OOD Detection on CIFAR10
98.3AUCCUPID Aleatoric
Evaluation results
| Method | Links | ||
|---|---|---|---|
| CUPID AleatoricUncertainty Type=Aleatoric2026.03 | 98.3 | 99.8 | |
| IGRUE2026.03 | 97.8 | 99.8 | |
| CUPID EpistemicUncertainty Type=Epistemic2026.03 | 89.8 | 99.1 | |
| MC Dropout2026.03 | 88.7 | 98.6 | |
| PostNet EpistemicUncertainty Type=Epistemic2026.03 | 77.3 | 97.6 | |
| DEC2026.03 | 66 | 96.3 | |
| PostNet AleatoricUncertainty Type=Aleatoric2026.03 | 65.7 | 95.2 | |
| BNN2026.03 | 64.3 | 95.9 | |
| Rate-in2026.03 | 62 | 92.7 | |
| VIMBackbone=ViT-B/16, Evaluation Protocol=Fully Fine-tuned2024.05 | 0.9705 | — | |
| KNNBackbone=ViT-B/16, Evaluation Protocol=Fully Fine-tuned2024.05 | 0.9682 | — | |
| ReactBackbone=ViT-B/16, Evaluation Protocol=Fully Fine-tuned2024.05 | 0.9672 | — | |
| MDSBackbone=ViT-B/16, Evaluation Protocol=Fully Fine-tuned2024.05 | 0.967 | — | |
| ASHBackbone=ViT-B/16, Evaluation Protocol=Fully Fine-tuned2024.05 | 0.9654 | — | |
| RMDSBackbone=ViT-B/16, Evaluation Protocol=Fully Fine-tuned2024.05 | 0.9576 | — | |
| CLIPNBackbone=ViT-B/16, Evaluation Protocol=Zero-shot CLIP2024.05 | 0.9505 | — | |
| MSPBackbone=ViT-B/16, Evaluation Protocol=Fully Fine-tuned2024.05 | 0.9503 | — | |
| CLS-MBackbone=ViT-B/16, Evaluation Protocol=16-shot Prompt Learning2024.05 | 0.9322 | — | |
| CLS-EBackbone=ViT-B/16, Evaluation Protocol=16-shot Prompt Learning2024.05 | 0.9253 | — | |
| LoCoOp (GL-MCM)Backbone=ViT-B/16, Evaluation Protocol=16-shot Prompt Learning2024.05 | 0.9238 | — | |
| MCMBackbone=ViT-B/16, Evaluation Protocol=Zero-shot CLIP2024.05 | 0.9235 | — | |
| AdaNegBackbone=ViT-B/16, Evaluation Protocol=Zero-shot CLIP2024.05 | 0.9134 | — | |
| NegLabelBackbone=ViT-B/16, Evaluation Protocol=Zero-shot CLIP2024.05 | 0.9129 | — | |
| MaxLogitBackbone=ViT-B/16, Evaluation Protocol=16-shot Prompt Learning2024.05 | 0.9117 | — | |
| NegPromptBackbone=ViT-B/16, Evaluation Protocol=16-shot Prompt Learning2024.05 | 0.9074 | — | |
| ODINBackbone=ViT-B/16, Evaluation Protocol=Fully Fine-tuned2024.05 | 0.8997 | — | |
| EnergyBackbone=ViT-B/16, Evaluation Protocol=16-shot Prompt Learning2024.05 | 0.8958 | — | |
| GradNormBackbone=ViT-B/16, Evaluation Protocol=Fully Fine-tuned2024.05 | 0.8823 | — |