MaMuJoCo
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Multi-Agent Reinforcement Learning | MAMuJoCo HalfCheetah Extreme Env v2 (various) | Average Return2,757.2 | 24 | |
| Multi-Agent Reinforcement Learning | MAMuJoCo HalfCheetah Random Env v2 (various) | Average Return3,178.7 | 24 | |
| Multi-agent Continual Cooperation | MAMuJoCo Reward (Expert) | Forward Transfer36.48 | 14 | |
| Multi-Agent Reinforcement Learning | MAMuJoCo Walker2d 6x1 (test) | Average Episodic Return1,116.75 | 13 | |
| Multi-agent Reinforcement Learning | MaMuJoCo OMIGA 2-Ant (Medium) | Average Episode Reward1,798.05 | 11 | |
| Multi-agent Reinforcement Learning | MaMuJoCo OMIGA 2-Ant (Medium-Replay) | Average Episode Reward1,822.72 | 11 | |
| Multi-agent Reinforcement Learning | MaMuJoCo OMIGA 2-Ant (Medium-Expert) | Average Episode Reward2,002 | 11 | |
| Multi-agent Reinforcement Learning | MaMuJoCo OMIGA 3-Hopper (Medium) | Average Episode Reward3,360 | 11 | |
| Multi-agent Reinforcement Learning | MaMuJoCo (OMIGA) 6-HalfCheetah Medium | Average Episode Reward4,695 | 11 | |
| Multi-agent Reinforcement Learning | MaMuJoCo OMIGA 6-HalfCheetah (Medium-Replay) | Avg Episode Reward4,582 | 11 | |
| Multi-agent Reinforcement Learning | MaMuJoCo OMIGA 6-HalfCheetah (Medium-Expert) | Average Episode Reward5,237 | 11 | |
| Multi-agent Reinforcement Learning | MaMuJoCo OMIGA 6-HalfCheetah (Expert) | Average Episode Reward5,545 | 11 | |
| Multi-Agent Offline Reinforcement Learning | MaMuJoCo 2HalfCheetah Medium | Performance4,554.11 | 9 | |
| Dynamics task stream | MAMuJoCo Medium | Backward Transfer2.97 | 8 | |
| Dynamics task stream | MAMuJoCo (Expert) | Backward Transfer11.32 | 8 | |
| Reward task stream | MAMuJoCo Medium | Backward Transfer16.13 | 8 | |
| Reward task stream | MAMuJoCo (Expert) | Backward Transfer21.57 | 8 | |
| Multi-Agent Reinforcement Learning | MAMuJoCo Ant 8x1 (test) | Average Episodic Return45.06 | 8 | |
| Multi-Agent Reinforcement Learning | MAMuJoCo Hopper 3x1 (test) | Average Episodic Return31.02 | 8 | |
| Multi-Agent Reinforcement Learning | MAMuJoCo HalfCheetah 6x1 (test) | Average Episodic Return43.1 | 8 | |
| Multi-agent continuous control | MaMuJoCo HalfCheetah Expert v2 | Score118.5 | 8 | |
| Multi-agent continuous control | MaMuJoCo HalfCheetah v2 (Med-rep) | Score59.5 | 8 | |
| Multi-agent continuous control | MaMuJoCo HalfCheetah v2 (Random) | Score39.7 | 8 | |
| Multi-agent Continual Cooperation | MAMuJoCo Dynamics Medium | Forward Transfer1.26 | 7 | |
| Multi-agent Continual Cooperation | MAMuJoCo Reward Medium | Forward Transfer27.15 | 7 |
Showing 25 of 47 rows