List your company
ResearchFormerly WizWand

Fine-grained Image Classification on Oxford Flowers (test)

99.7Top-1 AccuracyARC*
67.92876.176584.42592.6735Feb 4, 2024Feb 12, 2024Feb 21, 2024Mar 1, 2024Mar 10, 2024Mar 19, 2024Mar 28, 2024
Updated 8mo ago

Evaluation results

MethodLinks
ARC*Backbone=ViT-B/16, Pre-training=ImageNet-21k, Params. (M)=0.25, AugReg=true2024.0399.7—
SSFBackbone=ViT-B/16, Pre-training=ImageNet-21k, Params. (M)=0.39, AugReg=true2024.0399.6—
RLRR*Backbone=ViT-B/16, Pre-training=ImageNet-21k, Params. (M)=0.47, AugReg=true2024.0399.6—
RLRRBackbone=ViT-B/16, Pre-training=ImageNet-21k, Params. (M)=0.47, AugReg=false2024.0399.5—
ARCBackbone=ViT-B/16, Pre-training=ImageNet-21k, Params. (M)=0.25, AugReg=false2024.0399.3—
LoRABackbone=ViT-B/16, Pre-training=ImageNet-21k, Params. (M)=0.44, AugReg=false2024.0399.2—
VPT-DeepBackbone=ViT-B/16, Pre-training=ImageNet-21k, Params. (M)=0.85, AugReg=false2024.0399—
Full fine-tuningBackbone=ViT-B/16, Pre-training=ImageNet-21k, Params. (M)=85.98, AugReg=false2024.0398.8—
BiasBackbone=ViT-B/16, Pre-training=ImageNet-21k, Params. (M)=0.28, AugReg=false2024.0398.8—
AdapterBackbone=ViT-B/16, Pre-training=ImageNet-21k, Params. (M)=0.41, AugReg=false2024.0398.5—
VPT-ShallowBackbone=ViT-B/16, Pre-training=ImageNet-21k, Params. (M)=0.25, AugReg=false2024.0398.4—
Linear probingBackbone=ViT-B/16, Pre-training=ImageNet-21k, Params. (M)=0.18, AugReg=false2024.0397.9—
SPT-DeepBackbone=ViT-B, Pre-training Method=MoCo-v3, Pre-training Dataset=ImageNet-1K, Tuning Depth=Deep2024.0296.1—
SPT-ShallowBackbone=ViT-B, Pre-training Method=MoCo-v3, Pre-training Dataset=ImageNet-1K, Tuning Depth=Shallow2024.0295.03—
FullBackbone=ViT-B, Pre-training Method=MoCo-v3, Pre-training Dataset=ImageNet-1K, Tuning Depth=Full2024.0294.52—
VPT-DeepBackbone=ViT-B, Pre-training Method=MoCo-v3, Pre-training Dataset=ImageNet-1K, Tuning Depth=Deep2024.0294.41—
GateVPTBackbone=ViT-B, Pre-training Method=MoCo-v3, Pre-training Dataset=ImageNet-1K2024.0293.71—
SPT-DeepBackbone=ViT-B, Pre-training Method=MAE, Pre-training Dataset=ImageNet-1K, Tuning Depth=Deep2024.0293.07—
FullBackbone=ViT-B, Pre-training Method=MAE, Pre-training Dataset=ImageNet-1K, Tuning Depth=Full2024.0291.71—
VPT-ShallowBackbone=ViT-B, Pre-training Method=MoCo-v3, Pre-training Dataset=ImageNet-1K, Tuning Depth=Shallow2024.0290.47—
SPT-ShallowBackbone=ViT-B, Pre-training Method=MAE, Pre-training Dataset=ImageNet-1K, Tuning Depth=Shallow2024.0289.47—
VPT-DeepBackbone=ViT-B, Pre-training Method=MAE, Pre-training Dataset=ImageNet-1K, Tuning Depth=Deep2024.0280.05—
GateVPTBackbone=ViT-B, Pre-training Method=MAE, Pre-training Dataset=ImageNet-1K2024.0278.55—
VPT-ShallowBackbone=ViT-B, Pre-training Method=MAE, Pre-training Dataset=ImageNet-1K, Tuning Depth=Shallow2024.0269.15—
AdaLNBackbone=ViT-L/162023.03—700
AdaLNBackbone=SWIN-L, % Trainable parameters=0.1%, No backbone backpropagation=false2023.03—40
BitFitBackbone=ViT-L/162023.03—720
BitFitBackbone=SWIN-L, % Trainable parameters=0.1%, No backbone backpropagation=false2023.03—40
Full FTBackbone=ViT-L/162023.03—60
Full FTBackbone=SWIN-L, % Trainable parameters=100%, No backbone backpropagation=false2023.03—50
In. LPBackbone=ViT-L/162023.03—60
In. LPBackbone=SWIN-L, % Trainable parameters=0.1%, No backbone backpropagation=true2023.03—50
In. MLP-3Backbone=ViT-L/162023.03—50
In. MLP-3Backbone=SWIN-L, % Trainable parameters=2.8%, No backbone backpropagation=true2023.03—50
InCABackbone=ViT-L/162023.03—30
InCABackbone=SWIN-L, % Trainable parameters=3.7%, No backbone backpropagation=true2023.03—30
InCA (last)Backbone=ViT-L/162023.03—40
InCA (last)Backbone=SWIN-L, % Trainable parameters=3.7%, No backbone backpropagation=true2023.03—40
LoRABackbone=ViT-L/162023.03—400
LoRABackbone=SWIN-L, % Trainable parameters=0.8%, No backbone backpropagation=false2023.03—40
LPBackbone=ViT-L/162023.03—110
LPBackbone=SWIN-L, % Trainable parameters=0.1%, No backbone backpropagation=false2023.03—50
MLP-3Backbone=ViT-L/162023.03—70
MLP-3Backbone=SWIN-L, % Trainable parameters=2.8%, No backbone backpropagation=false2023.03—50
VPTBackbone=ViT-L/162023.03—220