import pandas as pd import matplotlib.pyplot as plt from mpl_toolkits.mplot3d import Axes3D import argparse if __name__ == '__main__': parser = argparse.ArgumentParser() group = parser.add_mutually_exclusive_group(required=True) group.add_argument('-w', '--weights', help='Input file to load h5 model trained weights.') group.add_argument('-f', '--file', help='Input file to save trained model weights.') parser.add_argument('-p', '--plot', action='store_true', help='Plot the latent space in a 2D scatter (if the latent space dimesion is greater than 2, PCA will be applied).') args = parser.parse_args() x, y_true = kdd.get_dataset() x = MinMaxScaler().fit_transform(x) x_train, x_test = train_test_split(x, test_size=0.2) original_dim = x_train.shape[1] batch_size = 192 epochs = 100 vae = Vae([original_dim, 96, 64, 32, 16]) if args.weights: vae.model.load_weights(args.weights) else: vae_history = vae.model.fit(
import numpy as np from utils import result_info if __name__ == '__main__': parser = argparse.ArgumentParser() group = parser.add_mutually_exclusive_group(required=True) group.add_argument('-l', '--load', help='File to load SVC trained model.') group.add_argument('-s', '--save', help='File to save SVC trained model.') parser.add_argument('-e', '--encode', help='Encode the training data with a Variational Autoencoder.', action='store_true') args = parser.parse_args() x, y = kdd.get_dataset() y = y.cat.add_categories(['anormal']) y[y != 'normal'] = 'anormal' y = y.cat.remove_unused_categories() x = MinMaxScaler().fit_transform(x) if args.encode: vae = Vae([x.shape[1], 96, 64, 32, 16]) vae.model.load_weights('models/vae_full.h5') x, _, _ = vae.encoder.predict(x) x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.4)
'--weights', help='Input file to load h5 model trained weights.') group.add_argument('-f', '--file', help='Input file to save trained model weights.') parser.add_argument( '-p', '--plot', action='store_true', help= 'Plot the latent space in a 2D scatter (if the latent space dimesion is greater than 2, PCA will be applied).' ) args = parser.parse_args() x, _ = kdd.get_dataset(mode='normal') x = MinMaxScaler().fit_transform(x) x_train, x_test = train_test_split(x, test_size=0.2) original_dim = x_train.shape[1] batch_size = 48 epochs = 100 vae = Vae([original_dim, 64, 32, 16, 8]) if args.file: vae_history = vae.model.fit(x_train, x_train,