Recent research in dynamic convolution shows substantial performance boost for efficient CNNs, due to the adaptive aggregation of K static convolution kernels. It has two limitations: (a) it increases the number of convolutional weights by K-times, and (b) the joint optimization of dynamic attention and static convolution kernels is challenging. In this paper, we revisit it from a new perspective of matrix decomposition and reveal the key issue is that dynamic convolution applies dynamic attention over channel groups after projecting into a higher dimensional latent space. To address this issue, we propose dynamic channel fusion to replace dynamic attention over channel groups. Dynamic channel fusion not only enables significant dimension reduction of the latent space, but also mitigates the joint optimization difficulty. As a result, our method is easier to train and requires significantly fewer parameters without sacrificing accuracy. Source code is at https://github.com/liyunsheng13/dcd.
翻译:最近对动态共变的研究显示,由于K静态共变内核的适应性整合,高效CNN的性能得到大幅提升。它有两个局限性:(a)它增加了K-时间的累进权重,和(b)联合优化动态关注和静态共变内核具有挑战性。在本文中,我们从矩阵分解的新角度重新审视它,并揭示关键问题是动态共变在投射到更高维度的潜层后对频道群群群进行动态关注。为了解决这一问题,我们提议动态集道取代频道群群的动态关注。动态集道不仅能够显著减少潜在空间的维度,而且可以减轻联合优化的困难。因此,我们的方法更容易培训,而且不需要牺牲准确性,要求大大降低参数。源代码在 https://github.com/liyunsheng13/dcd。