List your company
ResearchFormerly WizWand

Reinforcement Learning on HalfCheetah (Average Return)

7,223.53Average ReturnDiff-SR
-2,027.8212373.97192,775.7655,177.5581Jun 23, 2024Oct 8, 2024Jan 24, 2025May 11, 2025Aug 27, 2025Dec 12, 2025Mar 30, 2026
Updated 7mo ago

Evaluation results

MethodLinks
Diff-SRRL Category=Representation RL, Steps=200K, Seeds=42024.067,223.53
LV-RepRL Category=Representation RL, Steps=200K, Seeds=42024.065,557.6
SACNumber of training seeds=5, Number of evaluation episodes=1002026.034,988.5
SPEDERL Category=Representation RL, Steps=200K, Seeds=42024.064,210.3
DeepSFRL Category=Representation RL, Steps=200K, Seeds=42024.064,180.4
SACRL Category=Model-Free RL, Layers=3-layer, Steps=200K, Seeds=42024.064,000.7
Best MBBLRL Category=Model-Based RL, Steps=200K, Seeds=42024.063,639
SMACbatch size=1000, seeds=52026.012,936
PETS-CEMRL Category=Model-Based RL, Steps=200K, Seeds=42024.062,795.3
PolyGRADRL Category=Model-Based RL, Steps=200K, Seeds=42024.062,563.5
CG-FPDNumber of training seeds=5, Number of evaluation episodes=1002026.032,407.88
ME-TRPORL Category=Model-Based RL, Steps=200K, Seeds=42024.062,283.7
DF-CWP-CPNumber of training seeds=5, Number of evaluation episodes=1002026.032,103.8
AC-CGbatch size=1000, seeds=52026.011,766
PPONumber of training seeds=5, Number of evaluation episodes=1002026.031,579.13
PETS-RSRL Category=Model-Based RL, Steps=200K, Seeds=42024.06966.9
A2CNumber of training seeds=5, Number of evaluation episodes=1002026.03795.91
AC-KFACbatch size=1000, seeds=52026.01513
PPORL Category=Model-Free RL, Steps=200K, Seeds=42024.0617.2
TRPORL Category=Model-Free RL, Steps=200K, Seeds=42024.06-12
AC-Adambatch size=1000, seeds=52026.01-646
AC-SGDbatch size=1000, seeds=52026.01-1,672