Image Classification on CIFAR-100 down-sampled to 32x32 (test)
81.8Median AccuracyAll layers
Evaluation results
| Method | Links | ||
|---|---|---|---|
| All layersepsilon=8, Fine-tuning algorithm=DP-SGD, Backbone=28-10 Wide-ResNet, Pre-training=ImageNet2022.04 | 81.8 | 0.07 | |
| All layersepsilon=4, Fine-tuning algorithm=DP-SGD, Backbone=28-10 Wide-ResNet, Pre-training=ImageNet2022.04 | 79.2 | 0.24 | |
| Classifier layerepsilon=8, Fine-tuning algorithm=DP-SGD, Backbone=28-10 Wide-ResNet, Pre-training=ImageNet2022.04 | 77.6 | 0.1 | |
| Classifier layerepsilon=4, Fine-tuning algorithm=DP-SGD, Backbone=28-10 Wide-ResNet, Pre-training=ImageNet2022.04 | 76.1 | 0.31 | |
| All layersepsilon=2, Fine-tuning algorithm=DP-SGD, Backbone=28-10 Wide-ResNet, Pre-training=ImageNet2022.04 | 74.7 | 0.15 | |
| Classifier layerepsilon=2, Fine-tuning algorithm=DP-SGD, Backbone=28-10 Wide-ResNet, Pre-training=ImageNet2022.04 | 73.9 | 0.32 | |
| Classifier layerepsilon=1, Fine-tuning algorithm=DP-SGD, Backbone=28-10 Wide-ResNet, Pre-training=ImageNet2022.04 | 70.3 | 0.09 | |
| All layersepsilon=1, Fine-tuning algorithm=DP-SGD, Backbone=28-10 Wide-ResNet, Pre-training=ImageNet2022.04 | 67.4 | 0.17 |