Domain-incremental Learning on DN4IL (test)
78.4Last AccuracyMIRA
Evaluation results
| Method | Links | |
|---|---|---|
| MIRABuffer size=0 (Exemplar-free), Method Type=Multi-adapter2025.11 | 78.4 | |
| DUCABuffer size=200 exemplars, Method Type=Multi-model2025.11 | 44.45 | |
| DARE++Buffer size=200 exemplars, Method Type=Multi-model2025.11 | 44.11 | |
| CLS-ERBuffer size=200 exemplars, Method Type=Multi-model2025.11 | 41.7 | |
| DAREBuffer size=200 exemplars, Method Type=Single-model2025.11 | 40.59 | |
| DER++Buffer size=200 exemplars, Method Type=Single-model2025.11 | 35.74 | |
| ERBuffer size=200 exemplars, Method Type=Single-model2025.11 | 27.45 |