Play C# and Tensorflow.NET with Jupyter Notebook (Part 2)

In the previous part, we explained how to run C# and TensorFlow.NET in Jupyter Notebook. In this part, we will go one step further to introduce how we do linear regression using TensorFlow.NET within Jupyter Notebook.

Adding required packages and namespaces

using Tensorflow;
using static Tensorflow.Python;
using PlotNET;
using NumSharp;

Declare variables

int training_epochs = 1000;

float learning_rate = 0.01f;
int display_step = 50;

NumPyRandom rng = np.random;
NDArray train_X, train_Y;
int n_samples;

Prepare sample data for training

train_X = np.array(3.3f, 4.4f, 5.5f, 6.71f, 6.93f, 4.168f, 9.779f, 6.182f, 7.59f, 2.167f,
         7.042f, 10.791f, 5.313f, 7.997f, 5.654f, 9.27f, 3.1f);
train_Y = np.array(1.7f, 2.76f, 2.09f, 3.19f, 1.694f, 1.573f, 3.366f, 2.596f, 2.53f, 1.221f,
            2.827f, 3.465f, 1.65f, 2.904f, 2.42f, 2.94f, 1.3f);
n_samples = train_X.shape[0];

Construct the model

// tf Graph Input
var X = tf.placeholder(tf.float32);
var Y = tf.placeholder(tf.float32);

// Set model weights 
// We can set a fixed init value in order to debug
// var rnd1 = rng.randn<float>();
// var rnd2 = rng.randn<float>();
var W = tf.Variable(-0.06f, name: "weight");
var b = tf.Variable(-0.73f, name: "bias");

// Construct a linear model
var pred = tf.add(tf.multiply(X, W), b);

// Mean squared error
var cost = tf.reduce_sum(tf.pow(pred - Y, 2.0f)) / (2.0f * n_samples);

// Gradient descent
// Note, minimize() knows to modify W and b because Variable objects are trainable=True by default
var optimizer = tf.train.GradientDescentOptimizer(learning_rate).minimize(cost);

Start and iniialize Tensorflow session

// Create a TensorFlow session
var sess = tf.Session();

// Initialize the variables (i.e. assign their default value)
var init = tf.global_variables_initializer();

// Run the initializer
sess.run(init);

Fill training data

// Fit all training data
for (int epoch = 0; epoch < training_epochs; epoch++)
{
    foreach (var (x, y) in zip<float>(train_X, train_Y))
    {
        sess.run(optimizer, 
            new FeedItem(X, x),
            new FeedItem(Y, y));
    }

    // Display logs per epoch step
    if ((epoch + 1) % display_step == 0)
    {
        var c = sess.run(cost, 
            new FeedItem(X, train_X),
            new FeedItem(Y, train_Y));
        Console.WriteLine($"Epoch: {epoch + 1} cost={c} " + $"W={sess.run(W)} b={sess.run(b)}");
    }
}

Console.WriteLine("Optimization Finished!");

// Calculate training cost
var training_cost = sess.run(cost,
    new FeedItem(X, train_X),
    new FeedItem(Y, train_Y));

Draw chart base on the training data

var plotter = new Plotter();

plotter.Plot(
    train_X,
    train_Y,
    "Original data", ChartType.Scatter,"markers");
plotter.Plot(
    train_X,
    sess.run(W) * train_X + sess.run(b),
    "Fitted line", ChartType.Scatter, "Fitted line");

plotter.Show();

Draw chart base on the testing data

// Testing example
var test_X = np.array(6.83f, 4.668f, 8.9f, 7.91f, 5.7f, 8.7f, 3.1f, 2.1f);
var test_Y = np.array(1.84f, 2.273f, 3.2f, 2.831f, 2.92f, 3.24f, 1.35f, 1.03f);

Console.WriteLine("Testing... (Mean square loss Comparison)");

var testing_cost = sess.run(tf.reduce_sum(tf.pow(pred - Y, 2.0f)) / (2.0f * test_X.shape[0]),
    new FeedItem(X, test_X), 
    new FeedItem(Y, test_Y));

Console.WriteLine($"Testing cost={testing_cost}");

var diff = Math.Abs((float)training_cost - (float)testing_cost);
Console.WriteLine($"Absolute mean square loss difference: {diff}");

Draw chart base on the testing data

plotter.Plot(
    test_X,
    test_Y,
    "Testing data", ChartType.Scatter, "markers");

plotter.Plot(
    train_X,
    sess.run(W) * train_X + sess.run(b),
    "Fitted line", ChartType.Scatter);

plotter.Show();