ada = Adadelta(lr=0.1, rho=0.95, epsilon=1e-08) model.compile(loss='categorical_crossentropy', optimizer=ada, metrics=['accuracy']) model.summary() return model img_rows, img_cols = 48, 48 batch_size = 128 nb_classes = 7 nb_epoch = 1200 img_channels = 1 Train_x, Train_y, Val_x, Val_y = dataprocessing.load_data() Train_x = numpy.asarray(Train_x) Train_x = Train_x.reshape(Train_x.shape[0], img_rows, img_cols) Val_x = numpy.asarray(Val_x) Val_x = Val_x.reshape(Val_x.shape[0], img_rows, img_cols) Train_x = Train_x.reshape(Train_x.shape[0], 1, img_rows, img_cols) Val_x = Val_x.reshape(Val_x.shape[0], 1, img_rows, img_cols) Train_x = Train_x.astype('float32') Val_x = Val_x.astype('float32') Train_y = np_utils.to_categorical(Train_y, nb_classes) Val_y = np_utils.to_categorical(Val_y, nb_classes)
import linear_regression as rg import dataprocessing as dp import numpy as np import matplotlib.pyplot as plt # Loading training examples X = dp.load_data('linearX.csv') y = dp.load_data('linearY.csv') X = X.reshape((-1, 1)) #Normalizing the training examples X, meu, sigma = dp.normalize(X) #Initialising parameters for gradient descent m, n = X.shape init_theta = np.zeros(n + 1) epsilon = 1e-10 eta = [0.001, 0.005, 0.009, 0.013, 0.017] color = ['red', 'blue', 'green', 'yellow', 'magenta'] plt.ion() for i in range(len(eta)): # Executing gradient descent theta, iterations, theta_history, cost_history = rg.linear_reg( X, y, init_theta, eta[i], epsilon) plt.plot(list(range(0, iterations + 1)), cost_history, color=color[i], label='eta = ' + str(eta[i]))
import logistic_regression as log_rg import dataprocessing as dp import numpy as np # Loading training examples X = dp.load_data('logisticX.csv') y = dp.load_data('logisticY.csv', dtype='int') #Normalizing the training examples X, meu, sigma = dp.normalize(X) m, n = X.shape init_theta = np.zeros(n + 1) epsilon = 1e-10 theta, iterations = log_rg.newton(X, y, init_theta, epsilon) # Plotting the data and hypothesis dp.plot_classification_data(X, y, ['Negative', 'Positive']) print('No. of iterations = ', iterations) print('Theta = ', theta) input('Press Enter to draw hypothesis') #Plotting decision boundary x_test = np.linspace(np.amin(X[:, 0]) - 1, np.amax(X[:, 0]) + 1, num=500) dp.plot_linear_decision_boundary(x_test, theta) input('Press Enter to close') dp.plot_close()
model.add(Activation('softmax')) ada = Adadelta(lr=0.1, rho=0.95, epsilon=1e-08) model.compile(loss='categorical_crossentropy', optimizer=ada, metrics=['accuracy']) model.summary() return model img_rows, img_cols = 48, 48 batch_size = 128 nb_classes = 7 nb_epoch = 1200 img_channels = 1 Train_x, Train_y, Val_x, Val_y = dataprocessing.load_data() Train_x = numpy.asarray(Train_x) Train_x = Train_x.reshape(Train_x.shape[0],img_rows,img_cols) Val_x = numpy.asarray(Val_x) Val_x = Val_x.reshape(Val_x.shape[0],img_rows,img_cols) Train_x = Train_x.reshape(Train_x.shape[0], 1, img_rows, img_cols) Val_x = Val_x.reshape(Val_x.shape[0], 1, img_rows, img_cols) Train_x = Train_x.astype('float32') Val_x = Val_x.astype('float32') Train_y = np_utils.to_categorical(Train_y, nb_classes)