121.2
SPQRBase algorithm=SAC-Min2024.01118.2
SAC-NImplementation=Ours2021.10116.7
SAC-10Implementation Source=Reproduced, Ensemble Size=102022.06116.7
SAC-Min2024.01116.7
EDAC-10Implementation Source=Reproduced, Ensemble Size=102022.06115.4
EDACImplementation=Ours2021.10114.7
EDACImplementation Source=Paper2022.06114.7
EDAC2024.01114.7
GTP2025.10114.2
QQLDomain=Gym Locomotion, Hyperparameter Tuning=consistent2025.11113.2
IQLDomain=Gym Locomotion, Hyperparameter Tuning=individually tuned2025.11112.7
XQLDomain=Gym Locomotion, Hyperparameter Tuning=individually tuned2025.11112.7
FQL2025.10111.7
MXQLDomain=Gym Locomotion, Hyperparameter Tuning=individually tuned2025.11111.2
SSCQL2025.10111.1
QIPO-OT2025.10110.9
CQLImplementation=Original Paper2021.10110.1
PBRL2022.06110.1
TD3+BCDomain=Gym Locomotion, Hyperparameter Tuning=individually tuned2025.11110.1
QIPO-Diff2025.10110.1
CQLImplementation=Reproduced2021.10109.1
CQL2022.06109.1
CQL2021.06108.8
CQLDomain=Gym Locomotion, Hyperparameter Tuning=individually tuned2025.11108.8
Decision Transformer2021.06108.1
BCDomain=Gym Locomotion, Hyperparameter Tuning=individually tuned2025.11107.5
Trajectory Transformerdiscretization=quantile2021.06101.9
CQL-Min2024.0198.7
Trajectory Transformerdiscretization=uniform2021.0691
BCImplementation=Reproduced2021.1090.1
BC2022.0690.1
BC2024.0190.1
BRAC2021.0681.6
MBOP2021.0670.2
SACImplementation=Reproduced2021.1068.8
UWACImplementation=Reproduced2021.1055.2
Behavior Cloning2021.0636.6
REMImplementation=Reproduced2021.1027.9
BRACImplementation=Reproduced2021.1016
MORELImplementation=Reproduced2021.101.5
BEARImplementation=Reproduced2021.101.3
BCQImplementation=Reproduced2021.10-0.2