技術研討方向
https://zhuanlan.zhihu.com/p/58805980
Image Not Showing
Possible Reasons
- The image file may be corrupted
- The server hosting the image is unavailable
- The image path is incorrect
- The image format is not supported
Learn More →
Redundancy in Weights
https://www.cnblogs.com/wujianming-110117/p/12702802.html
1. 剪枝(Pruning)
90年代只有neural network:
- Magnitude-based: 對網絡中每個hidden unit與其絕對值相關的weight decay來最小化hidden unit數量
- Optimal brain damage(OBD)/Optimal brain surgeon(OBS):基于损失函数相對於權重的二階導數(對權重向量来說即Hessian矩陣)来衡量網絡中權重的重要程度,然後對其進行裁減
2012年後,Deep neural network崛起
- 非結構化剪枝(Unstructured pruning):早期的方法多屬於非結構化剪枝,裁減神經元,但kernel會變得很稀疏,得到中間很多元素為0的矩陣,因此很難實質的提升性能。
- 結構化剪枝(Structured pruning):近期研究集中在結構化剪枝,可進一步細分為:
- channel-wise
- filter-wise
- shape-wise
Image Not Showing
Possible Reasons
- The image file may be corrupted
- The server hosting the image is unavailable
- The image path is incorrect
- The image format is not supported
Learn More →
詳細介紹: 闲话模型压缩之网络剪枝(Network Pruning)篇
閒話模型壓縮(Model compression)之網絡剪枝
Filter pruning: (JN)
paper:
PRUNING FILTERS FOR EFFICIENT CONVNETS
論文中文講解1
論文中文講解2(中文講解也很多篇)
code:
Pruning Filters for Efficient Convnets
Image Not Showing
Possible Reasons
- The image file may be corrupted
- The server hosting the image is unavailable
- The image path is incorrect
- The image format is not supported
Learn More →
1200 (github有很多人寫這篇)github 搜尋pruning filters for efficient convnets
**Channel pruning:
paper:
Learning Efficient Convolutional Networks through Network Slimming
code:
Network Slimming
Image Not Showing
Possible Reasons
- The image file may be corrupted
- The server hosting the image is unavailable
- The image path is incorrect
- The image format is not supported
Learn More →
564
YOLOv3-model-pruning
Image Not Showing
Possible Reasons
- The image file may be corrupted
- The server hosting the image is unavailable
- The image path is incorrect
- The image format is not supported
Learn More →
1400
Code:
Pruning (Keras):
- 修剪非必要權重
- Pruning in Keras example
Deep-Compression-PyTorch
Image Not Showing
Possible Reasons
- The image file may be corrupted
- The server hosting the image is unavailable
- The image path is incorrect
- The image format is not supported
Learn More →
260 / 用簡單的 net 實作 Deep CompressionList of Weight and Filter pruning 很多paper及對應的github
package:
Distiller Distiller is an open-source Python package for neural network compression research.(PyTorch)
- Pruning包含:filter/channel
- Quantization
Image Not Showing
Possible Reasons
- The image file may be corrupted
- The server hosting the image is unavailable
- The image path is incorrect
- The image format is not supported
Learn More →
相關經典 paper:
- A Survey of Model Compression and Acceleration for Deep Neural Networks 介紹模型壓縮方法的各種來龍去脈
- Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding 作者將剪枝、量化和編碼等技術結合起來,在不顯著影響準確性的前提下,將存儲需求減少35x(AlexNet)至49x(VGG-19)
- Learning both Weights and Connections for Efficient Neural Networks Deep Compression 中剪枝的方法
- Single Shot Structured Pruning Before Training
Resource:
Pruning 介紹
https://kknews.cc/zh-tw/science/g8kmo8e.html
Three Dimensional Convolutional Neural Network Pruning with Regularization-Based Method
2.量化(Quantization)
概念:
量化就是將神經網絡的浮點算法轉換為定點。量化有若干相似的術語。低精度(Low precision)可能是最通用的概念。常規精度一般使用FP32(32位浮點,單精度)存儲模型權重;低精度則表示FP16(半精度浮點),INT8(8位的定點整數)等等數值格式。不過目前低精度往往指代INT8。
參考資源:
https://jackwish.net/2019/neural-network-quantization-introduction-chn.html
Paper with code:
What Do Compressed Deep Neural Networks Forget?
Image Not Showing
Possible Reasons
- The image file may be corrupted
- The server hosting the image is unavailable
- The image path is incorrect
- The image format is not supported
Learn More →
16,032Training with Quantization Noise for Extreme Model Compression
Image Not Showing
Possible Reasons
- The image file may be corrupted
- The server hosting the image is unavailable
- The image path is incorrect
- The image format is not supported
Learn More →
11,3003.(Low-rank factorization)
4.(Knowledge distillation)
參考資料大補帖
https://github.com/memoiry/Awesome-model-compression-and-acceleration/blob/master/README.md