Sequential Backward-Compatible Learning on RParis + ROxford + GLDv2 (test)
50.87PupNDPP
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| NDPP9% Data=phi1, 30% Data=phi2, Backbone=ResNet-182025.03 | 50.87 | 59.44 | 54.8 | |
| ODPP9% Data=phi1, 100% Data=phi3, Backbone=ResNet-182025.03 | 50.69 | 57.27 | 53.77 | |
| ODPP30% Data=phi2, 100% Data=phi3, Backbone=ResNet-182025.03 | 50.69 | 55.54 | 52.84 | |
| ODPP9% Data=phi1, 30% Data=phi2, Backbone=ResNet-182025.03 | 50.55 | 58.75 | 54.32 | |
| NDPP9% Data=phi1, 100% Data=phi3, Backbone=ResNet-182025.03 | 50.38 | 56.11 | 53.09 | |
| NDPP30% Data=phi2, 100% Data=phi3, Backbone=ResNet-182025.03 | 50.38 | 53.58 | 51.78 | |
| AdvBCT9% Data=phi1, 100% Data=phi3, Backbone=ResNet-182025.03 | 50.16 | 53.36 | 51.71 | |
| AdvBCT30% Data=phi2, 100% Data=phi3, Backbone=ResNet-182025.03 | 50.16 | 47.48 | 48.76 | |
| AdvBCT9% Data=phi1, 30% Data=phi2, Backbone=ResNet-182025.03 | 49.3 | 53.23 | 51.19 | |
| BCT9% Data=phi1, 100% Data=phi3, Backbone=ResNet-182025.03 | 48.41 | 55.06 | 51.52 | |
| BCT30% Data=phi2, 100% Data=phi3, Backbone=ResNet-182025.03 | 48.41 | 43.31 | 45.59 | |
| BCT9% Data=phi1, 30% Data=phi2, Backbone=ResNet-182025.03 | 47.75 | 55.34 | 51.23 |