Multi-agent Reinforcement Learning on SMAC 25m v1 (test)
98.3Test Win RateCommFormer
Evaluation Results
| Method | Links | |
|---|---|---|
| CommFormerEnvironment Type=homo. control, Training Budget=2M2026.05 | 98.3 | |
| QMIXEnvironment Type=homo. control, Training Budget=2M2026.05 | 98.1 | |
| MAGIEnvironment Type=homo. control, Training Budget=2M2026.05 | 97.5 | |
| HIBCGEnvironment Type=homo. control, Training Budget=2M2026.05 | 97.2 | |
| HIB-flatEnvironment Type=homo. control, Training Budget=2M2026.05 | 96.5 | |
| GACGEnvironment Type=homo. control, Training Budget=2M2026.05 | 96 | |
| BVMEEnvironment Type=homo. control, Training Budget=2M2026.05 | 95.9 | |
| AIB-onlyEnvironment Type=homo. control, Training Budget=2M2026.05 | 94.3 | |
| ExpoCommEnvironment Type=homo. control, Training Budget=2M2026.05 | 91.5 |