Image Classification on CIFAR-10 (Standard Top-1 Accuracy)
14.07Top-1 AccuracyResNet-18
Evaluation results
| Method | Links | |
|---|---|---|
| ResNet-18Backbone=ResNet-182026.03 | 14.07 | |
| Learnable PEPositional Encoding=Learnable2026.03 | 13.68 | |
| RoPEPositional Encoding=RoPE2026.03 | 12.51 | |
| No PEPositional Encoding=None2026.03 | 10.14 |