List your company
ResearchFormerly WizWand

Few-shot Classification on CUB200 (5-way 5-shot configuration)

95.19AccuracyOur Method (Improved EM)
48.826860.863472.984.9366Aug 11, 2021Aug 27, 2021Sep 13, 2021Sep 29, 2021Oct 16, 2021Nov 1, 2021Nov 18, 2021
Updated 8mo ago

Evaluation results

MethodLinks
Our Method (Improved EM)Setting=Transductive2021.0895.19
Our Method (EM)Setting=Transductive2021.0895.08
Conference VersionSetting=Transductive2021.0894.9
Our MethodSetting=Inductive2021.0894.05
ICISetting=Transductive2021.0892.38
EPNetSetting=Transductive2021.0891.32
TIM-GDSetting=Transductive2021.0890.8
RAP-LaplacianShotSetting=Transductive2021.0890.77
BD-CSPNSetting=Transductive2021.0890.22
DCSetting=Inductive2021.0889.58
Neg-CosineSetting=Inductive2021.0889.4
CPDESetting=Inductive2021.0889.28
LaplacianShotSetting=Transductive2021.0888.68
CentAlignSetting=Inductive2021.0888.65
RAP-ProtoNetSetting=Inductive2021.0888.29
CFASetting=Inductive2021.0886.8
MultiSemSetting=Inductive2021.0882.9
MAMLSetting=Inductive2021.0872.09
SimSiam + AugSelfBackbone=ResNet-50, Pre-training Dataset=ImageNet1002021.1166.27
Supervised + AugSelfBackbone=ResNet-50, Pre-training Dataset=ImageNet1002021.1165.31
SupervisedBackbone=ResNet-50, Pre-training Dataset=ImageNet1002021.1163.69
SimSiamBackbone=ResNet-50, Pre-training Dataset=ImageNet1002021.1162.48
MoCo v2 + AugSelfBackbone=ResNet-50, Pre-training Dataset=ImageNet1002021.1157.35
MoCo v2Backbone=ResNet-50, Pre-training Dataset=ImageNet1002021.1156.92
SimSiam + AugSelfBackbone=ResNet-18, Pre-training Dataset=STL102021.1156.33
MoCo v2 + AugSelfBackbone=ResNet-18, Pre-training Dataset=STL102021.1151.93
MoCo v2Backbone=ResNet-18, Pre-training Dataset=STL102021.1150.71
SimSiamBackbone=ResNet-18, Pre-training Dataset=STL102021.1150.61