Multi-label Classification on OpenImages V3 (test)
79.6mAP (G1)SCPNet
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| SCPNet2023.03 | 79.6 | 81.8 | 85.3 | 87.9 | 92.1 | 85.3 | |
| LL-CtBackbone=ResNet-101, Batch size=288, Input Resolution=224x2242022.06 | 77.7 | 79.3 | 82.1 | 84.7 | 89.4 | 82.6 | |
| LargeLoss2023.03 | 77.7 | 79.3 | 82.1 | 84.7 | 89.4 | 82.6 | |
| LL-CpBackbone=ResNet-101, Batch size=288, Input Resolution=224x2242022.06 | 77.6 | 79.1 | 81.9 | 84.6 | 89.4 | 82.5 | |
| LL-RBackbone=ResNet-101, Batch size=288, Input Resolution=224x2242022.06 | 77.4 | 79.1 | 82 | 84.5 | 89.5 | 82.5 | |
| Naive ANBackbone=ResNet-101, Batch size=288, Input Resolution=224x2242022.06 | 77.1 | 78.7 | 81.5 | 84.1 | 88.8 | 82 | |
| Naive AN2023.03 | 77.1 | 78.7 | 81.5 | 84.1 | 88.1 | 82 | |
| P-ASL2023.03 | 73.2 | 78.6 | 85.1 | 87.7 | 90.6 | 83 | |
| WANBackbone=ResNet-101, Batch size=288, Input Resolution=224x2242022.06 | 71.8 | 72.8 | 76.3 | 79.7 | 84.7 | 77 | |
| WAN2023.03 | 71.8 | 72.8 | 76.3 | 79.7 | 84.7 | 77 | |
| IMCLBackbone=ResNet-101, Batch size=288, Input Resolution=224x2242022.06 | 71 | 72.6 | 77.6 | 81.8 | 87.3 | 78.1 | |
| IMCL2023.03 | 71 | 72.6 | 77.6 | 81.8 | 87.3 | 78.1 | |
| CurriculumBackbone=ResNet-101, Batch size=288, Input Resolution=224x2242022.06 | 70.4 | 71.3 | 76.2 | 80.5 | 86.8 | 77.1 | |
| CL2023.03 | 70.4 | 71.3 | 76.2 | 80.5 | 86.8 | 77.1 | |
| Naive IUBackbone=ResNet-101, Batch size=288, Input Resolution=224x2242022.06 | 69.5 | 70.3 | 74.8 | 79.2 | 85.5 | 75.9 | |
| LSANBackbone=ResNet-101, Batch size=288, Input Resolution=224x2242022.06 | 68.4 | 69.3 | 73.7 | 77.9 | 85.6 | 75 | |
| LSAN2023.03 | 68.4 | 69.3 | 73.7 | 77.9 | 85.6 | 75 |