Image Classification on ImageNet-1K (val) (Accuracy, Time & FLOPs Efficiency)
80.18AccuracySViTE+-Small
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| SViTE+-SmallTokens (%)=95%, Sparsity Type=Unstructured, Sparsity Ratio=50%2021.06 | 80.18 | 4.4 | 49.32 | |
| SViTE+-SmallTokens (%)=90%, Sparsity Type=Unstructured, Sparsity Ratio=50%2021.06 | 79.91 | 7.63 | 52.38 | |
| SViTE+-SmallTokens (%)=100%, Sparsity Type=Unstructured, Sparsity Ratio=50%2021.06 | 79.72 | 0 | 46.26 | |
| S2ViTE+-SmallTokens (%)=100%, Sparsity Type=Structured, Sparsity Ratio=40%2021.06 | 79.22 | 22.65 | 31.63 | |
| S2ViTE+-SmallTokens (%)=95%, Sparsity Type=Structured, Sparsity Ratio=40%2021.06 | 78.44 | 27.17 | 37.76 | |
| S2ViTE+-SmallTokens (%)=90%, Sparsity Type=Structured, Sparsity Ratio=40%2021.06 | 78.16 | 29.21 | 41.5 | |
| SViTE+-SmallTokens (%)=70%, Sparsity Type=Unstructured, Sparsity Ratio=50%2021.06 | 77.9 | 19.77 | 63.95 | |
| SigLIxPTrain-free=true, Shots=16, Backbone=ViT-B/162024.11 | 77.9 | — | — | |
| S2ViTE+-SmallTokens (%)=70%, Sparsity Type=Structured, Sparsity Ratio=40%2021.06 | 74.77 | 39.1 | 54.96 | |
| DMNTrain-free=false, Shots=16, Backbone=ViT-B/162024.11 | 74.7 | — | — | |
| APE-TTrain-free=false, Shots=16, Backbone=ViT-B/162024.11 | 74.3 | — | — | |
| CasPLTrain-free=false, Shots=16, Backbone=ViT-B/162024.11 | 74.2 | — | — | |
| Tip-Adapter-FTrain-free=false, Shots=16, Backbone=ViT-B/162024.11 | 73.7 | — | — | |
| PromptSRCTrain-free=false, Shots=16, Backbone=ViT-B/162024.11 | 73.2 | — | — | |
| DMN-TFTrain-free=true, Shots=16, Backbone=ViT-B/162024.11 | 72.6 | — | — | |
| MaPLeTrain-free=false, Shots=16, Backbone=ViT-B/162024.11 | 72.3 | — | — | |
| APETrain-free=true, Shots=16, Backbone=ViT-B/162024.11 | 72.1 | — | — | |
| TIP-XTrain-free=true, Shots=16, Backbone=ViT-B/162024.11 | 71.1 | — | — | |
| Clip-AdapterTrain-free=false, Shots=16, Backbone=ViT-B/162024.11 | 71.1 | — | — | |
| Linear ProbeTrain-free=false, Shots=16, Backbone=ViT-B/162024.11 | 67.3 | — | — |