List your company
ResearchFormerly WizWand

Offline Reinforcement Learning on D4RL MuJoCo halfcheetah-medium-expert

104.5Normalized ScoreROMI
40.33256.99173.6590.309Feb 20, 2026Mar 10, 2026Mar 28, 2026Apr 15, 2026May 3, 2026May 21, 2026Jun 9, 2026
Updated 4mo ago

Evaluation results

MethodLinks
ROMItraining steps=1M, seeds=5, version=v22026.03104.5
FACPolicy Type=Flow, Seeds=82026.02101.9
ReBRACPolicy Type=Gaussian, Seeds=82026.02101.1
COUNTtraining steps=1M, seeds=5, version=v22026.03100
FQLPolicy Type=One-Step Flow/Diffusion Policy2026.0699.8
FQLPolicy Type=Flow, Seeds=82026.0299.6
BFQPolicy Type=One-Step Flow/Diffusion Policy2026.0698.6
SPAR-PROJLearning Paradigm=Ours2026.0597
Diff-QLLearning Paradigm=Policy Gradient Guidance2026.0596.8
SORLPolicy Type=One-Step Flow/Diffusion Policy2026.0696.5
MOBILEtraining steps=1M, seeds=5, version=v22026.0396.1
IDQLPolicy Type=Diffusion, Seeds=82026.0295.9
EDPPolicy Type=Diffusion Policy2026.0695.8
EPQPolicy Type=Gaussian, Seeds=82026.0295.7
ISEP-FMFlow Matching=true2026.0595.7
DQLPolicy Type=Diffusion Policy2026.0695.5
OFQLPolicy Type=One-Step Flow/Diffusion Policy2026.0695.2
QIPO-OTPolicy Type=Diffusion, Seeds=82026.0294.5
DQL2026.0594.4
QIPO2026.0594.4
IDQLLearning Paradigm=In-Support Learning2026.0594.4
PLASLearning Paradigm=Policy Gradient Guidance2026.0594.3
LAPOLearning Paradigm=In-Support Learning2026.0594.2
RAMBOtraining steps=1M, seeds=5, version=v22026.0393.9
QGPO2026.0593.5
ISEPFlow Matching=false2026.0592.4
MOPOtraining steps=1M, seeds=5, version=v22026.0392.3
SRPOPolicy Type=Diffusion, Seeds=82026.0292.2
SRPOPolicy Type=One-Step Flow/Diffusion Policy2026.0692.2
CQLPolicy Type=Gaussian, Seeds=82026.0291.6
CQLLearning Paradigm=Policy Gradient Guidance2026.0591.6
TD3-BCPolicy Type=Gaussian Policy2026.0691.5
IDQLPolicy Type=Diffusion Policy2026.0691.3
TD3+BCPolicy Type=Gaussian, Seeds=82026.0290.7
TD3+BCLearning Paradigm=Policy Gradient Guidance2026.0590.7
EQLLearning Paradigm=In-Support Learning2026.0590.6
IDQL2026.0590.3
BCQLearning Paradigm=Policy Gradient Guidance2026.0589.1
IQLPolicy Type=Gaussian Policy2026.0688.3
MCQPolicy Type=Gaussian, Seeds=82026.0287.5
SPOTPolicy Type=Gaussian, Seeds=82026.0286.9
IQLPolicy Type=Gaussian, Seeds=82026.0286.7
IQL2026.0586.7
IQLLearning Paradigm=In-Support Learning2026.0586.7
IQLtraining steps=1M, seeds=5, version=v22026.0386.3
SPAR-MLPLearning Paradigm=Ours2026.0585.2
CACPolicy Type=Diffusion, Seeds=82026.0284.3
CQL-AWLearning Paradigm=In-Support Learning2026.0584
CQLtraining steps=1M, seeds=5, version=v22026.0375.6
BaseLearning Paradigm=Ours2026.0563.1
CQL2026.0562.4
BCPolicy Type=Gaussian Policy2026.0660.2
BC2026.0555.2
AWACLearning Paradigm=In-Support Learning2026.0542.8