List your company
ResearchFormerly WizWand

Image Classification on CIFAR-10 resized to 224 × 224 (test)

97.4AccuracyBEiT large
92.51293.78195.0596.319May 21, 2022
Updated 8mo ago

Evaluation results

MethodLinks
BEiT largeepsilon (ε)=8, pretrained_dataset=ImageNet21k, parameters=303M, clipping_method=mixed ghost clipping2022.0597.4
BEiT largeepsilon (ε)=4, pretrained_dataset=ImageNet21k, parameters=303M, clipping_method=mixed ghost clipping2022.0597.2
BEiT largeepsilon (ε)=2, pretrained_dataset=ImageNet21k, parameters=303M, clipping_method=mixed ghost clipping2022.0597.1
BEiT largeepsilon (ε)=1, pretrained_dataset=ImageNet21k, parameters=303M, clipping_method=mixed ghost clipping2022.0596.7
De et al.epsilon (ε)=8, pretrained_dataset=ImageNet1k2022.0596.6
CrossViT baseepsilon (ε)=8, pretrained_dataset=ImageNet1k, parameters=104M, clipping_method=mixed ghost clipping2022.0596.5
CrossViT baseepsilon (ε)=4, pretrained_dataset=ImageNet1k, parameters=104M, clipping_method=mixed ghost clipping2022.0596.2
De et al.epsilon (ε)=4, pretrained_dataset=ImageNet1k2022.0596.1
CrossViT baseepsilon (ε)=2, pretrained_dataset=ImageNet1k, parameters=104M, clipping_method=mixed ghost clipping2022.0596.1
CrossViT baseepsilon (ε)=1, pretrained_dataset=ImageNet1k, parameters=104M, clipping_method=mixed ghost clipping2022.0595.5
De et al.epsilon (ε)=2, pretrained_dataset=ImageNet1k2022.0595.4
Yu et al.epsilon (ε)=2, pretrained_dataset=ImageNet1k2022.0594.8
De et al.epsilon (ε)=1, pretrained_dataset=ImageNet1k2022.0594.8
Yu et al.epsilon (ε)=1, pretrained_dataset=ImageNet1k2022.0594.3
Tramer et al.epsilon (ε)=2, pretrained_dataset=ImageNet1k2022.0592.7