Involution: Inverting the Inherence of Convolution for Visual Recognition

Open paper
Benchmark Model Rank Results
image-classification-on-imagenetRedNet-152#803Top 1 Accuracy: 79.3%Number of params: 34MGFLOPs: 6.8
image-classification-on-imagenetRedNet-101#815Top 1 Accuracy: 79.1%Number of params: 25.6MGFLOPs: 4.7
image-classification-on-imagenetRedNet-50#875Top 1 Accuracy: 78.4%Number of params: 15.5MGFLOPs: 2.7
image-classification-on-imagenetRedNet-38#906Top 1 Accuracy: 77.6%Number of params: 12.4MGFLOPs: 2.2
image-classification-on-imagenetRedNet-26#968Top 1 Accuracy: 75.9%Number of params: 9.2MGFLOPs: 1.7

Rank counts only results with a code link.