Online Class-Incremental Learning on Mini-ImageNet Split
17.08Final Average AccuracyER-ACE w/ Ours
Evaluation Results
| Method | Links | |
|---|---|---|
| ER-ACE w/ OursMemory size |M|=5000, Batch size |B|=102023.10 | 17.08 | |
| ER-ACE w/ OursMemory size |M|=2000, Batch size |B|=102023.10 | 16.41 | |
| ER-ACE w/ CNMemory size |M|=2000, Batch size |B|=102023.10 | 15.76 | |
| ER-ACE w/ CNMemory size |M|=5000, Batch size |B|=102023.10 | 15.39 | |
| ER-ACE w/ BNMemory size |M|=5000, Batch size |B|=102023.10 | 14.53 | |
| ER-ACE w/ BNMemory size |M|=2000, Batch size |B|=102023.10 | 14.41 | |
| DER++ w/ OursMemory size |M|=5000, Batch size |B|=102023.10 | 10.85 | |
| DER++ w/ OursMemory size |M|=2000, Batch size |B|=102023.10 | 10.8 | |
| DER++ w/ CNMemory size |M|=2000, Batch size |B|=102023.10 | 10.16 | |
| DER++ w/ CNMemory size |M|=5000, Batch size |B|=102023.10 | 8.94 | |
| DER++ w/ BNMemory size |M|=2000, Batch size |B|=102023.10 | 7.04 | |
| DER++ w/ BNMemory size |M|=5000, Batch size |B|=102023.10 | 6.59 | |
| ER-ACE w/ GNMemory size |M|=5000, Batch size |B|=102023.10 | 6.29 | |
| ER-ACE w/ LNMemory size |M|=2000, Batch size |B|=102023.10 | 6.28 | |
| ER-ACE w/ LNMemory size |M|=5000, Batch size |B|=102023.10 | 6.21 | |
| ER-ACE w/ GNMemory size |M|=2000, Batch size |B|=102023.10 | 5.33 | |
| ER-ACE w/ INMemory size |M|=5000, Batch size |B|=102023.10 | 5.13 | |
| ER-ACE w/ INMemory size |M|=2000, Batch size |B|=102023.10 | 5 | |
| DER++ w/ LNMemory size |M|=2000, Batch size |B|=102023.10 | 2.94 | |
| DER++ w/ GNMemory size |M|=2000, Batch size |B|=102023.10 | 2.8 | |
| DER++ w/ GNMemory size |M|=5000, Batch size |B|=102023.10 | 2.72 | |
| DER++ w/ INMemory size |M|=5000, Batch size |B|=102023.10 | 2.33 | |
| DER++ w/ LNMemory size |M|=5000, Batch size |B|=102023.10 | 2.21 | |
| DER++ w/ INMemory size |M|=2000, Batch size |B|=102023.10 | 1.88 |