信步闲庭v - 简书

IP属地：云南

深度学习调参技巧
从粗到细实践中，一般先进行初步范围搜索，然后根据好结果出现的地方，再缩小范围进行更精细的搜索。先参考相关论文，以论文中给出的参数作为初始参数...

1109 0 0
Knowledge transfer summary-2
所谓“知识精炼”我的理解就是将一个训练好的复杂模型的“知识”迁移到一个结构更为简单的网络中，或者通过简单的网络去学习复杂模型中“知识”或模仿复杂...

0.3 2031 0 2

网络量化小结
网络量化作为一种重要的模型压缩方法，大致可以分为两类：直接降低参数精度典型的工作包括二值网络，三值网络以及XNOR-Net. HORQ和Net...

1151 0 0
Speeding-up CNN using CP-Decomposition
Approach We propose a simple two-step approach for speeding up convoluti...

786 0 0
Exploiting Linear Structure Within CNN
Approach Matrix Decomposition Higher Order Tensor Approximations Monochr...

713 0 0
Dynamic Network Surgery
Approach Han song recently propose to compress DNN by deleting unimporta...

457 0 0
Speed up CNN with Low Rank Expansions
Approach Experiment References:Speeding up Convolutional Neural Networks...

734 0 2

Learning Structured Sparsity in DNN
Approach The optimization target of learning the filter-wise and channel...

698 0 0
Fixed-point Factorized Networks
Approach Fixed-point Factorization Full-precision Weights Recovery The q...

576 0 0