Out-of-Distribution Detection on CIFAR-100 (ID) vs Textures (Far-OOD) (OpenOOD v1.5, test)
35.86FPR@95ViM
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ViMBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 35.86 | 91.91 | |
| MDSBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 38.52 | 90.14 | |
| RMDSBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 43.03 | 86.84 | |
| EPDBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 45.21 | 87.93 | |
| KNNBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 47.88 | 87.96 | |
| ReActBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 57.11 | 87.31 | |
| GradNormBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 58.1 | 86.65 | |
| GENBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 58.78 | 86.89 | |
| EBOBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 59.26 | 86.88 | |
| MLSBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 59.27 | 86.51 | |
| TempScaleBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 61.7 | 84.23 | |
| MSPBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 64.86 | 82.25 | |
| SHEBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 65.28 | 86.85 | |
| DICEBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 73.32 | 82.57 | |
| OpenMaxBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 75.71 | 79.82 | |
| ASHBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 82.07 | 58.53 |