Classification on miniImagenet (test)
98.63AccuracyDeiT-B/16 (Last)
Evaluation results
| Method | Links | |
|---|---|---|
| DeiT-B/16 (Last)Model=DeiT-B/16, Selection Strategy=Last Layer2026.05 | 98.63 | |
| DeiT-B/16 (LOES)Model=DeiT-B/16, Selection Strategy=LOES2026.05 | 98.44 | |
| DINOv2-S (LOES)Model=DINOv2-S, Selection Strategy=LOES2026.05 | 94.27 | |
| DINOv2-S (Last)Model=DINOv2-S, Selection Strategy=Last Layer2026.05 | 93.72 | |
| DINOv3-S/16 (LOES)Model=DINOv3-S/16, Selection Strategy=LOES2026.05 | 93.6 | |
| DINOv3-S/16 (Last)Model=DINOv3-S/16, Selection Strategy=Last Layer2026.05 | 93.37 | |
| ViT-IN21k-B/16 (Last)Model=ViT-IN21k-B/16, Selection Strategy=Last Layer2026.05 | 93.17 | |
| ViT-IN21k-B/16 (LOES)Model=ViT-IN21k-B/16, Selection Strategy=LOES2026.05 | 93.15 | |
| CLIP-B32 (LOES)Model=CLIP-B32, Selection Strategy=LOES2026.05 | 90.93 | |
| CLIP-B32 (Last)Model=CLIP-B32, Selection Strategy=Last Layer2026.05 | 89.21 | |
| EPNetSSLBackbone=WRN-28-10, Number of shots=10-shot2020.03 | 89.02 | |
| EPNetSSLBackbone=RESNET-12, Number of shots=10-shot2020.03 | 87.34 | |
| EPNetBackbone=WRN-28-10, Number of shots=10-shot2020.03 | 87.03 | |
| EPNetBackbone=RESNET-12, Number of shots=10-shot2020.03 | 85.39 | |
| MAE-B/16 (LOES)Model=MAE-B/16, Selection Strategy=LOES2026.05 | 84.22 | |
| MAE-B/16 (Last)Model=MAE-B/16, Selection Strategy=Last Layer2026.05 | 81.82 | |
| TADAMBackbone=RESNET-12, Number of shots=10-shot2020.03 | 80.8 | |
| DiscriminativeBackbone=RESNET-12, Number of shots=10-shot2020.03 | 78.5 |