Multi-label Classification on COCO 2014
81.5mAPTAI-DPT
Evaluation Results
| Method | Links | |
|---|---|---|
| TAI-DPTExtra Training Data=10% Data2023.12 | 81.5 | |
| DualCoOpExtra Training Data=10% Data2023.12 | 78.7 | |
| SARBExtra Training Data=10% Data2023.12 | 75.5 | |
| TagCLIPExtra Training Data=None2023.12 | 68.8 | |
| BCPBackbone=ViT-B/16, BP-free=true2026.06 | 65.47 | |
| TAI-DPTExtra Training Data=COCO captions2023.12 | 65.1 | |
| DPTExtra Training Data=None, Softmax=true2023.12 | 64.3 | |
| SCABackbone=ViT-B/16, BP-free=true2026.06 | 63.31 | |
| CLIPExtra Training Data=None, Softmax=true2023.12 | 63.3 | |
| BCPBackbone=ViT-B/32, BP-FREE=true2026.06 | 62.97 | |
| BCPBackbone=RN-101, BP-FREE=true2026.06 | 62.84 | |
| ADAPTBackbone=ViT-B/16, BP-free=true2026.06 | 61.91 | |
| BCPBackbone=RN-50, BP-free=true2026.06 | 61.69 | |
| SSGBackbone=ViT-B/16, BP-free=true2026.06 | 61.49 | |
| CLIPSurgeryExtra Training Data=None2023.12 | 61.2 | |
| STSBackbone=ViT-B/16, BP-free=false2026.06 | 59.8 | |
| DPTExtra Training Data=None, Softmax=false2023.12 | 59.6 | |
| BEMBackbone=ViT-B/16, BP-free=false2026.06 | 57.52 | |
| DOTABackbone=ViT-B/16, BP-free=true2026.06 | 56.29 | |
| DOTABackbone=RN-50, BP-free=true2026.06 | 55.69 | |
| TDABackbone=ViT-B/16, BP-free=true2026.06 | 55.21 | |
| DARTBackbone=ViT-B/16, BP-free=false2026.06 | 54.73 | |
| DOTABackbone=ViT-B/32, BP-FREE=true2026.06 | 54.66 | |
| CLIPBackbone=ViT-B/16, BP-free=true2026.06 | 54.42 | |
| RLCFBackbone=ViT-B/16, BP-free=false2026.06 | 54.21 | |
| CLIPExtra Training Data=None, Softmax=false2023.12 | 54.2 | |
| DiffTPTBackbone=ViT-B/16, BP-free=false2026.06 | 53.91 | |
| TPTBackbone=ViT-B/16, BP-free=false2026.06 | 53.32 | |
| BEMBackbone=RN-101, BP-FREE=false2026.06 | 52.92 | |
| BEMBackbone=ViT-B/32, BP-FREE=false2026.06 | 52.83 | |
| DMNBackbone=ViT-B/16, BP-free=true2026.06 | 52.52 | |
| BEMBackbone=RN-50, BP-free=false2026.06 | 51.58 | |
| TDABackbone=ViT-B/32, BP-FREE=true2026.06 | 51.23 | |
| CLIPBackbone=ViT-B/32, BP-FREE=true2026.06 | 50.31 | |
| RLCFBackbone=ViT-B/32, BP-FREE=false2026.06 | 50.28 | |
| TDABackbone=RN-101, BP-FREE=true2026.06 | 50.19 | |
| TPTBackbone=RN-101, BP-FREE=false2026.06 | 49.71 | |
| DIFFTPTBackbone=RN-101, BP-FREE=false2026.06 | 49.45 | |
| DMNBackbone=ViT-B/32, BP-FREE=true2026.06 | 49.32 | |
| TDABackbone=RN-50, BP-free=true2026.06 | 48.91 | |
| CLIPBackbone=RN-101, BP-FREE=true2026.06 | 48.83 | |
| DIFFTPTBackbone=ViT-B/32, BP-FREE=false2026.06 | 48.73 | |
| DiffTPTBackbone=RN-50, BP-free=false2026.06 | 48.56 | |
| TPTBackbone=RN-50, BP-free=false2026.06 | 48.52 | |
| DOTABackbone=RN-101, BP-FREE=true2026.06 | 48.46 | |
| TPTBackbone=ViT-B/32, BP-FREE=false2026.06 | 48.12 | |
| CLIPBackbone=RN-50, BP-free=true2026.06 | 47.53 | |
| DMNBackbone=RN-101, BP-FREE=true2026.06 | 46.28 | |
| DMNBackbone=RN-50, BP-free=true2026.06 | 44.54 | |
| RLCFBackbone=RN-101, BP-FREE=false2026.06 | 40.53 | |
| RLCFBackbone=RN-50, BP-free=false2026.06 | 36.87 | |
| Plugin NetworksBackbone=ResNet-18, Inference time [s]=25.722019.01 | 27.97 | |
| F. Prop (RF)Backbone=ResNet-18, Inference time [s]=103.272019.01 | 25.7 | |
| F. Prop (LF)Backbone=ResNet-18, Inference time [s]=93.362019.01 | 25.26 | |
| Base modelBackbone=ResNet-18, Inference time [s]=25.642019.01 | 23 |