Image Classification on ImageNet-100 1.0 (test)
92.5CLIP ScoreFull Dataset
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Full DatasetTrain Set=Full2026.02 | 92.5 | 95.2 | 94.1 | 89.4 | 92.8 | |
| Ours (SFM+CI)Train Set=1 Img/Cls2026.02 | 92.2 | 95.1 | 93.1 | 87.5 | 92 | |
| Ours (SFM)Train Set=1 Img/Cls2026.02 | 86.7 | 90.6 | 88.9 | 84.7 | 87.7 | |
| LGMTrain Set=1 Img/Cls2026.02 | 84.9 | 91.5 | 89 | 83.4 | 87.2 | |
| CentroidsTrain Set=1 Img/Cls2026.02 | 77.1 | 86.9 | 80.9 | 77.7 | 80.6 | |
| Neighbors (SFM)Train Set=1 Img/Cls2026.02 | 72.2 | 88 | 80.4 | 77.8 | 79.6 | |
| Neighbors (LGM)Train Set=1 Img/Cls2026.02 | 67.8 | 86 | 78.8 | 77.1 | 77.4 | |
| RandomTrain Set=1 Img/Cls2026.02 | 56.6 | 74.8 | 64.5 | 61.4 | 64.3 |