Model Explainability on ImageNet
70.76Pixel AccuracyQUEST
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| QUESTBackbone=ViT-B/16, Training Protocol=DeiT-3, Epochs=400 + 20, Explainability Method=AG-CAM2026.03 | 70.76 | 53.54 | 65.97 | |
| Standard AttentionBackbone=ViT-B/16, Training Protocol=DeiT-3, Epochs=400 + 20, Explainability Method=AG-CAM2026.03 | 65.22 | 35.53 | 48.7 |