Offline Reinforcement Learning on D4RL MuJoCo hopper-medium-expert
112.3Normalized ScoreISEP-FM
Evaluation results
| Method | Links | |
|---|---|---|
| ISEP-FMFlow Matching=true2026.05 | 112.3 | |
| QIPO2026.05 | 112.1 | |
| ISEPFlow Matching=false2026.05 | 112 | |
| IDQL2026.05 | 111.9 | |
| MCQPolicy Type=Gaussian, Seeds=82026.02 | 111.2 | |
| Diff-QLLearning Paradigm=Policy Gradient Guidance2026.05 | 111.1 | |
| DQLPolicy Type=Diffusion Policy2026.06 | 111.1 | |
| CQL2026.05 | 111 | |
| LAPOLearning Paradigm=In-Support Learning2026.05 | 111 | |
| EDPPolicy Type=Diffusion Policy2026.06 | 110.8 | |
| BFQPolicy Type=One-Step Flow/Diffusion Policy2026.06 | 110.5 | |
| OFQLPolicy Type=One-Step Flow/Diffusion Policy2026.06 | 110.2 | |
| IDQLPolicy Type=Diffusion Policy2026.06 | 110.1 | |
| EPQPolicy Type=Gaussian, Seeds=82026.02 | 108.8 | |
| SPAR-PROJLearning Paradigm=Ours2026.05 | 108.7 | |
| IDQLPolicy Type=Diffusion, Seeds=82026.02 | 108.6 | |
| QIPO-OTPolicy Type=Diffusion, Seeds=82026.02 | 108 | |
| QGPO2026.05 | 108 | |
| ReBRACPolicy Type=Gaussian, Seeds=82026.02 | 107 | |
| EQLLearning Paradigm=In-Support Learning2026.05 | 105.5 | |
| CQLPolicy Type=Gaussian, Seeds=82026.02 | 105.4 | |
| CQLLearning Paradigm=Policy Gradient Guidance2026.05 | 105.4 | |
| DQL2026.05 | 105.3 | |
| IDQLLearning Paradigm=In-Support Learning2026.05 | 105.3 | |
| FACPolicy Type=Flow, Seeds=82026.02 | 104.2 | |
| TD3-BCPolicy Type=Gaussian Policy2026.06 | 101.6 | |
| CACPolicy Type=Diffusion, Seeds=82026.02 | 100.4 | |
| SRPOPolicy Type=Diffusion, Seeds=82026.02 | 100.1 | |
| SRPOPolicy Type=One-Step Flow/Diffusion Policy2026.06 | 100.1 | |
| SPOTPolicy Type=Gaussian, Seeds=82026.02 | 99.3 | |
| TD3+BCPolicy Type=Gaussian, Seeds=82026.02 | 98 | |
| TD3+BCLearning Paradigm=Policy Gradient Guidance2026.05 | 98 | |
| PLASLearning Paradigm=Policy Gradient Guidance2026.05 | 94.7 | |
| IQLPolicy Type=Gaussian, Seeds=82026.02 | 91.5 | |
| IQL2026.05 | 91.5 | |
| IQLLearning Paradigm=In-Support Learning2026.05 | 91.5 | |
| CQL-AWLearning Paradigm=In-Support Learning2026.05 | 91 | |
| FQLPolicy Type=One-Step Flow/Diffusion Policy2026.06 | 86.2 | |
| FQLPolicy Type=Flow, Seeds=82026.02 | 83.1 | |
| BCQLearning Paradigm=Policy Gradient Guidance2026.05 | 81.8 | |
| IQLPolicy Type=Gaussian Policy2026.06 | 76.6 | |
| BCPolicy Type=Gaussian Policy2026.06 | 67.2 | |
| AWACLearning Paradigm=In-Support Learning2026.05 | 55.8 | |
| BaseLearning Paradigm=Ours2026.05 | 55.2 | |
| BC2026.05 | 52.5 | |
| SORLPolicy Type=One-Step Flow/Diffusion Policy2026.06 | 45.9 | |
| SPAR-MLPLearning Paradigm=Ours2026.05 | 1.1 |