FasterViT: Fast Vision Transformers with Hierarchical Attention

Open paper
Benchmark Model Rank Results
image-classification-on-imagenetFasterViT-6#210Top 1 Accuracy: 85.8%Number of params: 1360MGFLOPs: 142
image-classification-on-imagenetFasterViT-5#229Top 1 Accuracy: 85.6%Number of params: 957.5MGFLOPs: 113
image-classification-on-imagenetFasterViT-4#250Top 1 Accuracy: 85.4%Number of params: 424.6MGFLOPs: 36.6
image-classification-on-imagenetFasterViT-3#299Top 1 Accuracy: 84.9%Number of params: 159.5MGFLOPs: 18.2
image-classification-on-imagenetFasterViT-2#368Top 1 Accuracy: 84.2%Number of params: 75.9MGFLOPs: 8.7
image-classification-on-imagenetFasterViT-1#491Top 1 Accuracy: 83.2%Number of params: 53.4MGFLOPs: 5.3
image-classification-on-imagenetFasterViT-0#614Top 1 Accuracy: 82.1%Number of params: 31.4MGFLOPs: 3.3

Rank counts only results with a code link.