List your company
ResearchFormerly WizWand

Image Classification on ImageNet-1k (val) (Top-1 Accuracy, FLOPs, and Throughput)

2.6FLOPs (G)SPVIT
-3.99240.50485129.496May 29, 2023
Updated 8mo ago

Evaluation results

MethodLinks
SPVITBackbone Model=ViT-S (DeiT)2023.052.679.34—
DynamicViTBackbone Model=ViT-S (DeiT)2023.052.979.36,527
ToMeBackbone Model=ViT-S (DeiT)2023.052.979.496,712
ATSBackbone Model=ViT-S (DeiT)2023.052.979.7—
DiffRateBackbone Model=ViT-S (DeiT), Fine-tuning Protocol=None2023.052.979.586,744
DiffRateBackbone Model=ViT-S (DeiT), Fine-tuning Protocol=30 epochs2023.052.979.836,744
Evo-ViTBackbone Model=ViT-S (DeiT)2023.05379.46,679
EVITBackbone Model=ViT-S (DeiT)2023.05379.56,807
BaselineBackbone Model=ViT-S (DeiT)2023.054.679.825,039
EVITBackbone Model=ViT-B (DeiT)2023.0511.581.32,886
ToMeBackbone Model=ViT-B (DeiT)2023.0511.581.412,834
DiffRateBackbone Model=ViT-B (DeiT), Fine-tuning Protocol=None2023.0511.581.52,865
DiffRateBackbone Model=ViT-B (DeiT), Fine-tuning Protocol=30 epochs2023.0511.581.712,865
ToMeBackbone Model=ViT-B (MAE)2023.0511.582.942,834
DiffRateBackbone Model=ViT-B (MAE), Fine-tuning Protocol=None2023.0511.582.912,865
DiffRateBackbone Model=ViT-B (MAE), Fine-tuning Protocol=30 epochs2023.0511.583.252,865
BaselineBackbone Model=ViT-B (DeiT)2023.0517.681.832,130
BaselineBackbone Model=ViT-B (MAE)2023.0517.683.722,130
ToMeBackbone Model=ViT-L (MAE)2023.0542.385.591,033
DiffRateBackbone Model=ViT-L (MAE), Fine-tuning Protocol=None2023.0542.385.561,045
DiffRateBackbone Model=ViT-L (MAE), Fine-tuning Protocol=30 epochs2023.0542.385.711,045
BaselineBackbone Model=ViT-L (MAE)2023.0561.685.95758
ToMeBackbone Model=ViT-H (MAE)2023.05103.486.51442
DiffRateBackbone Model=ViT-H (MAE), Fine-tuning Protocol=None2023.05103.486.72450
BaselineBackbone Model=ViT-H (MAE)2023.05167.486.88299