Out-of-Distribution Detection on ImageNet-R (AUROC, FPR95)
88.08AUROCMM++
Evaluation results
| Method | Links | ||
|---|---|---|---|
| MM++Backbone=ConvNeXt-T, In-Distribution Dataset=ImageNet-LT2026.06 | 88.08 | 49.12 | |
| Maha++Backbone=ConvNeXt-T, In-Distribution Dataset=ImageNet-LT2026.06 | 87.32 | 52.25 | |
| rMaha++Backbone=ConvNeXt-T, In-Distribution Dataset=ImageNet-LT2026.06 | 87.29 | 52.37 | |
| MahaBackbone=ConvNeXt-T, In-Distribution Dataset=ImageNet-LT2026.06 | 87.03 | 51.92 | |
| rMahaBackbone=ConvNeXt-T, In-Distribution Dataset=ImageNet-LT2026.06 | 85.55 | 52.62 | |
| MSPBackbone=ConvNeXt-T, In-Distribution Dataset=ImageNet-LT2026.06 | 77.58 | 67.02 | |
| KNNBackbone=ConvNeXt-T, In-Distribution Dataset=ImageNet-LT2026.06 | 65.56 | 91.85 | |
| ReActBackbone=ConvNeXt-T, In-Distribution Dataset=ImageNet-LT2026.06 | 64.86 | 89.88 | |
| EnergyBackbone=ConvNeXt-T, In-Distribution Dataset=ImageNet-LT2026.06 | 57.71 | 91.89 | |
| ODINBackbone=ConvNeXt-T, In-Distribution Dataset=ImageNet-LT2026.06 | 55.07 | 92.33 |