Image Classification on FashionMNIST (Standard Accuracy, Robust Accuracy (AutoAttack))
91.2Standard AccuracyWideResNet-50
Evaluation results
| Method | Links | ||
|---|---|---|---|
| WideResNet-50Setup=Standard Baseline (fst), Compression=None2024.03 | 91.2 | 4.96 | |
| WideResNet-50Setup=Standard Fine-tuning (Tst), Compression=INT8 Quantization2024.03 | 89.72 | 6.6 | |
| WideResNet-50Setup=Standard Fine-tuning (Tst), Compression=l1-Pruning (80% sparsity)2024.03 | 89.05 | 5.66 | |
| WideResNet-50Setup=Robust Baseline (frb), Compression=None2024.03 | 85.81 | 9.44 | |
| WideResNet-50Setup=Adversarial Fine-tuning (Tad), Compression=INT8 Quantization2024.03 | 85.69 | 11.25 | |
| ViTSetup=Standard Baseline (fst), Compression=None2024.03 | 85.22 | 13.84 | |
| ViTSetup=Standard Fine-tuning (Tst), Compression=INT8 Quantization2024.03 | 84.62 | 16.53 | |
| WideResNet-50Setup=Adversarial Fine-tuning (Tad), Compression=l1-Pruning (80% sparsity)2024.03 | 83.07 | 9.56 | |
| ViTSetup=Standard Fine-tuning (Tst), Compression=l1-Pruning (80% sparsity)2024.03 | 82.17 | 14.45 | |
| ViTSetup=Robust Baseline (frb), Compression=None2024.03 | 80.91 | 24.22 | |
| ViTSetup=Adversarial Fine-tuning (Tad), Compression=INT8 Quantization2024.03 | 79.17 | 22.1 | |
| ViTSetup=Adversarial Fine-tuning (Tad), Compression=l1-Pruning (80% sparsity)2024.03 | 79.11 | 24.44 |