参考博客:
安装:
https://zhuanlan.zhihu.com/p/22129880?refer=startdl (知乎专栏)
可视化显示(特征或者热力图):
caffe:
https://blog.csdn.net/jacke121/article/details/78136926
https://blog.csdn.net/liyaohhh/article/details/50990927
https://www.cnblogs.com/denny402/p/5103425.html caffemodel可视化(系列文章-内容很棒)
https://www.cnblogs.com/denny402/p/5105911.html 模型各层数据和参数可视化(系列文章-内容很棒)
keras:
https://blog.csdn.net/Einstellung/article/details/82858974?utm_source=blogxgwz9
Deep Visualization Toolbox
https://blog.csdn.net/zchang81/article/details/78095378
https://github.com/yosinski/deep-visualization-toolbox
https://www.bilibili.com/video/av7405645/ (使用教程)
热力图制作介绍:
http://bindog.github.io/blog/2018/02/10/model-explanation/
http://www.sohu.com/a/216216094_473283
第一个提出的作者:https://github.com/ramprs/grad-cam/
详细步骤:https://blog.csdn.net/qq_32799915/article/details/83503464
详细步骤博客的参考代码:https://github.com/jacobgil/keras-grad-cam
keras实现:https://github.com/keras-team/keras/issues/8447
https://jacobgil.github.io/deeplearning/class-activation-maps
3.darknet 可视化
参考博客:
yolo系列之yolo v3【深度解析】
模型结构可视化神器——Netron(支持tf, caffe, keras,mxnet等多种框架)