Multi-label Classification on MS-COCO 2014 (test)
91.3mAPQ2L-CvT
Evaluation results
| Method | Links | ||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Q2L-CvTBackbone=CvT-w24, Resolution=384x384, Pre-training=ImageNet-22k2021.07 | 91.3 | — | — | — | — | — | — | 88.8 | 83.2 | 85.9 | 89.2 | 84.6 | 86.8 | 92.8 | 71.6 | 80.8 | 93.9 | 72.1 | 81.6 | — | — | — | — | — | — | — | — | — | |
| ML-VPTBackbone=DINOv2/B, Pre-trained Data=LVD-142M, Resolution=448x4482025.04 | 90.6 | 84.2 | 84.8 | 84.5 | 86 | 86.9 | 86.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CvT-w24Backbone=CvT-w24, Resolution=384x384, Pre-training=ImageNet-22k2021.07 | 90.5 | — | — | — | — | — | — | 89.4 | 81.7 | 85.4 | 89.6 | 83.8 | 86.6 | 93.3 | 70.5 | 80.3 | 94.1 | 71.5 | 81.3 | — | — | — | — | — | — | — | — | — | |
| Q2L-SwinLBackbone=Swin-L, Resolution=384x384, Pre-training=ImageNet-22k2021.07 | 90.5 | — | — | — | — | — | — | 89.4 | 81.7 | 85.4 | 89.8 | 83.2 | 86.4 | 93.9 | 70.4 | 80.5 | 94.8 | 71 | 81.2 | — | — | — | — | — | — | — | — | — | |
| VPTBackbone=DINOv2/B, Pre-trained Data=LVD-142M, Resolution=448x4482025.04 | 89.7 | 83.2 | 83.8 | 83.5 | 85.2 | 86.1 | 85.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| E2VPTBackbone=DINOv2/B, Pre-trained Data=LVD-142M, Resolution=448x4482025.04 | 89.6 | 83.1 | 83.7 | 83.4 | 85.2 | 86.1 | 85.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Swin-LBackbone=Swin-L, Resolution=384x384, Pre-training=ImageNet-22k2021.07 | 89.6 | — | — | — | — | — | — | 89.9 | 80.2 | 84.8 | 90.4 | 82.1 | 86.1 | 93.6 | 69.9 | 80 | 94.3 | 71.1 | 81.1 | — | — | — | — | — | — | — | — | — | |
| Q2L-TResLBackbone=TResNetL, Resolution=448x448, Pre-training=ImageNet-22k2021.07 | 89.2 | — | — | — | — | — | — | 86.3 | 81.4 | 83.8 | 86.5 | 83.3 | 84.9 | 91.6 | 69.4 | 79 | 92.9 | 70.5 | 80.2 | — | — | — | — | — | — | — | — | — | |
| GateVPTBackbone=DINOv2/B, Pre-trained Data=LVD-142M, Resolution=448x4482025.04 | 89.1 | 82.5 | 83.1 | 82.8 | 84.5 | 85.5 | 85 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MITr-1Backbone=MLTr-1, Resolution=384x384, Pre-training=ImageNet-22k2021.07 | 88.5 | — | — | — | — | — | — | 86 | 81.4 | 83.3 | 86.5 | 83.4 | 84.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TResNetLBackbone=TResNetL, Resolution=448x448, Pre-training=ImageNet-22k2021.07 | 88.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ML-VPTBackbone=DINOv2/B, Pre-trained Data=LVD-142M, Resolution=224x2242025.04 | 87.5 | 80.8 | 81.4 | 81.1 | 83 | 83.9 | 83.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ML-VPTBackbone=DINOv2/S, Pre-trained Data=LVD-142M, Resolution=448x4482025.04 | 87.4 | 80.8 | 81.4 | 81.1 | 83.2 | 84.1 | 83.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Q2L-TResLBackbone=TResNetL, Resolution=448x4482021.07 | 87.3 | — | — | — | — | — | — | 87.6 | 76.5 | 81.6 | 88.4 | 78.5 | 83.1 | 91.9 | 66.2 | 77 | 93.5 | 67.6 | 78.5 | — | — | — | — | — | — | — | — | — | |
| CSRABackbone=ViT-L16, RandAugment=true2021.08 | 86.9 | — | — | — | — | — | — | 89.1 | 74.2 | 81 | 89.6 | 77.1 | 82.9 | 92.5 | 65.8 | 76.9 | 93.4 | 68.1 | 78.8 | — | — | — | — | — | — | — | — | — | |
| ASLBackbone=TResNetL, Resolution=448x4482021.07 | 86.6 | — | — | — | — | — | — | 87.2 | 76.4 | 81.4 | 88.2 | 79.2 | 81.8 | 91.8 | 63.4 | 75.1 | 92.9 | 66.4 | 77.4 | — | — | — | — | — | — | — | — | — | |
| Q2L-R101Backbone=ResNet101, Resolution=576x5762021.07 | 86.5 | — | — | — | — | — | — | 85.8 | 76.7 | 81 | 87 | 78.9 | 82.8 | 90.4 | 66.3 | 76.5 | 92.4 | 67.9 | 78.3 | — | — | — | — | — | — | — | — | — | |
| ASLData Augmentation=RandAugment [5]2021.08 | 86.5 | — | — | — | — | — | — | 87.2 | 76.4 | 81.4 | 88.2 | 79.2 | 81.8 | 91.8 | 63.4 | 75.1 | 92.9 | 66.4 | 77.4 | — | — | — | — | — | — | — | — | — | |
| CSRABackbone=ViT-L162021.08 | 86.5 | — | — | — | — | — | — | 88.2 | 74.4 | 80.8 | 88.5 | 77.4 | 82.6 | 91.9 | 65.8 | 76.7 | 92.6 | 68.2 | 78.5 | — | — | — | — | — | — | — | — | — | |
| ML-VPTBackbone=ViT-B-21k, Pre-trained Data=ImageNet 21K, Resolution=448x4482025.04 | 86.4 | 79.7 | 80.3 | 80 | 81.6 | 82.5 | 82 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| E2VPTBackbone=DINOv2/B, Pre-trained Data=LVD-142M, Resolution=224x2242025.04 | 86.3 | 79.8 | 80.4 | 80.1 | 82.2 | 83.1 | 82.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| VPTBackbone=DINOv2/B, Pre-trained Data=LVD-142M, Resolution=224x2242025.04 | 86.1 | 79.5 | 80.1 | 79.8 | 81.9 | 82.8 | 82.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GateVPTBackbone=DINOv2/B, Pre-trained Data=LVD-142M, Resolution=224x2242025.04 | 85.6 | 79 | 79.6 | 79.3 | 81.6 | 82.5 | 82 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CSRABackbone=ResNet-cut2021.08 | 85.6 | — | — | — | — | — | — | 86.2 | 74.9 | 80.1 | 86.6 | 78 | 82.1 | 90.1 | 65.7 | 76 | 91.4 | 67.9 | 77.9 | — | — | — | — | — | — | — | — | — | |
| E2VPTBackbone=ViT-B-21k, Pre-trained Data=ImageNet 21K, Resolution=448x4482025.04 | 85.2 | 78.6 | 79.2 | 78.9 | 80.7 | 81.6 | 81.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ADD-GCNBackbone=ResNet101, Resolution=576x5762021.07 | 85.2 | — | — | — | — | — | — | 84.7 | 75.9 | 80.1 | 84.9 | 79.4 | 82 | 88.8 | 66.2 | 75.8 | 90.3 | 68.5 | 77.9 | — | — | — | — | — | — | — | — | — | |
| C-TransBackbone=ResNet101, Resolution=576x5762021.07 | 85.1 | — | — | — | — | — | — | 86.3 | 74.3 | 79.9 | 87.7 | 76.5 | 81.7 | 90.1 | 65.7 | 76 | 92.1 | 71.4 | 77.6 | — | — | — | — | — | — | — | — | — | |
| VPTBackbone=ViT-B-21k, Pre-trained Data=ImageNet 21K, Resolution=448x4482025.04 | 85 | 78.4 | 79 | 78.7 | 80.5 | 81.4 | 81 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Q2L-R101Backbone=ResNet101, Resolution=448x4482021.07 | 84.9 | — | — | — | — | — | — | 84.8 | 74.5 | 79.3 | 86.6 | 76.9 | 81.5 | 78 | 69.1 | 73.3 | 80.7 | 70.8 | 75.4 | — | — | — | — | — | — | — | — | — | |
| GateVPTBackbone=ViT-B-21k, Pre-trained Data=ImageNet 21K, Resolution=448x4482025.04 | 84.5 | 77.8 | 78.4 | 78.1 | 80.1 | 80.9 | 80.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| VPTBackbone=DINOv2/S, Pre-trained Data=LVD-142M, Resolution=448x4482025.04 | 84.5 | 78.2 | 78.8 | 78.5 | 81.1 | 82 | 81.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| E2VPTBackbone=DINOv2/S, Pre-trained Data=LVD-142M, Resolution=448x4482025.04 | 84.3 | 78.1 | 78.8 | 78.5 | 81 | 81.9 | 81.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MS-CMABackbone=ResNet101, Resolution=448x4482021.07 | 83.8 | — | — | — | — | — | — | 82.9 | 74.4 | 78.4 | 84.4 | 77.9 | 81 | 86.7 | 64.9 | 74.3 | 90.9 | 67.2 | 77.2 | — | — | — | — | — | — | — | — | — | |
| MCARBackbone=ResNet101, Resolution=448x4482021.07 | 83.8 | — | — | — | — | — | — | 85 | 72.1 | 78 | 88 | 73.9 | 80.3 | 88.1 | 65.5 | 75.1 | 91 | 66.3 | 76.7 | — | — | — | — | — | — | — | — | — | |
| SSGRLBackbone=ResNet101, Resolution=576x5762021.07 | 83.8 | — | — | — | — | — | — | 89.9 | 68.5 | 76.8 | 91.3 | 70.8 | 79.7 | 91.9 | 62.5 | 72.7 | 93.8 | 64.1 | 76.2 | — | — | — | — | — | — | — | — | — | |
| MS-CMABackbone=ResNet-1012021.08 | 83.8 | — | — | — | — | — | — | 82.9 | 74.4 | 78.4 | 84.4 | 77.9 | 81 | 88.2 | 65 | 74.9 | 90.2 | 67.4 | 77.1 | — | — | — | — | — | — | — | — | — | |
| MCARBackbone=ResNet-1012021.08 | 83.8 | — | — | — | — | — | — | 85 | 72.1 | 78 | 88 | 73.9 | 80.3 | 88.1 | 65.5 | 75.1 | 91 | 66.3 | 76.7 | — | — | — | — | — | — | — | — | — | |
| KSSNetBackbone=ResNet101, Resolution=448x4482021.07 | 83.7 | — | — | — | — | — | — | 84.6 | 73.2 | 77.2 | 87.8 | 76.2 | 81.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| KSSNetBackbone=ResNet-1012021.08 | 83.7 | — | — | — | — | — | — | 84.6 | 73.2 | 77.2 | 87.8 | 76.2 | 81.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ML-VPTBackbone=ViT-B, Pre-trained Data=ImageNet 1K, Resolution=448x4482025.04 | 83.6 | 76.9 | 77.5 | 77.2 | 79.1 | 79.9 | 79.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CSRABackbone=ResNet-1012021.08 | 83.5 | — | — | — | — | — | — | 84.1 | 72.5 | 77.9 | 85.6 | 75.7 | 80.3 | 88.5 | 64.2 | 74.4 | 90.4 | 66.4 | 76.5 | — | — | — | — | — | — | — | — | — | |
| ML-VPTBackbone=DINOv2/S, Pre-trained Data=LVD-142M, Resolution=224x2242025.04 | 83.4 | 76.6 | 77.1 | 76.8 | 79.6 | 80.5 | 80 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GateVPTBackbone=DINOv2/S, Pre-trained Data=LVD-142M, Resolution=448x4482025.04 | 83.3 | 77 | 77.7 | 77.3 | 80 | 80.9 | 80.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ML-VPTBackbone=ViT-B-21k, Pre-trained Data=ImageNet 21K, Resolution=224x2242025.04 | 83 | 76.2 | 76.7 | 76.4 | 78.6 | 79.5 | 79 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ML-GCNBackbone=ResNet101, Resolution=448x4482021.07 | 83 | — | — | — | — | — | — | 85.1 | 72 | 78 | 85.8 | 75.4 | 80.3 | 87.2 | 64.6 | 74.2 | 89.1 | 66.7 | 76.3 | — | — | — | — | — | — | — | — | — | |
| ML-GCNBackbone=ResNet-1012021.08 | 83 | — | — | — | — | — | — | 85.1 | 72 | 78 | 85.8 | 75.4 | 80.3 | 89.2 | 64.1 | 74.6 | 90.5 | 66.5 | 76.7 | — | — | — | — | — | — | — | — | — | |
| VPTBackbone=ViT-B, Pre-trained Data=ImageNet 1K, Resolution=448x4482025.04 | 82.6 | 76.1 | 76.7 | 76.4 | 78.2 | 79.1 | 78.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CADMBackbone=ResNet101, Resolution=448x4482021.07 | 82.3 | — | — | — | — | — | — | 82.5 | 72.2 | 77 | 84 | 75.6 | 79.6 | 87.1 | 63.6 | 73.5 | 89.4 | 66 | 76 | — | — | — | — | — | — | — | — | — | |
| ResNet-cutBackbone=ResNet-cut2021.08 | 82.1 | — | — | — | — | — | — | 86.2 | 68.7 | 76.4 | 88.9 | 73.1 | 80.3 | 88.7 | 61.3 | 72.5 | 92.1 | 65.2 | 76.3 | — | — | — | — | — | — | — | — | — | |
| E2VPTBackbone=ViT-B-21k, Pre-trained Data=ImageNet 21K, Resolution=224x2242025.04 | 81.9 | 75.4 | 76 | 75.7 | 77.9 | 78.8 | 78.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| E2VPTBackbone=ViT-B, Pre-trained Data=ImageNet 1K, Resolution=448x4482025.04 | 81.7 | 75.3 | 75.9 | 75.6 | 77.7 | 78.5 | 78.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| VPTBackbone=ViT-B-21k, Pre-trained Data=ImageNet 21K, Resolution=224x2242025.04 | 81 | 74.5 | 75.1 | 74.8 | 77.3 | 78.1 | 77.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GateVPTBackbone=ViT-B-21k, Pre-trained Data=ImageNet 21K, Resolution=224x2242025.04 | 80.8 | 74.2 | 74.8 | 74.5 | 77.3 | 78.1 | 77.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| E2VPTBackbone=DINOv2/S, Pre-trained Data=LVD-142M, Resolution=224x2242025.04 | 80.6 | 74.2 | 74.8 | 74.5 | 77.5 | 78.4 | 77.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GateVPTBackbone=ViT-B, Pre-trained Data=ImageNet 1K, Resolution=448x4482025.04 | 80.4 | 74.2 | 74.7 | 74.4 | 76.6 | 77.5 | 77 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ViT-L16Backbone=ViT-L162021.08 | 80.4 | — | — | — | — | — | — | 83.8 | 67 | 74.5 | 86.6 | 72 | 78.6 | 86.8 | 60 | 70.1 | 90.3 | 64.7 | 75.4 | — | — | — | — | — | — | — | — | — | |
| ML-VPTBackbone=MoCo v3, Pre-trained Data=ImageNet 1K, Resolution=448x4482025.04 | 80.3 | 73.6 | 74.3 | 74 | 76.5 | 77.3 | 76.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| VPTBackbone=DINOv2/S, Pre-trained Data=LVD-142M, Resolution=224x2242025.04 | 80.1 | 73.8 | 74.4 | 74.1 | 77.1 | 77.9 | 77.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ML-VPTBackbone=ViT-B, Pre-trained Data=ImageNet 1K, Resolution=224x2242025.04 | 79.6 | 73.1 | 73.6 | 73.3 | 75.7 | 76.5 | 76.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GateVPTBackbone=DINOv2/S, Pre-trained Data=LVD-142M, Resolution=224x2242025.04 | 79.4 | 73.2 | 73.8 | 73.5 | 76.5 | 77.3 | 76.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ResNet-101Backbone=ResNet-1012021.08 | 79.4 | — | — | — | — | — | — | 83.4 | 66.6 | 74 | 86.8 | 71.1 | 78.2 | 86.2 | 59.7 | 70.6 | 90.5 | 63.7 | 74.8 | — | — | — | — | — | — | — | — | — | |
| ML-VPTBackbone=MAE, Pre-trained Data=ImageNet 1K, Resolution=448x4482025.04 | 78.8 | 72.8 | 73.4 | 73.1 | 75.2 | 76 | 75.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ResNet-101Backbone=ResNet101, Resolution=224x2242021.07 | 78.3 | — | — | — | — | — | — | 80.2 | 66.7 | 72.8 | 83.9 | 70.8 | 76.8 | 84.1 | 59.4 | 69.7 | 89.1 | 62.8 | 73.6 | — | — | — | — | — | — | — | — | — | |
| VPTBackbone=ViT-B, Pre-trained Data=ImageNet 1K, Resolution=224x2242025.04 | 78 | 71.7 | 72.3 | 72 | 74.6 | 75.4 | 75 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| E2VPTBackbone=ViT-B, Pre-trained Data=ImageNet 1K, Resolution=224x2242025.04 | 77.3 | 71.6 | 72.2 | 71.9 | 74.5 | 75.4 | 75 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SRNBackbone=ResNet101, Resolution=224x2242021.07 | 77.1 | — | — | — | — | — | — | 81.6 | 65.4 | 71.2 | 82.7 | 69.9 | 75.8 | 85.2 | 58.8 | 67.4 | 87.4 | 62.5 | 72.9 | — | — | — | — | — | — | — | — | — | |
| E2VPTBackbone=MoCo v3, Pre-trained Data=ImageNet 1K, Resolution=448x4482025.04 | 76.5 | 70.3 | 70.9 | 70.6 | 74.2 | 75 | 74.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| VPTBackbone=MoCo v3, Pre-trained Data=ImageNet 1K, Resolution=448x4482025.04 | 75.9 | 70.2 | 70.8 | 70.5 | 73.4 | 74.2 | 73.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GateVPTBackbone=ViT-B, Pre-trained Data=ImageNet 1K, Resolution=224x2242025.04 | 75.6 | 69.7 | 70.3 | 70 | 72.9 | 73.7 | 73.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ML-VPTBackbone=MAE, Pre-trained Data=ImageNet 1K, Resolution=224x2242025.04 | 75.2 | 69.3 | 69.9 | 69.6 | 74 | 74.9 | 74.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ML-VPTBackbone=MoCo v3, Pre-trained Data=ImageNet 1K, Resolution=224x2242025.04 | 75.1 | 69.1 | 69.7 | 69.4 | 74.4 | 75.2 | 74.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GateVPTBackbone=MoCo v3, Pre-trained Data=ImageNet 1K, Resolution=448x4482025.04 | 74.9 | 69.4 | 70 | 69.7 | 73.4 | 74.2 | 73.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| E2VPTBackbone=MAE, Pre-trained Data=ImageNet 1K, Resolution=448x4482025.04 | 73 | 67.8 | 68.4 | 68.1 | 72.8 | 73.6 | 73.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| VPTBackbone=MoCo v3, Pre-trained Data=ImageNet 1K, Resolution=224x2242025.04 | 73 | 67.6 | 68.1 | 67.9 | 71.8 | 72.6 | 72.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| E2VPTBackbone=MoCo v3, Pre-trained Data=ImageNet 1K, Resolution=224x2242025.04 | 72.7 | 67.3 | 67.9 | 67.6 | 71.7 | 72.5 | 72.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| VPTBackbone=MAE, Pre-trained Data=ImageNet 1K, Resolution=448x4482025.04 | 72.2 | 66.9 | 67.6 | 67.3 | 72.3 | 73.1 | 72.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| VPTBackbone=MAE, Pre-trained Data=ImageNet 1K, Resolution=224x2242025.04 | 71 | 66 | 66.6 | 66.3 | 71.1 | 72 | 71.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GateVPTBackbone=MoCo v3, Pre-trained Data=ImageNet 1K, Resolution=224x2242025.04 | 70.9 | 65.7 | 66.3 | 66 | 70 | 70.9 | 70.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| E2VPTBackbone=MAE, Pre-trained Data=ImageNet 1K, Resolution=224x2242025.04 | 69.8 | 65 | 65.6 | 65.3 | 70.4 | 71.2 | 70.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GateVPTBackbone=MAE, Pre-trained Data=ImageNet 1K, Resolution=448x4482025.04 | 69.1 | 64.5 | 65.1 | 64.8 | 70.1 | 70.9 | 70.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GateVPTBackbone=MAE, Pre-trained Data=ImageNet 1K, Resolution=224x2242025.04 | 66.5 | 62.4 | 62.9 | 62.6 | 68.2 | 69 | 68.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Algorithm 1Backbone=ResNet-502026.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 69.73 | 63.87 | 61.16 | |
| ASLSetting=MLR-PL2022.05 | — | — | — | 47.9 | — | — | 46.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Asym. Loss (ASL)Backbone=ResNet-502026.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 69.85 | 64.55 | 61.38 | |
| Binary Relevance (BCE)Backbone=ResNet-502026.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 69.64 | 63.49 | 61.03 | |
| CLSetting=MLR-PL2022.05 | — | — | — | 48.3 | — | — | 61.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GCN-MLSetting=MLR-PL2022.05 | — | — | — | 68.4 | — | — | 73.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| HSTSetting=MLR-PL2022.05 | — | — | — | 72.6 | — | — | 76.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| KGGRSetting=MLR-PL2022.05 | — | — | — | 69.7 | — | — | 73.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Macro-ThresBackbone=ResNet-502026.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 69.42 | 64.12 | 60.94 | |
| MMO AlgorithmBackbone=ResNet-50, tau=02026.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 70.06 | 64.7 | 61.61 | |
| Partial BCESetting=MLR-PL2022.05 | — | — | — | 68.8 | — | — | 74 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SigmoidF1Backbone=ResNet-502026.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 69.52 | 64.01 | 61.12 | |
| SSGRLSetting=MLR-PL2022.05 | — | — | — | 68.1 | — | — | 73.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SSTSetting=MLR-PL2022.05 | — | — | — | 71.2 | — | — | 75.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — |