List your company
ResearchFormerly WizWand

Image Classification on 11 Fine-Grained Datasets Average

9.2Mean Top-1 ErrorInCA
8.53613.01817.521.982Mar 7, 2023
Updated 8mo ago

Evaluation results

MethodLinks
InCAArchitecture=ViT-L/16, Resolution=@384, Pre-training data=In21K (DeiT), Model size=304.7M2023.039.2-0.61
InCAArchitecture=CLIP-ViT-L/14, Resolution=@336, Pre-training data=400M Im-Text, Model size=304.2M2023.039.2—
Full FTArchitecture=SWIN-L, Pre-training data=In21K, Model size=196.5M2023.039.30
InCAArchitecture=ViT-H/14, Pre-training data=2B Im-Text, Model size=632.8M2023.039.4—
Full FTArchitecture=ConvNext-B, Pre-training data=In21K, Model size=88.5M2023.039.40
InCAArchitecture=ViT-G/14, Pre-training data=2B Im-Text, Model size=1884.9M2023.039.6—
InCAArchitecture=ViT-L/16, Pre-training data=In21K (DeiT), Model size=304.3M2023.039.8-2.3
Full FTArchitecture=ViT-L/16, Pre-training data=In21K (DeiT), Model size=304.3M2023.03100
InCAArchitecture=SWIN-L, Pre-training data=In21K, Model size=196.5M2023.0310-3.6
InCA (last)Architecture=ViT-H/14, Pre-training data=2B Im-Text, Model size=632.8M, backbone_backprop=false2023.0310.4—
InCA (last)Architecture=ViT-G/14, Pre-training data=2B Im-Text, Model size=1884.9M, backbone_backprop=false2023.0310.4—
InCA (last)Architecture=CLIP-ViT-L/14, Resolution=@336, Pre-training data=400M Im-Text, Model size=304.2M, backbone_backprop=false2023.0310.6—
InCAArchitecture=ConvNext-B, Pre-training data=In21K, Model size=88.5M2023.0310.7-7.4
Full FTArchitecture=ResNext-101, Pre-training data=IG-3.5B, Model size=468.5M2023.0311.40
InCA (last)Architecture=ViT-L/16, Pre-training data=In21K (DeiT), Model size=304.3M, backbone_backprop=false2023.0311.5-7
InCA (last)Architecture=ViT-L/16, Resolution=@384, Pre-training data=In21K (DeiT), Model size=304.7M, backbone_backprop=false2023.0311.7-9.11
InCAArchitecture=ResNext-101, Pre-training data=IG-3.5B, Model size=468.5M2023.0312-8.7
InCA (last)Architecture=SWIN-L, Pre-training data=In21K, Model size=196.5M, backbone_backprop=false2023.0312.4-9.5
InCA (last)Architecture=ConvNext-B, Pre-training data=In21K, Model size=88.5M, backbone_backprop=false2023.0312.5-12.6
Full FTArchitecture=ViT-B/16, Pre-training data=In21K, Model size=86.5M2023.03130
InCAArchitecture=ViT-B/16, Pre-training data=ALBEF (CC14M), Model size=85.9M2023.0313.5-4.2
Full FTArchitecture=ViT-B/16, Pre-training data=ALBEF (CC14M), Model size=85.9M2023.0313.80
inter. LPArchitecture=ViT-H/14, Pre-training data=2B Im-Text, Model size=632.8M2023.0314—
InCA (last)Architecture=ViT-B/16, Pre-training data=ALBEF (CC14M), Model size=85.9M, backbone_backprop=false2023.0314.8-9.3
LPArchitecture=ViT-H/14, Pre-training data=2B Im-Text, Model size=632.8M2023.0315.2—
inter. LPArchitecture=ViT-G/14, Pre-training data=2B Im-Text, Model size=1884.9M2023.0315.3—
inter. LPArchitecture=SWIN-L, Pre-training data=In21K, Model size=196.5M2023.0315.8-31.3
InCAArchitecture=ViT-B/16, Pre-training data=In21K, Model size=86.5M2023.0315.9-7.6
LPArchitecture=ViT-G/14, Pre-training data=2B Im-Text, Model size=1884.9M2023.0316.8—
inter. LPArchitecture=ViT-L/16, Pre-training data=In21K (DeiT), Model size=304.3M2023.0317.2-35.7
inter. LPArchitecture=ViT-L/16, Resolution=@384, Pre-training data=In21K (DeiT), Model size=304.7M2023.0317.3-38.1
InCA (last)Architecture=ResNext-101, Pre-training data=IG-3.5B, Model size=468.5M, backbone_backprop=false2023.0317.3-27.1
InCA (last)Architecture=ViT-B/16, Pre-training data=In21K, Model size=86.5M, backbone_backprop=false2023.0317.5-16.4
LPArchitecture=SWIN-L, Pre-training data=In21K, Model size=196.5M2023.0318.3-40.5
inter. LPArchitecture=ConvNext-B, Pre-training data=In21K, Model size=88.5M2023.0319.1-44.2
LPArchitecture=ConvNext-B, Pre-training data=In21K, Model size=88.5M2023.0319.4-44.2
inter. LPArchitecture=CLIP-ViT-L/14, Resolution=@336, Pre-training data=400M Im-Text, Model size=304.2M2023.0319.6—
inter. LPArchitecture=ResNext-101, Pre-training data=IG-3.5B, Model size=468.5M2023.0320.1-38.8
LPArchitecture=ResNext-101, Pre-training data=IG-3.5B, Model size=468.5M2023.0321.3-39.7
LPArchitecture=CLIP-ViT-L/14, Resolution=@336, Pre-training data=400M Im-Text, Model size=304.2M2023.0321.8—
LPArchitecture=ViT-L/16, Pre-training data=In21K (DeiT), Model size=304.3M2023.0321.9-52.5
LPArchitecture=ViT-L/16, Resolution=@384, Pre-training data=In21K (DeiT), Model size=304.7M2023.0322-54.4
inter. LPArchitecture=ViT-B/16, Pre-training data=In21K, Model size=86.5M2023.0323.9-32.4
LPArchitecture=ViT-B/16, Pre-training data=In21K, Model size=86.5M2023.0324.3-32.4
inter. LPArchitecture=ViT-B/16, Pre-training data=ALBEF (CC14M), Model size=85.9M2023.0324.7-42.6
LPArchitecture=ViT-B/16, Pre-training data=ALBEF (CC14M), Model size=85.9M2023.0325.8-42.6