Understanding Neural Networks Through Deep Visualization实现
来源:互联网 发布:网络新媒体是什么 编辑:程序博客网 时间:2024/06/03 21:44
这是Jason Yosinski, Jeff Clune, Anh Nguyen, Thomas Fuchs, and Hod Lipson在2015年出版在Computer Science上的一篇论文,主要讲述的是卷积神经网络可视化。https://github.com/yosinski/deep-visualization-toolbox 为本论文的公开代码,先讲述一下代码的运行环境及步骤,随后会贴出对本论文的理解。
运行环境:linux+caffe
步骤:
Step 0: Compile master branch of caffe
本代码运行的前提是,配置过caffe。因为配置caffe的过程中会出现一些依赖库,正是本代码所需要的。http://blog.csdn.net/u011204487/article/details/51596471是配置caffe的过程。注意Makefile.config中的CPU_ONLY := 1设置。
Step 1: Compile the deconv-deep-vis-toolbox branch of caffe
以下运行是在caffe根目录下
$ git remote add yosinski https://github.com/yosinski/caffe.git
$ git fetch --all
$ git checkout --track -b deconv-deep-vis-toolbox yosinski/deconv-deep-vis-toolbox
#这个地方很有可能报错,提示你更新代码会覆盖本地
#$ git checkout --track -b yosinski/deconv-deep-vis-toolbox
#$ git checkout --track -b deconv-deep-vis-toolbox
#只需将两个工具包分开下载就好
$ < edit Makefile.config to suit your system if not already done in Step 0 >
#编辑Makefile.config以适合自己的caffe
$ make clean
$ make -j
#这将会是一个漫长的等待,$ make -j4 可以稍微加快一下速度
$ make -j pycaffe
Step 2: Install prerequisites
这是要安装三个python-opencv scipy python-skimage东西,在安装 scipy 的时候可能会出现问题,只说可能,可以通过先安装pip来解决。
$ sudo apt-get install python-opencv scipy python-skimage
Step 3: Download and configure Deep Visualization Toolbox code
$ git clone https://github.com/yosinski/deep-visualization-toolbox
$ cd deep-visualization-toolbox
$ cp models/caffenet-yos/settings_local.template-caffenet-yos.py settings_local.py
$ < edit settings_local.py >
#在打开的setting_local.py文件中,有几处要修改成自己的文件路径,一定要看清楚有几处,这个特别重要
$ cd models/caffenet-yos/
$ ./fetch.sh
#这将会是一个极其漫长的等待
$ cd ../..
Step 4: Run it!
$ ./run_toolbox.py
运行结果:
当然还可以可视化自己的脸
- Understanding Neural Networks Through Deep Visualization实现
- Understanding Neural Networks Through Deep Visualization论文
- Understanding Neural Networks Through Deep Visualization
- [深度学习论文笔记][Visualizing] Understanding Neural Networks Through Deep Visualization
- Understanding the difficulty of training deep feedforward neural networks
- Understanding the difficulty of training deep feedforward neural networks (Xavier)
- 《Understanding the difficulty of training deep feedforward neural networks》笔记
- Understanding the difficulty of training deep feedforward neural networks
- 学习摘要:Methods for interpreting and understanding deep neural networks
- 【Deep Learning】笔记:Understanding the difficulty of training deep feedforward neural networks
- Deep learning-------------Neural networks
- 神经网络不同激活函数比较--读《Understanding the difficulty of training deep feedforward neural networks》
- 有效感受野--Understanding the Effective Receptive Field in Deep Convolutional Neural Networks
- deep learning 卷积神经网络的实现(Convolution Neural Networks)
- deep learning 卷积神经网络的实现(Convolution Neural Networks)
- deep learning 卷积神经网络的实现(Convolution Neural Networks)
- Neural Networks and Deep Learning
- Neural Networks and Deep Learning
- 关于android获取相册有些机型路径为空的解决办法
- 强大的DataGrid组件[9]_自定义头模板(HeaderTemplate)——Silverlight学习笔记[17]
- <contex<context:component-scan>t:component-scan>
- 深度学习FPGA实现基础知识8(Deep Learning(深度学习)Matlab实现--简单清晰的实验)
- 【补充知识点】e.target指代什么?
- Understanding Neural Networks Through Deep Visualization实现
- Java中如何遍历Map对象的4种方法
- 机器学习整理笔记——使用k-近邻算法对手写识别系统的测试
- Python正则匹配 -> 基本规则
- 回调函数
- Android 隐藏App的图标
- TCP 状态变迁图说明 【转】
- 网页WEB打印控件制作 开放源码可以调试
- greenDao分页加载