List your company
ResearchFormerly WizWand

1-shot classification on tiered-ImageNet (test)

71.71AccuracyRobust 20 Full
61.746864.333466.9269.5066Mar 27, 2019Oct 14, 2019May 3, 2020Nov 21, 2020Jun 10, 2021Dec 29, 2021Jul 19, 2022
Updated 8mo ago

Evaluation results

MethodLinks
Robust 20 FullInput size=224, Network=ResNet2019.0371.71
RFS+distBackbone=ResNet-12, Number of Parameters=8.0M, Input Size=84, MACs=3.5G, Supervision Type=sup.2022.0771.52
CC+RotBackbone=WRN-28-10, Number of Parameters=36.5M, Input Size=84, MACs=41G, Supervision Type=sup.+ssl2022.0770.53
Robust 20-distInput size=224, Network=ResNet2019.0370.44
Robust+distBackbone=ResNet-18, Number of Parameters=11.2M, Input Size=224, MACs=1.8G, Supervision Type=sup.2022.0770.44
FEATBackbone=WRN-28-10, Number of Parameters=36.5M, Input Size=84, MACs=41G, Supervision Type=sup.2022.0770.41
UniSiam+distBackbone=ResNet-50, Number of Parameters=23.5M, Input Size=224, MACs=4.1G, Supervision Type=unsup.2022.0769.6
Centroid AlignmentBackbone=ResNet-18, Number of Parameters=11.2M, Input Size=224, MACs=1.8G, Supervision Type=sup.2022.0769.29
UniSiamBackbone=ResNet-50, Number of Parameters=23.5M, Input Size=224, MACs=4.1G, Supervision Type=unsup.2022.0769.11
BMLBackbone=ResNet-12, Number of Parameters=8.0M, Input Size=84, MACs=3.5G, Supervision Type=sup.2022.0768.99
UniSiam+distBackbone=ResNet-34, Number of Parameters=21.3M, Input Size=224, MACs=3.6G, Supervision Type=unsup.2022.0768.65
Mean CentroidInput size=224, Network=ResNet2019.0368.33
UniSiamBackbone=ResNet-34, Number of Parameters=21.3M, Input Size=224, MACs=3.6G, Supervision Type=unsup.2022.0767.57
UniSiam+distBackbone=ResNet-18, Number of Parameters=11.2M, Input Size=224, MACs=1.8G, Supervision Type=unsup.2022.0767.01
LEOInput size=80, Network=WideResNet2019.0366.33
LEOBackbone=WRN-28-10, Number of Parameters=36.5M, Input Size=84, MACs=41G, Supervision Type=sup.2022.0766.33
MetaOptNetBackbone=ResNet-12, Number of Parameters=8.0M, Input Size=84, MACs=3.5G, Supervision Type=sup.2022.0765.99
UniSiamBackbone=ResNet-18, Number of Parameters=11.2M, Input Size=224, MACs=1.8G, Supervision Type=unsup.2022.0765.18
SimSiamBackbone=ResNet-18, Number of Parameters=11.2M, Input Size=224, MACs=1.8G, Supervision Type=unsup., implementation=authors2022.0764.05
SimCLRBackbone=ResNet-18, Number of Parameters=11.2M, Input Size=224, MACs=1.8G, Supervision Type=unsup., implementation=authors2022.0763.38
TADAMInput size=84, Network=ResNet2019.0362.13