Image Classification on Oxford-III (test)
76.23Top-1 AccuracyiiANET-L
Evaluation results
| Method | Links | ||
|---|---|---|---|
| iiANET-LSize=299², Train=90, Params=50.9M, FLOPs=13.45G2024.07 | 76.23 | 94.54 | |
| iiANET-BSize=299², Train=90, Params=25.2M, FLOPs=8.22G2024.07 | 74.04 | 93.98 | |
| Cross-ViT-BSize=224², Train=120, Params=105M, FLOPs=20.1G2024.07 | 73.1 | — | |
| Swin-TSize=224², Train=300, Params=29M, FLOPs=4.5G2024.07 | 72.3 | — | |
| DeiT-BSize=224², Train=120, Params=86M, FLOPs=17.5G2024.07 | 71 | — | |
| DiNAT-BSize=224², Train=90, Params=90M, FLOPs=13.7G2024.07 | 69.37 | 92.09 | |
| VMamba-TSize=224², Train=90, Params=30M, FLOPs=4.9G2024.07 | 67.8 | — | |
| CoAtNet-3Size=224², Train=90, Params=168M, FLOPs=32.53G2024.07 | 67.57 | 91.36 | |
| DilateFormer-BSize=224², Train=120, Params=48M, FLOPs=9.96G2024.07 | 64.85 | 88.18 | |
| BoT50Size=256², Train=90, Params=25.6M, FLOPs=3.18G2024.07 | 64.13 | 90 | |
| Dense201Size=224², Train=90, Params=20.0M, FLOPs=7.35G2024.07 | 62.55 | 86.84 | |
| ResNet-101Size=224², Train=90, Params=44.5M, FLOPs=14.58G2024.07 | 59.25 | 87.38 | |
| ResNet-50Size=224², Train=90, Params=25.6M, FLOPs=7.71G2024.07 | 59.07 | 86.61 | |
| ViT-L/16Size=224², Train=150, Params=304.3M, FLOPs=59.69G2024.07 | 53.19 | 83.28 | |
| MobileViT-SSize=256², Train=90, Params=6M, FLOPs=2G2024.07 | 51.89 | 84.52 | |
| ViT-B/16Size=224², Train=150, Params=86.6M, FLOPs=16.86G2024.07 | 51.23 | 82.31 | |
| EffNet-B5Size=224², Train=90, Params=30.4M, FLOPs=4.49G2024.07 | 47.54 | 78.92 |