RepViT: Revisiting mobile CNN from the perspective of ViT

Summary

https://arxiv.org/pdf/2307.09283.pdf
In recent years, lightweight visual transformers (ViTs) have shown better performance on resource-constrained mobile devices than lightweight convolutional neural networks (CNN). high sex

Guess you like

Origin blog.csdn.net/hhhhhhhhhhwwwwwwwwww/article/details/132841455