Image Classification on CIFAR-100 59 (test)
0.846Top-1 AccuracyVVT-S
Evaluation results
| Method | Links | |
|---|---|---|
| VVT-SStyle=Linear Transformers, Multi-scale output=true, Resolution=224 x 224, Training Protocol=trained from scratch2022.06 | 0.846 | |
| PVTv2-B2Style=Transformers, Multi-scale output=true, Resolution=224 x 224, Training Protocol=trained from scratch2022.06 | 0.839 | |
| Twins-SVT-SStyle=Hybrid, Multi-scale output=true, Resolution=224 x 224, Training Protocol=trained from scratch2022.06 | 0.816 | |
| ResNet50Style=ConvNets, Multi-scale output=true, Resolution=224 x 224, Training Protocol=trained from scratch2022.06 | 0.792 | |
| PVT-SmallStyle=Transformers, Multi-scale output=true, Resolution=224 x 224, Training Protocol=trained from scratch2022.06 | 0.763 | |
| Swin-TStyle=Transformers, Multi-scale output=true, Resolution=224 x 224, Training Protocol=trained from scratch2022.06 | 0.748 | |
| TNT-SStyle=Transformers, Multi-scale output=false, Resolution=224 x 224, Training Protocol=trained from scratch2022.06 | 0.715 | |
| ViT-SmallStyle=Transformers, Multi-scale output=false, Resolution=224 x 224, Training Protocol=trained from scratch2022.06 | 0.474 |