Continual Reinforcement Learning on Meta-World MT50 v2
81.7APAMC
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| AMCMemory (MB)=3802026.04 | 81.7 | 15.2 | 6.3 | 0.215 | |
| PNNMemory (MB)=50002026.04 | 76.4 | 11.3 | 2.1 | 0.015 | |
| PackNetMemory (MB)=50002026.04 | 73.8 | 9.2 | 8.4 | 0.015 | |
| HERMemory (MB)=10002026.04 | 71.2 | 8.4 | 21.3 | 0.071 | |
| EWCMemory (MB)=10002026.04 | 69.1 | 6.8 | 19.7 | 0.069 | |
| PERMemory (MB)=10002026.04 | 68.7 | 5.7 | 24.1 | 0.069 | |
| NECMemory (MB)=15002026.04 | 64.5 | 4.1 | 28.7 | 0.043 | |
| EWCMemory (MB)=380, Memory Matched=true2026.04 | 64.2 | 5.9 | 22.1 | 0.169 | |
| PERMemory (MB)=380, Memory Matched=true2026.04 | 63.6 | 4.8 | 26.3 | 0.167 | |
| VRMemory (MB)=10002026.04 | 62.3 | 3.2 | 31.4 | 0.062 | |
| VRMemory (MB)=380, Memory Matched=true2026.04 | 55.8 | 2.1 | 34.2 | 0.147 |