Image Classification on ImageNet (Clean and PGD Robustness Evaluation)
81.11Clean AccuracyMixedNUTS
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| MixedNUTSModel=(e)2025.03 | 81.11 | 67.52 | 67.52 | 33,784,085 | 2,758,488 | 91.83 | — | |
| Swin-LModel=(b)2025.03 | 78.19 | 59.27 | 59.27 | 29,396,885 | 2,646,549 | 91 | 91.02 | |
| ConvNeXt-LModel=(a)2025.03 | 77.48 | 58.15 | 58.15 | 29,110,933 | 2,681,126 | 90.79 | 90.53 | |
| ViT-B-ConvStemModel=(d)2025.03 | 76.12 | 55.03 | 55.03 | 27,555,582 | 16,073,533 | 41.67 | 40.4 | |
| XCiT-S12Model=(c)2025.03 | 72.54 | 42.7 | 42.7 | 21,409,926 | 2,160,645 | 89.91 | 89.44 |