List your company
ResearchFormerly WizWand

Class-incremental Learning on Tiny ImageNet i-Blurry-50-10 (test)

53.05Average AccuracyStandard
15.599625.322335.04544.7677Jul 31, 2025
Updated 4mo ago

Evaluation results

MethodLinks
Standard2025.0753.05
EB (w/ DMA)Number of tasks (t)=5, Pre-trained initialization (PR)=ImageNet, i-Blurry setting (n/m)=50/102025.0747.68
EB (w/o DMA)Number of tasks (t)=5, Pre-trained initialization (PR)=ImageNet, i-Blurry setting (n/m)=50/102025.0740.21
FOSTERNumber of tasks (t)=5, Pre-trained initialization (PR)=ImageNet, i-Blurry setting (n/m)=50/102025.0733.93
iCaRLNumber of tasks (t)=5, Pre-trained initialization (PR)=ImageNet, i-Blurry setting (n/m)=50/102025.0725.41
CLIBNumber of tasks (t)=5, Pre-trained initialization (PR)=ImageNet, i-Blurry setting (n/m)=50/102025.0725.05
BiCNumber of tasks (t)=5, Pre-trained initialization (PR)=ImageNet, i-Blurry setting (n/m)=50/102025.0724.9
FTFNumber of tasks (t)=5, Pre-trained initialization (PR)=ImageNet, i-Blurry setting (n/m)=50/102025.0724.58
ER-MIRNumber of tasks (t)=5, Pre-trained initialization (PR)=ImageNet, i-Blurry setting (n/m)=50/102025.0724.54
EWC++Number of tasks (t)=5, Pre-trained initialization (PR)=ImageNet, i-Blurry setting (n/m)=50/102025.0724.39
GDumbNumber of tasks (t)=5, Pre-trained initialization (PR)=ImageNet, i-Blurry setting (n/m)=50/102025.0718.69
RMNumber of tasks (t)=5, Pre-trained initialization (PR)=ImageNet, i-Blurry setting (n/m)=50/102025.0717.04