BriefGPT.xyz
May, 2017
深度学习的贝叶斯压缩
Bayesian Compression for Deep Learning
HTML
PDF
Christos Louizos, Karen Ullrich, Max Welling
TL;DR
本研究通过采用贝叶斯视角,使用稀疏感知先验来修剪网络,使用Hierarchical priors修剪节点并使用后验不确定性确定编码权重的最优固定点精度,使得压缩率达到了最佳水平,并且仍然具有与优化速度或能量效率的方法相竞争的性能。
Abstract
compression
and
computational efficiency
in
deep learning
have become a problem of great significance. In this work, we argue that the mos
→