Fast Vision Transformers with HiLo Attention

Open paper
Benchmark Model Rank Results
image-classification-on-imagenetLITv2-B|384#290Top 1 Accuracy: 84.7%Number of params: 87MGFLOPs: 39.7
image-classification-on-imagenetLITv2-B#403Top 1 Accuracy: 83.6%GFLOPs: 13.2
image-classification-on-imagenetLITv2-M#432Top 1 Accuracy: 83.3%Number of params: 49MGFLOPs: 7.5
image-classification-on-imagenetLITv2-S#568Top 1 Accuracy: 82%Number of params: 28MGFLOPs: 3.7

Rank counts only results with a code link.