Image Classification on ImageNet-1K (val) (Top-1, Top-5, and ReaL Accuracy)
84.7Top-1 AccuracyMaxViT
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| MaxViTModel=MaxViT2026.06 | 84.7 | 96.9 | 88.5 | |
| ConvNeXtModel=ConvNeXt2026.06 | 83.7 | 96.7 | 88.3 | |
| SwinModel=Swin2026.06 | 83.1 | 96.4 | 87.7 | |
| ViT-B + DnAParameters=100.7M, GFLOPs=21.12026.06 | 81.9 | 95.8 | — | |
| ViT-B + Diff. Attn.Parameters=86.6M, GFLOPs=17.62026.06 | 81.5 | 95.6 | — | |
| ViT-B + Cog Attn.Parameters=86.6M, GFLOPs=17.62026.06 | 81.5 | 95.7 | — | |
| ViT-BParameters=86.6M, GFLOPs=17.62026.06 | 81.1 | 95.5 | — | |
| ResNet-50Model=ResNet-502026.06 | 80.3 | 95.1 | 86.1 | |
| EfficientNetModel=EfficientNet2026.06 | 78.4 | 94.3 | 84.7 |