main_cup.py

import sys
from os import makedirs

from neuralnetwork.activationfunctions.linear import Linear

sys.path.insert(0, "../../")
sys.path.insert(0, "./")
from neuralnetwork.model.neuralnetwork import NeuralNetwork
from neuralnetwork.optimizer import MGD, LBFGS
from neuralnetwork.datasets.dataset_utils import load_monk_dataset, load_cup
from neuralnetwork.utils.utils import horizontal_plots_loss, print_results_to_table, plot_single_loss, \
    plot_convergence_rates, plot_single_convergence_rate, plot_losses
import numpy as np


def main():
    np.random.seed(seed=123)
    ds = "results/CUP" #<--- cambiamo qui così propaghiamo il dataset
    makedirs(ds, exist_ok=True)

    print("Loading datasets...")
    X_train, Y_train = load_cup()

    print("Building LBFGS model ... ")
    model = NeuralNetwork(LBFGS(m=25,
                                c1=1e-4,
                                c2=0.9,
                                threshold_gradient_norm=0.000001,
                                threshold_loss=1),
                          regulariz_lambda=0.0001,
                          loss_type="regularized_mse")
    model.add_layer(100, features = X_train.shape[1])
    model.add_layer(100)
    model.add_layer(2, activation=Linear())
    model.learn(X_train, Y_train, iterations=500)

    print("Building Momentum Descent model ... ")
    model2 = NeuralNetwork(MGD(learning_rate=0.06, momentum=0.6,
                               threshold_gradient_norm=0.000001,
                               threshold_loss=1),
                               regulariz_lambda=0.0001,
                               loss_type="regularized_mse")
    model2.add_layer(100, features = X_train.shape[1])
    model2.add_layer(100)
    model2.add_layer(2, activation=Linear())
    model2.learn(X_train, Y_train, iterations=500)

    # Print plots
    plot_losses(model.loss, model2.loss,ds)
    horizontal_plots_loss(model.loss, model2.loss, ds, algo1="LBFGS", algo2="Momentum Descent")
    plot_single_loss(model.loss, ds, "LBFGS")
    plot_single_loss(model2.loss, ds, "Momentum Descent")
    plot_convergence_rates(model.loss, model2.loss, ds)
    plot_single_convergence_rate(model.loss, ds, "LBFGS")
    plot_single_convergence_rate(model2.loss, ds,"Momentum Descent")
    print_results_to_table(model, model2, ds)
if __name__ == '__main__':
    main()