High-Performance Large-Scale Image Recognition Without Normalization

Open paper
Benchmark Model Rank Results
image-classification-on-imagenetNFNet-F4+#25Top 1 Accuracy: 89.2%Number of params: 527MGFLOPs: 367
image-classification-on-imagenetNFNet-F6 w/ SAM#145Top 1 Accuracy: 86.5%Number of params: 438.4MGFLOPs: 377.28
image-classification-on-imagenetNFNet-F5 w/ SAM#166Top 1 Accuracy: 86.3%Number of params: 377.2MGFLOPs: 289.76
image-classification-on-imagenetNFNet-F5#189Top 1 Accuracy: 86.0%Number of params: 377.2MGFLOPs: 289.76
image-classification-on-imagenetNFNet-F4#198Top 1 Accuracy: 85.9%Number of params: 316.1MGFLOPs: 215.24
image-classification-on-imagenetNFNet-F3#217Top 1 Accuracy: 85.7%Number of params: 254.9MGFLOPs: 114.76
image-classification-on-imagenetNFNet-F2#272Top 1 Accuracy: 85.1%Number of params: 193.8MGFLOPs: 62.59
image-classification-on-imagenetNFNet-F1#315Top 1 Accuracy: 84.7%Number of params: 132.6MGFLOPs: 35.54
image-classification-on-imagenetNFNet-F0#440Top 1 Accuracy: 83.6%Number of params: 71.5MGFLOPs: 12.38

Rank counts only results with a code link.