Image Classification on CIFAR-100 (Top-1 Accuracy, Time)
78.34Top-1 AccuracyLLQR
Evaluation results
| Method | Links | ||
|---|---|---|---|
| LLQRBackbone=ResNet-18, Batch Size=256, Epochs=200, Optimizer structure=E-KFAC block2026.05 | 78.34 | 1.25 | |
| SGDMBackbone=ResNet-18, Batch Size=256, Epochs=2002026.05 | 77.52 | 1 | |
| AdaFisher†Backbone=ResNet-18, Batch Size=256, Epochs=2002026.05 | 77.28 | 1.33 | |
| Shampoo†Backbone=ResNet-18, Batch Size=256, Epochs=2002026.05 | 76.78 | 10.82 | |
| SGDM†Backbone=ResNet-18, Batch Size=256, Epochs=2002026.05 | 76.56 | 1 | |
| K-FAC†Backbone=ResNet-18, Batch Size=256, Epochs=2002026.05 | 76.03 | 2.32 | |
| Adam†Backbone=ResNet-18, Batch Size=256, Epochs=2002026.05 | 75.74 | 1.16 |