1 paper · 2 filters
Vladimir Korviakov, Denis Koposov
Most of the computer vision architectures nowadays are built upon the well-known foundation operations: fully-connected layers, convolutions and multi-head self-attention blocks. I…