Wave-ViT: Unifying Wavelet and Transformers for Visual Representation Learning

Benchmark Model Rank Results
image-classification-on-imagenetWave-ViT-L#238Top 1 Accuracy: 85.5%Number of params: 57.5MGFLOPs: 14.8
image-classification-on-imagenetWave-ViT-B#308Top 1 Accuracy: 84.8%Number of params: 33.5MGFLOPs: 7.2
image-classification-on-imagenetWave-ViT-S#407Top 1 Accuracy: 83.9%Number of params: 22.7MGFLOPs: 4.7