OOD Detection on CIFAR-10 OpenOOD v1.5 (test)
91.34AUROCKNN
Evaluation results
| Method | Links | ||
|---|---|---|---|
| KNNBackbone=ResNet-502026.05 | 91.34 | 46.45 | |
| MahaVarBackbone=ResNet-502026.05 | 90.91 | 46.09 | |
| Mahalanobis++Backbone=ResNet-502026.05 | 90.36 | 50.21 | |
| GENBackbone=ResNet-502026.05 | 90.01 | 43.18 | |
| NCIBackbone=ResNet-502026.05 | 90 | 50.49 | |
| NCI+filterBackbone=ResNet-502026.05 | 90 | 50.49 | |
| NNGuideBackbone=ResNet-502026.05 | 89.72 | 41.1 | |
| EnergyBackbone=ResNet-502026.05 | 89.64 | 40.38 | |
| ViMBackbone=ResNet-502026.05 | 89.58 | 58.37 | |
| MaxLogitBackbone=ResNet-502026.05 | 89.55 | 40.74 | |
| MSPBackbone=ResNet-502026.05 | 89.51 | 51.72 | |
| ReActBackbone=ResNet-502026.05 | 89.32 | 43.44 | |
| IDBDBackbone=ResNet-502026.05 | 88.19 | 54.03 | |
| MahalanobisBackbone=ResNet-502026.05 | 87.65 | 67.04 | |
| KL-matchingBackbone=ResNet-502026.05 | 86.46 | 52.12 | |
| SCALEBackbone=ResNet-502026.05 | 86.21 | 44.13 | |
| ASH-SBackbone=ResNet-502026.05 | 85.23 | 45.04 |