Domain-incremental learning on PACS
90.24APFOGO
Evaluation results
| Method | Links | ||
|---|---|---|---|
| FOGORegime=Replay, CL Method=ER, Memory=Buffer + Centroids, Mem. (MB)=58.0, Time (m)=10.442026.06 | 90.24 | 4.21 | |
| MuonRegime=Replay, CL Method=ER, Memory=Buffer, Mem. (MB)=57.4, Time (m)=8.882026.06 | 89.2 | 7.56 | |
| FOGORegime=Free, CL Method=None, Memory=Centroids, Mem. (MB)=0.55, Time (m)=9.842026.06 | 87.89 | 6.77 | |
| MuonRegime=Free, CL Method=None, Memory=None, Mem. (MB)=< 0.1, Time (m)=8.152026.06 | 87.81 | 9.44 | |
| AdamRegime=Replay, CL Method=ER, Memory=Buffer, Mem. (MB)=57.4, Time (m)=8.212026.06 | 87.5 | 9.81 | |
| CBPRegime=Replay, CL Method=ER, Memory=Buffer, Mem. (MB)=57.4, Time (m)=13.602026.06 | 84.77 | 13.01 | |
| AdamRegime=Free, CL Method=None, Memory=None, Mem. (MB)=< 0.1, Time (m)=7.492026.06 | 83.34 | 15.45 | |
| FIRERegime=Replay, CL Method=ER, Memory=Buffer, Mem. (MB)=57.4, Time (m)=11.522026.06 | 82.79 | 13.36 | |
| CBPRegime=Free, CL Method=None, Memory=None, Mem. (MB)=< 0.1, Time (m)=12.102026.06 | 82.78 | 15.94 | |
| FIRERegime=Free, CL Method=None, Memory=None, Mem. (MB)=< 0.1, Time (m)=11.142026.06 | 75.94 | 20.98 |