List your company
ResearchFormerly WizWand

Image Classification on iNaturalist 2017 (val)

72.2Top-1 AccuracyViT-B w/ AGLU
57.515261.327665.1468.9524Aug 9, 2020Apr 4, 2021Nov 29, 2021Jul 26, 2022Mar 21, 2023Nov 15, 2023Jul 11, 2024
Updated 8mo ago

Evaluation results

MethodLinks
ViT-B w/ AGLUBackbone=ViT-B, AGLU=true, Crop Size=2242024.0772.2——
ViT-BBackbone=ViT-B, Crop Size=2242024.0772.1——
SE-ResNet50 w/ APA* + AGLUBackbone=ResNet-50, SE=true, APA=true, AGLU=true, Crop Size=2242024.0769.7——
SAMixBackbone=ResNeXt-1012021.1168.26——
SAMixPBackbone=ResNeXt-1012021.1168.23——
AutoMixBackbone=ResNeXt-1012021.1168.03——
PuzzleMixBackbone=ResNeXt-1012021.1167.72——
CutMixBackbone=ResNeXt-1012021.1167.59——
SE-ResNet50Backbone=ResNet-50, SE=true, Crop Size=2242024.0767.3——
SaliencyMixBackbone=ResNeXt-1012021.1167.2——
ResizeMixBackbone=ResNeXt-1012021.1166.82——
FMixBackbone=ResNeXt-1012021.1166.64——
Feature Space AugmentationBackbone=ResNet-1522020.0866.58——
MixupBackbone=ResNeXt-1012021.1166.27——
ManifoldMixBackbone=ResNeXt-1012021.1166.08——
BaselineBackbone=ResNet-1522020.0865.12——
CBBackbone=ResNet-1522020.0864.75——
Feature Space AugmentationBackbone=ResNet-1012020.0864.16——
VanillaBackbone=ResNeXt-1012021.1163.7——
SAMixPBackbone=ResNet-502021.1163.38——
SAMixBackbone=ResNet-502021.1163.32——
AutoMixBackbone=ResNet-502021.1163.08——
PuzzleMixBackbone=ResNet-502021.1162.66——
SaliencyMixBackbone=ResNet-502021.1162.51——
CutMixBackbone=ResNet-502021.1162.34——
ResizeMixBackbone=ResNet-502021.1162.29——
Feature Space AugmentationBackbone=ResNet-502020.0861.96——
FMixBackbone=ResNet-502021.1161.9——
BaselineBackbone=ResNet-1012020.0861.81——
ManifoldMixBackbone=ResNet-502021.1161.47——
MixupBackbone=ResNet-502021.1161.22——
CBBackbone=ResNet-101, Source=Literature2020.0860.94——
BaselineBackbone=ResNet-502020.0860.5——
VanillaBackbone=ResNet-502021.1160.23——
CBBackbone=ResNet-50, Source=Literature2020.0858.08——
ResNet-101Loss=Softmax, Input Size=224 x 2242019.01—42.5720.42
ResNet-101Loss=CB Focal, beta=0.999, gamma=0.5, Input Size=224 x 2242019.01—39.0618.96
ResNet-101Loss=CB Focal, beta=0.999, gamma=0.5, Input Size=320 x 3202019.01—34.9615.9
ResNet-152Loss=Softmax, Input Size=224 x 2242019.01—41.4219.47
ResNet-152Loss=CB Focal, beta=0.999, gamma=0.5, Input Size=224 x 2242019.01—38.0618.42
ResNet-152Loss=CB Focal, beta=0.999, gamma=0.5, Input Size=320 x 3202019.01—33.7314.96
ResNet-50Loss=Softmax, Input Size=224 x 2242019.01—45.3822.67
ResNet-50Loss=CB Focal, beta=0.999, gamma=0.5, Input Size=224 x 2242019.01—41.9220.92
ResNet-50Loss=CB Focal, beta=0.999, gamma=0.5, Input Size=320 x 3202019.01—38.1618.28