Image Classification on STL10 standard (test)
72.19AccuracyE2E
Evaluation results
| Method | Links | ||
|---|---|---|---|
| E2EBackbone=Swin-Tiny Transformer, Training Protocol=End-to-End2023.09 | 72.19 | 0 | |
| Parallel TRGLBackbone=Swin-Tiny Transformer, Number of modules=4, Training Protocol=Parallel2023.09 | 67.92 | 50 | |
| Parallel VanGLBackbone=Swin-Tiny Transformer, Number of modules=4, Training Protocol=Parallel2023.09 | 67 | 55 | |
| InfoProLBackbone=Swin-Tiny Transformer, Number of modules=4, Training Protocol=Parallel2023.09 | 66.89 | 45 | |
| InfoProBackbone=Swin-Tiny Transformer, Number of modules=4, Training Protocol=Parallel2023.09 | 64.61 | 38 |