Network trimming: a data driven neuron pruning approach towards efficient deep architectures paper translation / Notes

Network Trimming: A Data-Driven Neuron Pruning Approach towards Efficient Deep Architectures Thesis translation / note _weixin_43998119 The blog of -CSDN Blog The theme of the graduation thesis is proposed in this paper APoZ The method is applied to ResNet The effect of pruning , I hereby read the abstract of the catalogue , Introduction 2 , Related work III , Zero activation and network pruning 3.1 VGG-16 Zero activation on 3.2 Network pruning and retraining IV , Experimental Abstract: neural networks at the forefront are becoming deeper and wider , Although their performance improves with the increase of the number of layers and neurons , But designing an efficient network with low requirements for computing power and storage is not simple , Because it means a lot of experiments and fine adjustments . This paper introduces a method to evaluate the importance of neurons based on their output to a large data set , And eliminate unimportant neurons accordingly , Pruning algorithm to fine tune this iterative process . This algorithm https://blog.csdn.net/weixin_43998119/article/details/115779336

  Code link ：

GitHub - jjeamin/Network_Trimming_Pytorch: Implementation network trimming using pytorchImplementation network trimming using pytorch. Contribute to jjeamin/Network_Trimming_Pytorch development by creating an account on GitHub.https://github.com/jjeamin/Network_Trimming_Pytorch

https://github.com/rezabonyadi/deep_networks_trimminghttps://github.com/rezabonyadi/deep_networks_trimming