public void ARange()
{
var np = new NDArray <int>();
np.ARange(3);
Assert.IsTrue(np.ToString().Equals("array([0, 1, 2])"));
np.ARange(7, 3);
Assert.IsTrue(np.ToString().Equals("array([3, 4, 5, 6])"));
np.ARange(7, 3, 2);
Assert.IsTrue(np.ToString().Equals("array([3, 5])"));
}
public void ReShape()
{
var np = new NDArray <int>();
var np2 = np.ARange(6).ReShape(3, 2);
Assert.IsTrue(np2.ToString().Equals("array([[0, 1], [2, 3], [4, 5]])"));
}
public void ReshapeNegative()
{
var np = new NDArray <int>();
np.ARange(12);
np.ReShape(-1, 2);
Assert.IsTrue(np.Shape[0] == 6);
Assert.IsTrue(np.Shape[1] == 2);
np.ARange(12);
np.ReShape(2, -1);
Assert.IsTrue(np.Shape[0] == 2);
Assert.IsTrue(np.Shape[1] == 6);
np.ARange(12);
np.ReShape(1, 3, 4);
np.ReShape(-1, 3);
Assert.IsTrue(np.Shape[0] == 4);
Assert.IsTrue(np.Shape[1] == 3);
np.ARange(12);
np.ReShape(1, 3, 4);
np.ReShape(3, -1);
Assert.IsTrue(np.Shape[0] == 3);
Assert.IsTrue(np.Shape[1] == 4);
np.ARange(100 * 100 * 3);
np.ReShape(100, 100, 3);
np.ReShape(-1, 3);
Assert.IsTrue(np.Shape[0] == 10000);
Assert.IsTrue(np.Shape[1] == 3);
np.ARange(15801033);
np.ReShape(2531, 2081, 3);
np.ReShape(-1, 3);
Assert.IsTrue(np.Shape[0] == 5267011);
Assert.IsTrue(np.Shape[1] == 3);
}
public void SeedsClassificationTest()
{
var filename = @"data/seeds_dataset.csv";
var np = new NDArray <int>();
var n_hidden_nodes = 5; // nodes in hidden layers
var l_rate = 0.6; // learning rate
var n_epochs = 1000; // number of training epochs
var n_folds = 4; // number of folds for cross-validation
/* =================================
* Read data (X,y) and normalize X
* ================================= */
Console.WriteLine($"Reading '{filename}'...");
(var X, var y) = Utils.ReadCsv(filename); // read as matrix of floats and int
// normalize
X.Normalize();
// extract shape of X
(var N, var d) = X.Shape;
var nClasses = y.Unique().Length;
Console.WriteLine($"X.shape = {X.Shape}, y.shape = {y.Shape}");
Console.WriteLine($"size = {X.Data.Count}, dimesion = {X.NDim}, number of classes = {nClasses}");
Console.WriteLine($"hidden nodes = {n_hidden_nodes}, learning rate = {l_rate}, epochs = {n_epochs}");
/* ===================================
* Create cross-validation folds
* These are a list of a list of indices for each fold
* =================================== */
var idx_all = np.ARange(N, 0);
var idx_folds = Utils.CrossValFolds(N, 4);
/* ===================================
* Train and evaluate the model on each fold
* =================================== */
// List<acc_train, acc_test = list(), list() # training/test accuracy score
Console.WriteLine("Training and cross-validating...");
List <double> acc_train = new List <double>();
List <double> acc_test = new List <double>();
for (int i = 0; i < n_folds; i++)
{
// Collect training and test data from folds
var idx_test = idx_folds[i];
var idx_train = idx_all.Delete(idx_test);
var X_train = X[idx_train];
var y_train = y[idx_train];
var X_test = X[idx_test];
var y_test = y[idx_test];
// Build neural network classifier model and train
NetworkModel model = new NetworkModel();
model.Layers.Add(new NeuralLayer(d, "INPUT"));
model.Layers.Add(new NeuralLayer(n_hidden_nodes, "HIDDEN"));
model.Layers.Add(new NeuralLayer(nClasses, "OUTPUT"));
model.Build();
model.Train(X_train, y_train, iterations: n_epochs, learningRate: l_rate);
// Make predictions for training and test data
var y_train_predict = model.Predict(X_train);
var y_test_predict = model.Predict(X_test);
acc_train.Add(100 * y_train.Sum(y_train_predict) / y_train.Length);
acc_test.Add(100 * y_test.Sum(y_test_predict) / y_test.Length);
Console.WriteLine($"Fold {i + 1}/{n_folds}: train acc = {acc_train.Last()}%, test acc = {acc_test.Last()}% (n_train = {X_train.Length}, n_test = {X_test.Length})");
}
Console.WriteLine($"Avg train acc = {acc_train.Average()}%, avg test acc = {acc_test.Average()}%");
}