神经网络压缩加速

系统管理员 2022-07-14 10:27 196阅读 0赞

[深度学习模型压缩方法综述（一）][Link 1]

[深度网络模型压缩 - CNN Compression][- CNN Compression]

[神经网络压缩：Deep Compression][Deep Compression]

[神经网络压缩（1）：Deep Compression][1_Deep Compression]

[Deep Learning（深度学习）之（六）【深度神经网络压缩】Deep Compression （ICLR2016 Best Paper）][Deep Learning_Deep Compression _ICLR2016 Best Paper]

[CNN 模型压缩与加速算法综述][CNN]

pruning

Pruning Convolutional Neural Networks for Resource Efficient Inference

[神经网络压缩（3）：Learning both Weights and Connections for Efficient Neural Network][3_Learning both Weights and Connections for Efficient Neural Network]

[Dynamic Network Surgery for Efficient DNNs][]

pytorch基于卷积层通道剪枝的方法

# Network Slimming #

# Iterative Pruning #

Quantization

[神经网络压缩（8）Incremental Network Quantization: Towards Lossless CNNs with Low-precision Weights][8_Incremental Network Quantization_ Towards Lossless CNNs with Low-precision Weights]

https://github.com/Zhouaojun/Incremental-Network-Quantization

[网络压缩-量化方法对比][-]

[CNN网络量化 - Quantized Convolutional Neural Networks for Mobile Devices][CNN_ - Quantized Convolutional Neural Networks for Mobile Devices]

[tensorflow模型量化][tensorflow]

[tensorflow模型的定点化][tensorflow 1]

[Why are Eight Bits Enough for Deep Neural Networks?][Why are Eight Bits Enough for Deep Neural Networks]

[tensorflow 量化和裁剪的资料][tensorflow 2]

Binary Network

[【论文笔记】二值化神经网络（Binarized Neural Network）][Binarized Neural Network]

[CNN网络二值化--XNOR-Net： ImageNet Classification Using Binary Convolutional Neural Networks][CNN_--XNOR-Net_ ImageNet Classification Using Binary Convolutional Neural Networks]

[深度学习——缩减+召回加速网络训练][Link 2]

[【网络优化】超轻量级网络SqueezeNet算法详解][SqueezeNet]

[BinaryConnect： Training Deep Neural Networks with binary weights during propagations][BinaryConnect_ Training Deep Neural Networks with binary weights during propagations]

[CNN网络结构 - Refining Architectures of Deep Convolutional Neural Networks][CNN_ - Refining Architectures of Deep Convolutional Neural Networks]

[CNN网络分解--Factorized Convolutional Neural Networks][CNN_--Factorized Convolutional Neural Networks]

[论文笔记：DeepRebirth——从非权重层入手来进行模型压缩][DeepRebirth]

[论文笔记：ThiNet——一种filter级的模型裁剪算法][ThiNet_filter]

http://www.jianshu.com/u/f5c90c3856bb

https://github.com/yihui-he/channel-pruning

https://github.com/jacobgil/pytorch-pruning

https://github.com/eeric/channel\_prune

https://github.com/tostq/Caffe-Python-Tutorial/blob/master/prune.py

https://github.com/tostq/Caffe-Python-Tutorial/blob/master/quantize.py

https://github.com/IntelLabs/SkimCaffe

https://github.com/WWLoveBasketball/PRUNE-BASED-ON-TENSOTFLOW

https://github.com/Aaron-Zhao123/LeNet5-testing

https://github.com/garion9013/impl-pruning-TF

https://github.com/garion9013/impl-pruning-caffemodel

https://github.com/shekkizh/TensorflowProjects/tree/master/Model\_Pruning

https://github.com/ex4sperans/pruning\_with\_tensorflow

https://github.com/DNNToolBox/Net-Trim-v1

https://github.com/DAVIDNEWGATE/Project

https://github.com/yiwenguo/Dynamic-Network-Surgery

https://github.com/ZhouYuSong/caffe-pruned

https://github.com/sh0416/pruning

https://github.com/liuzhuang13/slimming

[Link 1]: http://blog.csdn.net/wspba/article/details/75671573
[- CNN Compression]: http://blog.csdn.net/linolzhang/article/details/78231341?locationNum=3&fps=1
[Deep Compression]: http://blog.csdn.net/shuzfan/article/details/51383809
[1_Deep Compression]: http://blog.csdn.net/cookie_234/article/details/70904905
[Deep Learning_Deep Compression _ICLR2016 Best Paper]: http://blog.csdn.net/boon_228/article/details/51718521
[CNN]: http://blog.csdn.net/QcloudCommunity/article/details/77719498
[3_Learning both Weights and Connections for Efficient Neural Network]: http://blog.csdn.net/cookie_234/article/details/72904452
[Dynamic Network Surgery for Efficient DNNs]: http://blog.csdn.net/zhangjunhit/article/details/53218211
[8_Incremental Network Quantization_ Towards Lossless CNNs with Low-precision Weights]: http://blog.csdn.net/cookie_234/article/details/75386737
[-]: http://blog.csdn.net/shuzfan/article/details/51678499
[CNN_ - Quantized Convolutional Neural Networks for Mobile Devices]: http://blog.csdn.net/cv_family_z/article/details/52300275
[tensorflow]: http://blog.csdn.net/u011961856/article/details/76736103
[tensorflow 1]: http://blog.csdn.net/baidu_37967613/article/details/72678003
[Why are Eight Bits Enough for Deep Neural Networks]: https://petewarden.com/2015/05/23/why-are-eight-bits-enough-for-deep-neural-networks/
[tensorflow 2]: https://www.cnblogs.com/powercsj/p/7580737.html
[Binarized Neural Network]: http://blog.csdn.net/elaine_bao/article/details/50950969
[CNN_--XNOR-Net_ ImageNet Classification Using Binary Convolutional Neural Networks]: http://blog.csdn.net/cv_family_z/article/details/51917143
[Link 2]: http://blog.csdn.net/shuzfan/article/details/51439672
[SqueezeNet]: http://blog.csdn.net/shenxiaolu1984/article/details/51444525
[BinaryConnect_ Training Deep Neural Networks with binary weights during propagations]: http://blog.csdn.net/cv_family_z/article/details/53158782
[CNN_ - Refining Architectures of Deep Convolutional Neural Networks]: http://blog.csdn.net/cv_family_z/article/details/52957065
[CNN_--Factorized Convolutional Neural Networks]: http://blog.csdn.net/cv_family_z/article/details/52457476
[DeepRebirth]: http://blog.csdn.net/wspba/article/details/76098493
[ThiNet_filter]: http://blog.csdn.net/wspba/article/details/77427960