Which Transformer to Favor: A Comparative Analysis of Efficiency in Vision Transformers

Open paper
Benchmark Model Rank Results
image-classification-on-imagenetCaiT-S24#292Top 1 Accuracy: 84.91%
image-classification-on-imagenetXCiT-S#434Top 1 Accuracy: 83.65%
image-classification-on-imagenetWave-ViT-S#436Top 1 Accuracy: 83.61%
image-classification-on-imagenetSwinV2-Ti#502Top 1 Accuracy: 83.09%
image-classification-on-imagenetViT-S#559Top 1 Accuracy: 82.54%
image-classification-on-imagenetEViT (delete)#591Top 1 Accuracy: 82.29%
image-classification-on-imagenetSTViT-Swin-Ti#592Top 1 Accuracy: 82.22%
image-classification-on-imagenetToMe-ViT-S#609Top 1 Accuracy: 82.11%
image-classification-on-imagenetEViT (fuse)#626Top 1 Accuracy: 81.96%
image-classification-on-imagenetGFNet-S#685Top 1 Accuracy: 81.33%
image-classification-on-imagenetDynamicViT-S#709Top 1 Accuracy: 81.09%
image-classification-on-imagenetTokenLearner-ViT-8#728Top 1 Accuracy: 80.66%
image-classification-on-imagenetCoaT-Ti#872Top 1 Accuracy: 78.42%
image-classification-on-imagenetPoly-SA-ViT-S#879Top 1 Accuracy: 78.34%
image-classification-on-imagenetEfficientFormer-V2-S0#1060Top 1 Accuracy: 71.53%

Rank counts only results with a code link.