CoAtNet: Marrying Convolution and Attention for All Data Sizes

Open paper
Benchmark Model Rank Results
image-classification-on-gashissdbCoAtNet-1#1Accuracy: 98.74Precision: 99.97F1-Score: 99.38
image-classification-on-imagenetCoAtNet-3 @384#39Top 1 Accuracy: 88.52%GFLOPs: 114
image-classification-on-imagenetCoAtNet-3 (21k)#77Top 1 Accuracy: 87.6%
image-classification-on-imagenetCoAtNet-2 (21k)#101Top 1 Accuracy: 87.1%
image-classification-on-imagenetCoAtNet-3#301Top 1 Accuracy: 84.5%Number of params: 168MGFLOPs: 34.7
image-classification-on-imagenetCoAtNet-2#340Top 1 Accuracy: 84.1%Number of params: 75MGFLOPs: 15.7
image-classification-on-imagenetCoAtNet-1#430Top 1 Accuracy: 83.3%Number of params: 42MGFLOPs: 8.4
image-classification-on-imagenetCoAtNet-0#602Top 1 Accuracy: 81.6%Number of params: 25MGFLOPs: 4.2

Rank counts only results with a code link.