List your company
ResearchFormerly WizWand

Image Classification on Tiny ImageNet 200 class (test)

87.2Top-1 AccuracySS-CA
14.71233.53152.3571.169Nov 15, 2025Dec 17, 2025Jan 18, 2026Feb 20, 2026Mar 24, 2026Apr 25, 2026May 28, 2026
Updated 4mo ago

Evaluation results

MethodLinks
SS-CABackbone=ViT, Evaluation Protocol=End-to-End Fine-Tuning2025.1187.2
Chen et al.Backbone=ViT, Evaluation Protocol=End-to-End Fine-Tuning2025.1186.83
Xiao et al.Backbone=ViT, Evaluation Protocol=End-to-End Fine-Tuning2025.1186.69
Conventional TrainingBackbone=ViT, Evaluation Protocol=End-to-End Fine-Tuning2025.1186.68
SS-CABackbone=ResNet-101, Evaluation Protocol=End-to-End Fine-Tuning2025.1175.94
Chen et al.Backbone=ResNet-101, Evaluation Protocol=End-to-End Fine-Tuning2025.1175.7
Conventional TrainingBackbone=ResNet-101, Evaluation Protocol=End-to-End Fine-Tuning2025.1175.67
Xiao et al.Backbone=ResNet-101, Evaluation Protocol=End-to-End Fine-Tuning2025.1175.57
SS-CABackbone=CLIP (ViT/32b), Evaluation Protocol=Linear Probing2025.1174.42
Chen et al.Backbone=CLIP (ViT/32b), Evaluation Protocol=Linear Probing2025.1173.8
Xiao et al.Backbone=CLIP (ViT/32b), Evaluation Protocol=Linear Probing2025.1173.4
Conventional TrainingBackbone=CLIP (ViT/32b), Evaluation Protocol=Linear Probing2025.1173.31
BPLoss Function=CE2026.0539.9
BPLoss Function=MSE2026.0536
SBD ExpLoss Function=CE, Score Expansion=true2026.0531.4
SBDLoss Function=CE, Score Expansion=false2026.0528.3
EBDLoss Function=MSE2026.0518.5
DFALoss Function=CE2026.0517.5