List your company
ResearchFormerly WizWand

Offline Reinforcement Learning on D4RL MuJoCo halfcheetah-medium-replay

0.62Normalized ScoreEPQ
0.336080.409790.48350.55721Feb 20, 2026Mar 10, 2026Mar 28, 2026Apr 15, 2026May 3, 2026May 21, 2026Jun 9, 2026
Updated 4mo ago

Evaluation results

MethodLinks
EPQPolicy Type=Gaussian, Seeds=82026.020.62
CACPolicy Type=Diffusion, Seeds=82026.020.587
MCQPolicy Type=Gaussian, Seeds=82026.020.568
FACPolicy Type=Flow, Seeds=82026.020.554
ISEP-FMFlow Matching=true2026.050.546
SPOTPolicy Type=Gaussian, Seeds=82026.020.522
BFQPolicy Type=One-Step Flow/Diffusion Policy2026.060.521
SRPOPolicy Type=Diffusion, Seeds=82026.020.514
SRPOPolicy Type=One-Step Flow/Diffusion Policy2026.060.514
OFQLPolicy Type=One-Step Flow/Diffusion Policy2026.060.512
FQLPolicy Type=One-Step Flow/Diffusion Policy2026.060.511
ReBRACPolicy Type=Gaussian, Seeds=82026.020.51
SPAR-MLPLearning Paradigm=Ours2026.050.509
FQLPolicy Type=Flow, Seeds=82026.020.493
CQL-AWLearning Paradigm=In-Support Learning2026.050.49
SORLPolicy Type=One-Step Flow/Diffusion Policy2026.060.483
QIPO-OTPolicy Type=Diffusion, Seeds=82026.020.48
QIPO2026.050.48
DQLPolicy Type=Diffusion Policy2026.060.479
DQL2026.050.478
Diff-QLLearning Paradigm=Policy Gradient Guidance2026.050.478
QGPO2026.050.476
EQLLearning Paradigm=In-Support Learning2026.050.472
IDQLPolicy Type=Diffusion Policy2026.060.465
ISEPFlow Matching=false2026.050.463
IDQLPolicy Type=Diffusion, Seeds=82026.020.459
PLASLearning Paradigm=Policy Gradient Guidance2026.050.457
CQLPolicy Type=Gaussian, Seeds=82026.020.455
CQL2026.050.455
CQLLearning Paradigm=Policy Gradient Guidance2026.050.455
IDQL2026.050.451
IDQLLearning Paradigm=In-Support Learning2026.050.451
EDPPolicy Type=Diffusion Policy2026.060.449
TD3+BCPolicy Type=Gaussian, Seeds=82026.020.446
TD3+BCLearning Paradigm=Policy Gradient Guidance2026.050.446
IQL2026.050.445
TD3-BCPolicy Type=Gaussian Policy2026.060.445
IQLPolicy Type=Gaussian, Seeds=82026.020.442
IQLLearning Paradigm=In-Support Learning2026.050.442
SPAR-PROJLearning Paradigm=Ours2026.050.436
IQLPolicy Type=Gaussian Policy2026.060.429
LAPOLearning Paradigm=In-Support Learning2026.050.419
AWACLearning Paradigm=In-Support Learning2026.050.405
BaseLearning Paradigm=Ours2026.050.405
BCQLearning Paradigm=Policy Gradient Guidance2026.050.404
BC2026.050.366
BCPolicy Type=Gaussian Policy2026.060.347