Class-incremental Continual Learning on CIFAR-10 Split (test)
92.2Average Accuracy (AT)MULTITASK
Evaluation results
| Method | Links | ||
|---|---|---|---|
| MULTITASKProtocol=Standard Continual Learning, Replay-buffer (B)=false2023.06 | 92.2 | — | |
| KD w/ BUFFERProtocol=Standard Continual Learning, Replay-buffer (B)=true2023.06 | 70.34 | 16.32 | |
| DER++Protocol=Standard Continual Learning, Replay-buffer (B)=true2023.06 | 67.23 | 26.13 | |
| DERProtocol=Standard Continual Learning, Replay-buffer (B)=true2023.06 | 66.08 | 27.4 | |
| KD w/ BUFFERProtocol=Heterogeneous Continual Learning, Replay-buffer (B)=true2023.06 | 65.4 | 10.13 | |
| iCaRLProtocol=Standard Continual Learning, Replay-buffer (B)=true2023.06 | 49.02 | 28.72 | |
| DER++Protocol=Heterogeneous Continual Learning, Replay-buffer (B)=true2023.06 | 48.82 | 50.11 | |
| ERProtocol=Standard Continual Learning, Replay-buffer (B)=true2023.06 | 48.56 | 58.11 | |
| DERProtocol=Heterogeneous Continual Learning, Replay-buffer (B)=true2023.06 | 44.13 | 57.64 | |
| ERProtocol=Heterogeneous Continual Learning, Replay-buffer (B)=true2023.06 | 38.77 | 69.61 | |
| KD w/ QDIProtocol=Heterogeneous Continual Learning, Replay-buffer (B)=false2023.06 | 33.89 | 31.73 | |
| KDProtocol=Heterogeneous Continual Learning, Replay-buffer (B)=false2023.06 | 30.21 | 27.54 | |
| KDProtocol=Standard Continual Learning, Replay-buffer (B)=false2023.06 | 22.73 | 87 | |
| DIProtocol=Standard Continual Learning, Replay-buffer (B)=false2023.06 | 22.72 | 77.65 | |
| A-GEMProtocol=Standard Continual Learning, Replay-buffer (B)=true2023.06 | 21.98 | 92.18 | |
| FINETUNEProtocol=Heterogeneous Continual Learning, Replay-buffer (B)=false2023.06 | 21.45 | 91.24 | |
| DIProtocol=Heterogeneous Continual Learning, Replay-buffer (B)=false2023.06 | 20.67 | 70.97 | |
| KD w/ QDIProtocol=Standard Continual Learning, Replay-buffer (B)=false2023.06 | 19.75 | 0.08 | |
| A-GEMProtocol=Heterogeneous Continual Learning, Replay-buffer (B)=true2023.06 | 19.67 | 89.26 | |
| LwFProtocol=Standard Continual Learning, Replay-buffer (B)=false2023.06 | 19.61 | 96.69 | |
| FINETUNEProtocol=Standard Continual Learning, Replay-buffer (B)=false2023.06 | 19.6 | 96.66 | |
| SIProtocol=Standard Continual Learning, Replay-buffer (B)=false2023.06 | 19.48 | 95.78 |