Rethinking Channel Dimensions for Efficient Model Design

Open paper
Benchmark Model Rank Results
image-classification-on-imagenetReXNet-R_3.0#300Top 1 Accuracy: 84.5%Number of params: 34.8M
image-classification-on-imagenetReXNet-R_2.0#437Top 1 Accuracy: 83.2%Number of params: 16.5M
image-classification-on-imagenetReXNet_3.0#477Top 1 Accuracy: 82.8%Number of params: 34.7MGFLOPs: 3.4
image-classification-on-imagenetReXNet_2.0#597Top 1 Accuracy: 81.6%Number of params: 19MGFLOPs: 1.5
image-classification-on-imagenetReXNet_1.5#674Top 1 Accuracy: 80.3%Number of params: 9.7MGFLOPs: 0.86
image-classification-on-imagenetReXNet_1.3#709Top 1 Accuracy: 79.5%Number of params: 7.6MGFLOPs: 0.66
image-classification-on-imagenetReXNet_1.0#808Top 1 Accuracy: 77.9%Number of params: 4.8MGFLOPs: 0.40
image-classification-on-imagenetReXNet_0.9#827Top 1 Accuracy: 77.2%Number of params: 4.1MGFLOPs: 0.35
image-classification-on-imagenetReXNet_0.6#913Top 1 Accuracy: 74.6%Number of params: 2.7M

Rank counts only results with a code link.