DCGAN的Tensorflow算法实现

Python配套使用资料，更直观的看到协议分析图，更好地学习。(Python supporting the use of data, see the protocol analysis more intuitive, better learning. )

利用数据，进行对出行者出行状态的蒙特卡洛模拟方法(metro Carlo simulation)

在具有至少十五个节点的网络中模拟洪泛路由。每个数据包应包含一个计数器（n = 10），该计数器在每一跳上递减。(Simulate flood routing in a network with at least fifteen nodes. Each packet should contain a counter (n=10) that is decremented on each hop. When the counter gets to zero, the packet is discarded. Time is discrete, and each link can handle only one packet per time interval (i.e. only one packet in total may traverse the link in either direction). Nodes should include a buffer to queue any additional packets. Present your results in terms of the number of duplicate packets produced, and the congestion at each node, and suggest a possible solution for reducing this problem.)

Keras是一个高层神经网络库， Keras由纯Python编写而成并基Tensorflow或Theano。 Keras 为支持快 速实验而生，能够把你的idea迅速转换为结果，如果你有如下需求，请选择Keras： 简易和快速的原型设计（keras具有高度模块化，极简，和可扩充特性） 支持CNN和RNN，或二者的结合 支持任意的链接方案（包括多输入和多输出训练） 无缝CPU和GPU切换(Keras is a high-level neural network library. Keras is written from pure Python and is based on Tensorflow or Theano. Keras support fast Quick experimentation can transform your idea into results quickly. If you have the following requirements, please choose Keras: Simple and fast prototyping (keras is highly modular, minimized, and extensible). Support for CNN and RNN or the combination of the two Supports arbitrary link schemes (including multiple input and multiple output training). Seamless CPU and GPU handoff)

说明：  18个项目，通过前9天的数据，预测第10天PM2.5的含量，采用线性回归，损失函数为均方根误差(Through the data of the first nine days, the PM2.5 content of the 10th day was predicted for 18 projects, and linear regression was adopted, and the loss function was root mean square error.)

a menu that is created for a segmentation program(a simple menu)

截取手机图片提取图片文字自动百度进行搜索(Capture text automatically Baidu search)

编写简单的神经网络程序，使得通过神经网络对手写数字（将手写每个数字转换成784个数字表示）进行识别。在文件里放了100组训练集和10组训练集，训练准确率达到80%。事实证明，增大训练集至1000组可以将准确率提升至95%(Write a simple neural network program to recognize handwritten digits (converting each handwritten digit into 784 digital representations) by neural network. In the document, there are 100 training sets and 10 training sets, and the training accuracy rate reaches 80%. It has been proved that increasing the training set to 1000 groups can increase the accuracy rate to 95%.)

加上Coolprop库后，可以在Python软件中进行ORC的计算(With the addition of the Coolprop library, the ORC calculation can be done in Python software)

说明：  这是一个Python编写的多孔介质代码，代码标记为英语，希望能够对学习的人有所帮助。(This is a porous media code written by python. The code is marked in English. I hope it can be helpful to the learners.)