ConViT: Improving Vision Transformers
with Soft Convolutional Inductive Biases
Stephane d’Ascoli 1 2 Hugo Touvron 2 Matthew L. Leavitt 2 Ari S. Morcos 2 Giulio Biroli 1 2 Levent Sagun 2
Abstract 1. Introduction
The success of deep learning over the last decade has largely
been fueled by models with strong inductive biases, al-
Convolutional architectures have proven ex-
...


雷达卡




京公网安备 11010802022788号







