Image Classification on SVHN (Corruption Robustness Evaluation)
97.36Accuracy (Original)Zubov-Net
Evaluation results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Zubov-NetActivation=Tanh2025.09 | 97.36 | 86.9 | 93.24 | 83.95 | 82.76 | 92.5 | 91.36 | 72.68 | 86.55 | 86.24 | |
| SODEFActivation=sin2025.09 | 97.17 | 86.42 | 92.84 | 83.51 | 82.13 | 92.46 | 91.07 | 73.33 | 86.27 | 86 | |
| Neural ODEActivation=Tanh2025.09 | 97.11 | 84.55 | 92.61 | 80.41 | 75.37 | 91.93 | 90.73 | 69.3 | 81.51 | 83.3 | |
| Proj-NODEActivation=Tanh2025.09 | 97.09 | 85.49 | 92.71 | 81.83 | 80.95 | 91.4 | 91 | 71.52 | 85.87 | 85.1 | |
| ResNet182025.09 | 97.03 | 84.37 | 92.47 | 80.66 | 75.23 | 91.8 | 90.37 | 69.18 | 81.08 | 83.14 | |
| FxTS-Net2025.09 | 96.92 | 86.39 | 92.87 | 82.28 | 81.3 | 92.21 | 90.96 | 71.62 | 84.28 | 85.24 | |
| LyaNet2025.09 | 96.76 | 86.18 | 92.74 | 82.63 | 80.37 | 92.07 | 90.6 | 71.72 | 83.88 | 85.02 |