Image Classification on CIFAR-100 (test) (Accuracy, Sparsity, and Effective Rank Profile)
71AccuracyPCD
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| PCDBackbone=ResNet-34, priority tolerance (τ)=0.022026.06 | 71 | 92.6 | 24.5 | |
| PCDBackbone=Inception, priority tolerance (τ)=0.022026.06 | 67.8 | 76.1 | 27.7 | |
| AuxiNashBackbone=ResNet-342026.06 | 48.7 | 98.8 | 184.8 | |
| CAGradBackbone=ResNet-342026.06 | 36.1 | 98.9 | 186.3 | |
| WSBackbone=ResNet-342026.06 | 34.2 | 90.6 | 137.9 | |
| AuxiNashBackbone=Inception2026.06 | 31.9 | 93.1 | 12.1 | |
| PCGradBackbone=ResNet-342026.06 | 30 | 92.2 | 160 | |
| WSBackbone=Inception2026.06 | 29.8 | 86.6 | 28.1 | |
| FAMOBackbone=ResNet-342026.06 | 20.3 | 98.4 | 190.5 | |
| CAGradBackbone=Inception2026.06 | 14.1 | 94.3 | 59 | |
| PCGradBackbone=Inception2026.06 | 12.4 | 86.8 | 52.4 | |
| MGDABackbone=Inception2026.06 | 4.3 | 94.5 | 70.3 | |
| MGDABackbone=ResNet-342026.06 | 1.1 | 98.9 | 227.6 | |
| FAMOBackbone=Inception2026.06 | 1 | 94.5 | 100.3 |