Image Classification on CIFAR-10 (test) (Anat & Arob Metrics)
81.61Anat ScoreADMM
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ADMMBackbone=ResNet-18, Pruning Ratio (p)=25%, Size [MB]=32.1, FPS=312021.10 | 81.61 | 42.67 | |
| ADMM BaselineBackbone=ResNet-18, Size [MB]=42.6, FPS=282021.10 | 80.65 | 47.05 | |
| ADMMBackbone=ResNet-18, Pruning Ratio (p)=50%, Size [MB]=21.7, FPS=602021.10 | 79.42 | 42.23 | |
| GDWSBackbone=ResNet-18, Beta (β)=0.75, Size [MB]=30.4, FPS=1052021.10 | 79.13 | 46.15 | |
| ADMMBackbone=VGG-16, Pruning Ratio (p)=25%, Size [MB]=31.6, FPS=262021.10 | 77.88 | 43.8 | |
| ADMM BaselineBackbone=VGG-16, Size [MB]=56.2, FPS=362021.10 | 77.45 | 45.78 | |
| GDWSBackbone=VGG-16, Beta (β)=0.5, Size [MB]=38.8, FPS=1192021.10 | 76.4 | 46.28 | |
| ADMMBackbone=VGG-16, Pruning Ratio (p)=50%, Size [MB]=14, FPS=1132021.10 | 75.33 | 42.93 | |
| ADMMBackbone=ResNet-18, Pruning Ratio (p)=75%, Size [MB]=11.2, FPS=742021.10 | 74.62 | 43.23 | |
| ADMMBackbone=VGG-16, Pruning Ratio (p)=75%, Size [MB]=3.5, FPS=1742021.10 | 70.39 | 41.07 |