Keras.NET（1）- 安裝

0.寫在前面：

作者基于tensorflow 2.X，利用自在的keras接口完成了神經網絡的訓練，現在想把模型跑在c#端，但使用TensorFlowsharp好像還是tensorflow 1.X的語言風格，改動起來較為困難。是以想尋找c#接口的keras。經過一番搜尋，找到兩個資源，分别為keras-sharp（https://github.com/tcmxx/keras-sharp）和Keras.NET（https://github.com/SciSharp/Keras.NET/），keras-sharp看起來還久沒有維護和更新了，是以放棄，嘗試使用Keras.NET。

1.依賴項的安裝：

• Python 2.7 - 3.7, Link: https://www.python.org/downloads/
• Install keras, numpy and one of the backend (Tensorflow/CNTK/Theano). Please see on how to configure: https://keras.io/backend/

 我是安裝的python3.7.0（别忘了點選添加到環境變量），

python3.7.0安裝完畢之後使用pip指令依次安裝numpy，tensorflow，keras即可。這裡注意：安裝之後最好更新一下pip，

python -m pip install --upgrade pip
           

然後我是安裝的tensorflow2.1 cpu版本，使用:

pip install tensorflow-cpu==2.1.0 -i https://mirrors.aliyun.com/pypi/simple
           

2.Keras.NET的安裝

2.1 通過Nuget指令行安裝

Install-Package Keras.NET
           

2.2 通過Nuget界面安裝

安裝完成之後自動添加了Keras，Numpy.Bare，Python.Runtime等

3.測試

拿着官方的minist例程測試一下，代碼如下：

using System;
using System.Collections.Generic;
using System.ComponentModel;
using System.Data;
using System.Drawing;
using System.Linq;
using System.Text;
using System.Threading.Tasks;
using System.Windows.Forms;

using Keras.Datasets;
using System;
using System.Collections.Generic;
using System.Text;
using Numpy;
using K = Keras.Backend;
using Keras;
using Keras.Models;
using Keras.Layers;
using Keras.Utils;
using Keras.Optimizers;
using System.IO;



namespace keras.net
{
    public partial class Form1 : Form
    {
        public Form1()
        {
            InitializeComponent();
        }

        private void Form1_Load(object sender, EventArgs e)
        {

        }

        private void test()
        {

            int batch_size = 200;
            int num_classes = 10;
            int epochs = 10;

            // input image dimensions
            int img_rows = 28, img_cols = 28;

            Shape input_shape = null;

            // the data, split between train and test sets
            var ((x_train, y_train), (x_test, y_test)) = MNIST.LoadData();

            if (K.ImageDataFormat() == "channels_first")
            {
                x_train = x_train.reshape(x_train.shape[0], 1, img_rows, img_cols);
                x_test = x_test.reshape(x_test.shape[0], 1, img_rows, img_cols);
                input_shape = (1, img_rows, img_cols);
            }
            else
            {
                x_train = x_train.reshape(x_train.shape[0], img_rows, img_cols, 1);
                x_test = x_test.reshape(x_test.shape[0], img_rows, img_cols, 1);
                input_shape = (img_rows, img_cols, 1);
            }

            x_train = x_train.astype(np.float32);
            x_test = x_test.astype(np.float32);
            x_train /= 255;
            x_test /= 255;
            Console.WriteLine("x_train shape: " + x_train.shape);
            Console.WriteLine(x_train.shape[0] + " train samples");
            Console.WriteLine(x_test.shape[0] + " test samples");

            // convert class vectors to binary class matrices
            y_train = Util.ToCategorical(y_train, num_classes);
            y_test = Util.ToCategorical(y_test, num_classes);

            // Build CNN model
            var model = new Sequential();
            model.Add(new Conv2D(32, kernel_size: (3, 3).ToTuple(),
                                 activation: "relu",
                                 input_shape: input_shape));
            model.Add(new Conv2D(64, (3, 3).ToTuple(), activation: "relu"));
            model.Add(new MaxPooling2D(pool_size: (2, 2).ToTuple()));
            model.Add(new Dropout(0.25));
            model.Add(new Flatten());
            model.Add(new Dense(128, activation: "relu"));
            model.Add(new Dropout(0.5));
            model.Add(new Dense(num_classes, activation: "softmax"));

            model.Compile(loss: "categorical_crossentropy",
              optimizer: new Adadelta(), metrics: new string[] { "accuracy" });

            model.Fit(x_train, y_train,
                      batch_size: batch_size,
                      epochs: epochs,
                      verbose: 1,
                      validation_data: new NDarray[] { x_test, y_test });

            model.Save("model.h5");
            model.SaveTensorflowJSFormat("./");

            var score = model.Evaluate(x_test, y_test, verbose: 0);
            Console.WriteLine("Test loss:" + score[0]);
            Console.WriteLine("Test accuracy:" + score[1]);
        }

        private void button1_Click(object sender, EventArgs e)
        {
            test();
        }
    }
}
           

程式會下載下傳mnist資料集，非常慢又經常失敗，從控制台的輸出資訊可以找到下載下傳原網址如下：

https://s3.amazonaws.com/img-datasets/mnist.npz

我們可以随便從網上找一個用于keras的mnist資料集下載下傳即可，我是用的這個https://download.csdn.net/download/lsldd/11502401

下載下傳完成之後将mnist.npz檔案放到keras資料集路徑下（我的電腦是C:\Users\Administrator\.keras\datasets）

10個epoch之後損失和準确率如下：

注意：出現錯誤時，可以先試試解決方案平台改成X64，還不行的話删除所有依賴（删幹淨！）重新安裝，不要使用anaconda安裝依賴包！

ref：

https://blog.csdn.net/kinfey/article/details/99691593

https://scisharp.github.io/Keras.NET/index.html

https://github.com/SciSharp/Numpy.NET