List your company
ResearchFormerly WizWand

Offline Reinforcement Learning on D4RL HalfCheetah Med-Replay

72.1Normalized Avg ReturnMOPO
35.148844.741954.33563.9281Oct 30, 2023Feb 22, 2024Jun 17, 2024Oct 11, 2024Feb 4, 2025May 31, 2025Sep 24, 2025
Updated 5mo ago

Evaluation results

MethodLinks
MOPOImplementation=Original Paper2024.1172.1
MOBILEImplementation=Original Paper2024.1171.7
MOPOImplementation=Reproduced2024.1166.3
MOBILEImplementation=Reproduced2024.1163.7
C-LAPImplementation=Reproduced2024.1155.5
MOPO2023.1053.1
FQL+BCTraining timesteps=1M, Random seeds=102025.0946.23
CQL2023.1045.5
CQL2025.0945.5
IQL2023.1044.2
IQL2025.0944.2
PLASImplementation=Original Paper2024.1143.9
IQLTraining timesteps=1M, Random seeds=102025.0943.44
MBOP2023.1042.3
Diffuser2023.1042.2
TT2023.1041.9
TT2025.0941.9
PLASImplementation=Reproduced2024.1141.8
RGG+planning seeds=152023.1041.3
RGGplanning seeds=152023.1041
MOREL2023.1040.2
MOReL2025.0940.2
Decision Diffuser2025.0939.3
WFDiffuser2025.0938.1
Aaren2024.0537.91
BC2023.1036.6
DT2023.1036.6
BC2025.0936.6
DT2025.0936.6
Transformer2024.0536.57