Image Classification on ImageNet-1K (test) (Accuracy & Runtime)
68.93AccuracyResNet18
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ResNet18Method variant=Parametric2026.03 | 68.93 | 26.1 | |
| SPM-CMethod variant=SPM (Clustering), Subsampling percentage=100%2026.03 | 67.1 | 31.3 | |
| ResNet18-KNNMethod variant=KNN, Subsampling percentage=100%2026.03 | 66.9 | 63.7 | |
| SPM-CMethod variant=SPM (Clustering), Subsampling percentage=10%2026.03 | 66.6 | 31.3 | |
| SPM-CMethod variant=SPM (Clustering), Subsampling percentage=5%2026.03 | 66.1 | 31.3 | |
| ResNet18-KNNMethod variant=KNN, Subsampling percentage=10%2026.03 | 63.4 | 34.5 | |
| ResNet18-KNNMethod variant=KNN, Subsampling percentage=5%2026.03 | 61.5 | 34.4 | |
| SPM-RMethod variant=SPM (Retrieval), Subsampling percentage=100%2026.03 | 59.9 | 61.2 | |
| SPM-RMethod variant=SPM (Retrieval), Subsampling percentage=10%2026.03 | 54.8 | 34.2 | |
| SPM-RMethod variant=SPM (Retrieval), Subsampling percentage=5%2026.03 | 52 | 33.8 |