List your company
ResearchFormerly WizWand

Image Classification on CIFAR-100 coarse (test)

97.76Accuracy (Test)Base
-3.910422.484848.8875.2752Sep 10, 2025Oct 27, 2025Dec 14, 2025Jan 31, 2026Mar 20, 2026May 7, 2026Jun 24, 2026
Updated 4mo ago

Evaluation results

MethodLinks
BaseClass=Veg, Backbone=ViT2026.0697.76——95.79
BaseClass=People, Backbone=ViT2026.0697.42——95.81
BaseClass=Elec_Dev, Backbone=ViT2026.0696.78——95.84
BaseClass=Vehicle2, Backbone=ViT2026.0695.42——95.92
BaseClass=Nat_Scen, Backbone=ViT2026.0688.64——96.27
CalexNetForwarding Ratio (FR)=0.4, Batch size=1, Backbone=ResNet-182025.09832.54164—
"no align, no KD" ref.Forwarding Ratio (FR)=0.4, Batch size=1, Backbone=ResNet-182025.09823.19171—
PTEEnetForwarding Ratio (FR)=0.4, Batch size=1, Backbone=ResNet-182025.09801.94124—
ZTWForwarding Ratio (FR)=0.4, Batch size=1, Backbone=ResNet-182025.09792.11118—
CalexNetForwarding Ratio (FR)=0.6, Batch size=1, Backbone=ResNet-182025.09762.19120—
"no align, no KD" ref.Forwarding Ratio (FR)=0.6, Batch size=1, Backbone=ResNet-182025.09742.49142—
PTEEnetForwarding Ratio (FR)=0.6, Batch size=1, Backbone=ResNet-182025.09711.67102—
ZTWForwarding Ratio (FR)=0.6, Batch size=1, Backbone=ResNet-182025.09681.6694—
CalexNetForwarding Ratio (FR)=0.8, Batch size=1, Backbone=ResNet-182025.09591.3183—
"no align, no KD" ref.Forwarding Ratio (FR)=0.8, Batch size=1, Backbone=ResNet-182025.09551.6591—
PTEEnetForwarding Ratio (FR)=0.8, Batch size=1, Backbone=ResNet-182025.09541.376—
ZTWForwarding Ratio (FR)=0.8, Batch size=1, Backbone=ResNet-182025.09531.0765—
DFMUClass=Veg, Backbone=ViT2026.067.81——91.56
DFMUClass=Nat_Scen, Backbone=ViT2026.064.84——85.12
DFMUClass=Elec_Dev, Backbone=ViT2026.060.39——84.25
DFMUClass=People, Backbone=ViT2026.060.39——87.86
DFMUClass=Vehicle2, Backbone=ViT2026.060.2——81.29
SSDClass=Elec_Dev, Backbone=ViT2026.060——6.99
SSDClass=Vehicle2, Backbone=ViT2026.060——5.29
SSDClass=Veg, Backbone=ViT2026.060——18.57
SSDClass=People, Backbone=ViT2026.060——4.88
SSDClass=Nat_Scen, Backbone=ViT2026.060——5.26