Image Classification on ImageNet HR
92.8Top-1 AccuracyViT+Jumbo
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ViT+JumboSize=D=768, J=3, Throughput=3.1K, Resolution=224x2242025.02 | 92.8 | 99 | |
| ViT+RegistersSize=D=768, R=16, Throughput=2.9K, Resolution=224x2242025.02 | 92 | 98.9 | |
| ViT+JumboSize=D=384, J=6, Throughput=7.6K, Resolution=224x2242025.02 | 90.7 | 98.2 | |
| ViT+RegistersSize=D=384, R=16, Throughput=6.6K, Resolution=224x2242025.02 | 90 | 98 | |
| MobileNetV4Size=hybrid-medium, Throughput=8.8K, Resolution=224x2242025.02 | 87.2 | 96.8 | |
| ViT+JumboSize=D=192, J=6, Throughput=20.4K, Resolution=224x2242025.02 | 87 | 96.7 | |
| EfficientViTSize=B1, Throughput=9.8K, Resolution=224x2242025.02 | 85.8 | 96.4 | |
| MobileNetV4Size=conv-medium, Throughput=11.2K, Resolution=224x2242025.02 | 84.7 | 95.8 | |
| SHViTSize=S3, Throughput=23.7K, Resolution=224x2242025.02 | 84 | 95.4 | |
| ViT+RegistersSize=D=192, R=16, Throughput=16.7K, Resolution=224x2242025.02 | 83.7 | 95.6 | |
| ViT+JumboSize=D=128, J=6, Throughput=31.3K, Resolution=224x2242025.02 | 83.4 | 94.9 | |
| SHViTSize=S2, Throughput=34.5K, Resolution=224x2242025.02 | 80.9 | 93.9 | |
| ViT+JumboSize=D=96, J=6, Throughput=43.7K, Resolution=224x2242025.02 | 78.6 | 92.8 | |
| SHViTSize=S1, Throughput=42.1K, Resolution=224x2242025.02 | 77.6 | 92.7 | |
| EfficientViTSize=B0, Throughput=25.2K, Resolution=224x2242025.02 | 75.8 | 91 | |
| MobileNetV4Size=conv-small, Throughput=33.7K, Resolution=224x2242025.02 | 75.7 | 91.3 | |
| ViT+RegistersSize=D=128, R=16, Throughput=25.5K, Resolution=224x2242025.02 | 69.7 | 88.9 |