Image Classification on Flower-102 (test)
0.54NLL (N=510)MAP + GS
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MAP + GSBackbone=ConNeXt-Tiny, Model=L2-SP2025.02 | 0.54 | 0.28 | |
| DE ELBOBackbone=ConNeXt-Tiny, Model=L2-SP2025.02 | 0.58 | 0.3 | |
| MAP + LaplaceBackbone=ResNet-50, Model=L2-zero2025.02 | 0.65 | 0.35 | |
| MAP + LaplaceBackbone=ResNet-50, Model=L2-SP2025.02 | 0.65 | 0.33 | |
| MAP + GSBackbone=ResNet-50, Model=PTYL2025.02 | 0.65 | 0.33 | |
| MAP + GSBackbone=ViT-B/16, Model=L2-SP2025.02 | 0.69 | 0.42 | |
| DE ELBOBackbone=ResNet-50, Model=L2-zero2025.02 | 0.71 | 0.36 | |
| MAP + GSBackbone=ResNet-50, Model=PTYL (SSL)2025.02 | 0.71 | 0.38 | |
| DE ELBOBackbone=ResNet-50, Model=PTYL, Order=22025.02 | 0.71 | 0.38 | |
| MAP + GSBackbone=ResNet-50, Model=Linear probing2025.02 | 0.83 | 0.53 | |
| MAP + GSBackbone=ResNet-50, Model=L2-zero2025.02 | 0.83 | 0.53 | |
| MAP + GSBackbone=ResNet-50, Model=L2-SP2025.02 | 0.84 | 0.44 | |
| MAP + LaplaceBackbone=ResNet-50, Model=PTYL2025.02 | 0.84 | 0.49 | |
| DE ELBOBackbone=ViT-B/16, Model=L2-SP2025.02 | 0.9 | 0.51 | |
| DE ELBOBackbone=ResNet-50, Model=PTYL, Order=12025.02 | 1.1 | 0.72 | |
| DE ELBOBackbone=ResNet-50, Model=L2-SP2025.02 | 4.62 | 2.6 |