Fine-grained classification on CompCars
74.69W -> S AccuracyCFSG
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| CFSGBackbone=ResNet, Params=28M2026.01 | 74.69 | 20.62 | 47.66 | |
| S-CFSGBackbone=ResNet, Params=28M2026.01 | 73.12 | 17.19 | 45.16 | |
| S-FSDGBackbone=ResNet, Params=26M2026.01 | 53.44 | 10.83 | 32.14 | |
| FSDGBackbone=ResNet, Params=26M2026.01 | 51.78 | 11.3 | 31.54 | |
| SAGMBackbone=ResNet, Params=23M2026.01 | 49.55 | 8.58 | 29.07 | |
| PAN (DA)Backbone=ResNet, Params=103M2026.01 | 47.05 | 15.57 | 31.31 | |
| MIROBackbone=ResNet, Params=47M2026.01 | 46.01 | 7.88 | 26.95 | |
| SagNetBackbone=ResNet, Params=24M2026.01 | 45.33 | 8.89 | 27.11 | |
| MLDGBackbone=ResNet, Params=23M2026.01 | 44.94 | 7.56 | 26.25 | |
| ERMBackbone=ResNet, Params=24M2026.01 | 44.15 | 7.54 | 25.85 | |
| GroupDROBackbone=ResNet, Params=23M2026.01 | 43.6 | 7.75 | 25.68 | |
| MixupBackbone=ResNet, Params=23M2026.01 | 43.07 | 7.56 | 25.32 | |
| CORALBackbone=ResNet, Params=23M2026.01 | 43.05 | 7.97 | 25.51 | |
| MixStyleBackbone=ResNet, Params=23M2026.01 | 38.37 | 6.28 | 22.33 | |
| RIDGBackbone=ResNet, Params=24M2026.01 | 36.57 | 8.11 | 22.34 | |
| DANNBackbone=ResNet, Params=24M2026.01 | 35.1 | 6.8 | 20.95 | |
| ARMBackbone=ResNet, Params=24M2026.01 | 20.25 | 4.74 | 12.5 |