List your company
ResearchFormerly WizWand

Image Classification on CIFAR-100 (test) (Accuracy vs FLOPs Budget)

0.7071Top-1 Accuracy (25% FLOPs)LwF
0.5184440.5674220.61640.665378May 6, 2026
Updated 6mo ago

Evaluation results

MethodLinks
LwFBackbone=ResNet-34, lambda=100, rho=0.72026.050.70710.73840.73610.7355
EWCBackbone=ResNet-34, lambda=100, rho=0.72026.050.7050.74070.7390.7384
LwFBackbone=MSDNet, lambda=1000, rho=0.22026.050.6650.73040.73330.732
SeparateBackbone=ResNet-342026.050.66250.72570.72870.7282
JointBackbone=MSDNet2026.050.65930.72020.74730.7586
EWCBackbone=MSDNet, lambda=1000, rho=0.22026.050.6510.71030.71980.7189
SeparateBackbone=MSDNet2026.050.6490.67840.69020.6921
Branch-wiseBackbone=ResNet-342026.050.64190.67030.66880.6675
JointBackbone=ResNet-342026.050.62840.7280.74320.7417
Branch-wiseBackbone=MSDNet2026.050.58240.59730.60160.5997
DisjointBackbone=MSDNet2026.050.56740.63960.68590.7036
DisjointBackbone=ResNet-342026.050.52570.67560.73490.7379