List your company
ResearchFormerly WizWand

Offline Reinforcement Learning on Stochastic-D4RL HalfCheetah medium-replay

2,696.61Mean ReturnRAFMAC
-148.2684590.30581,328.882,067.4542Oct 3, 2025
Updated 4mo ago

Evaluation results

MethodLinks
RAFMACEvaluation Steps=1000, Seeds=52025.102,696.611,499.8
RADACEvaluation Steps=1000, Seeds=52025.102,674.721,401.03
DiffusionQLEvaluation Steps=1000, Seeds=52025.102,261.161,439.77
FQLEvaluation Steps=1000, Seeds=52025.101,909.57568.43
RADAC2025.10525.84278.65
FQLinstantiation=Flow2025.10434.33224.73
ORAACEvaluation Steps=1000, Seeds=52025.10315.87161.54
DiffusionQL2025.10279.9579.93
ORAAC-Diff.instantiation=Diffusion2025.1022044.93
CQL2025.1066.21127.09
ORAAC2025.1018.9934.09
CODAC2025.100.111.47
CODACEvaluation Steps=1000, Seeds=52025.10-0.12-1.48
CQLEvaluation Steps=1000, Seeds=52025.10-38.85-40.23