Image Labeling on ImageNet 1k (test)
8.17FDRSGR
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| SGRModel=CLIP-VIT-B/32, Alpha=0.1, Epsilon=0.22025.10 | 8.17 | 39.66 | 25.79 | — | |
| SGRModel=ResNet-34, Alpha=0.1, Epsilon=0.22025.10 | 8.91 | 77.31 | 62.21 | — | |
| SGRModel=DenseNet-161, Alpha=0.1, Epsilon=0.22025.10 | 8.92 | 82.68 | 70 | — | |
| FDR searchModel=CLIP-VIT-B/32, Alpha=0.12025.10 | 9.77 | 45.41 | 30.04 | — | |
| FDR searchModel=DenseNet-161, Alpha=0.12025.10 | 9.83 | 84.71 | 72.45 | — | |
| Conformal LabelingModel=CLIP-VIT-B/32, Alpha=0.12025.10 | 9.84 | 45.63 | 30.21 | — | |
| FDR searchModel=ResNet-34, Alpha=0.12025.10 | 9.85 | 79.63 | 64.75 | — | |
| Conformal LabelingModel=DenseNet-161, Alpha=0.12025.10 | 9.92 | 84.89 | 72.68 | — | |
| Conformal LabelingModel=ResNet-34, Alpha=0.12025.10 | 9.97 | 79.89 | 65.04 | — | |
| AI onlyModel=ResNet-342025.10 | — | — | — | 26.71 | |
| AI onlyModel=DenseNet-1612025.10 | — | — | — | 22.89 | |
| AI onlyModel=CLIP-VIT-B/322025.10 | — | — | — | 40.35 |