Global Filter Networks for Image Classification

Benchmark Model Rank Results
domain-generalization-on-imagenet-aGFNet-S#30Top-1 accuracy %: 14.3
domain-generalization-on-imagenet-cGFNet-S#30mean Corruption Error (mCE): 53.8
image-classification-on-cifar-10GFNet-H-B#25Percentage correct: 99.0
image-classification-on-cifar-100GFNet-H-B#23Percentage correct: 90.3PARAMS: 54M
image-classification-on-flowers-102GFNet-H-B#19Accuracy: 98.8PARAMS: 54M
image-classification-on-imagenetGFNet-H-B#471Top 1 Accuracy: 82.9%Number of params: 54MGFLOPs: 8.6
image-classification-on-stanford-carsGFNet-H-B#9Accuracy: 93.2