Incremental Learning on Tiny-ImageNet 10 task
55.25Average Incremental AccuracyUpper bound
Evaluation results
| Method | Links | |
|---|---|---|
| Upper boundModel Inversion=false, Backbone=ResNet-32, Number of runs=52026.05 | 55.25 | |
| REMIXModel Inversion=true, Backbone=ResNet-32, Number of runs=52026.05 | 33.39 | |
| PMIModel Inversion=true, Backbone=ResNet-32, Number of runs=52026.05 | 32.43 | |
| PMI w/o CFSModel Inversion=true, Backbone=ResNet-32, Number of runs=52026.05 | 32.09 | |
| R-DFCILModel Inversion=true, Backbone=ResNet-32, Number of runs=52026.05 | 29.05 | |
| ABDModel Inversion=true, Backbone=ResNet-32, Number of runs=52026.05 | 23.17 |