List your company
ResearchFormerly WizWand

Multi-Agent Reinforcement Learning on SMAC 3s5z v1

99.4Normalized Win RateMAST
31.38449.04266.784.358Sep 28, 2024
Updated 8mo ago

Evaluation results

MethodLinks
MASTAlg.=RES, Sparsity=85%, Total Size=0.158x, FLOPs (Train)=0.154x, FLOPs (Test)=0.150x2024.0999.4
MASTAlg.=Q-MIX, Sparsity=90%, Total Size=0.109x, FLOPs (Train)=0.101x, FLOPs (Test)=0.100x2024.0999
MASTAlg.=WQ-MIX, Sparsity=90%, Total Size=0.105x, FLOPs (Train)=0.100x, FLOPs (Test)=0.100x2024.0996.1
TinyAlg.=RES, Sparsity=85%, Total Size=0.158x, FLOPs (Train)=0.154x, FLOPs (Test)=0.150x2024.0995.1
RigLAlg.=RES, Sparsity=85%, Total Size=0.158x, FLOPs (Train)=0.154x, FLOPs (Test)=0.150x2024.0992.8
SETAlg.=RES, Sparsity=85%, Total Size=0.158x, FLOPs (Train)=0.154x, FLOPs (Test)=0.150x2024.0990.3
SSAlg.=RES, Sparsity=85%, Total Size=0.158x, FLOPs (Train)=0.154x, FLOPs (Test)=0.150x2024.0989
RLx2Alg.=RES, Sparsity=85%, Total Size=0.158x, FLOPs (Train)=0.154x, FLOPs (Test)=0.150x2024.0986.2
TinyAlg.=WQ-MIX, Sparsity=90%, Total Size=0.105x, FLOPs (Train)=0.100x, FLOPs (Test)=0.100x2024.0970.7
TinyAlg.=Q-MIX, Sparsity=90%, Total Size=0.109x, FLOPs (Train)=0.101x, FLOPs (Test)=0.100x2024.0968.2
SSAlg.=WQ-MIX, Sparsity=90%, Total Size=0.105x, FLOPs (Train)=0.100x, FLOPs (Test)=0.100x2024.0962.5
RLx2Alg.=WQ-MIX, Sparsity=90%, Total Size=0.105x, FLOPs (Train)=0.100x, FLOPs (Test)=0.100x2024.0960.7
SETAlg.=WQ-MIX, Sparsity=90%, Total Size=0.105x, FLOPs (Train)=0.100x, FLOPs (Test)=0.100x2024.0956
SETAlg.=Q-MIX, Sparsity=90%, Total Size=0.109x, FLOPs (Train)=0.101x, FLOPs (Test)=0.100x2024.0952.3
RigLAlg.=WQ-MIX, Sparsity=90%, Total Size=0.105x, FLOPs (Train)=0.100x, FLOPs (Test)=0.100x2024.0950.4
RLx2Alg.=Q-MIX, Sparsity=90%, Total Size=0.109x, FLOPs (Train)=0.101x, FLOPs (Test)=0.100x2024.0950.1
RigLAlg.=Q-MIX, Sparsity=90%, Total Size=0.109x, FLOPs (Train)=0.101x, FLOPs (Test)=0.100x2024.0945.2
SSAlg.=Q-MIX, Sparsity=90%, Total Size=0.109x, FLOPs (Train)=0.101x, FLOPs (Test)=0.100x2024.0934