autoencoder_training.py

import numpy as np  
import pandas as pd  
import os,sys,inspect
from support_functions import *

def read_and_train(read_func,parm=''):
    # Read the data
    if len(parm) == 0: # no param
        AnomalyData, data_train, data_test, labels_train, labels_test=read_func()
    else:
        AnomalyData, data_train, data_test, labels_train, labels_test=read_func(parm)

    # Merge the data
    data = np.concatenate((data_train, data_test))
    labels = np.concatenate((labels_train, labels_test))

    if AnomalyData.replicate_for_training >0:# If we only has very few data, we replicates the data before training
        data = np.tile(data, (AnomalyData.replicate_for_training,1))
        labels = np.tile(labels, AnomalyData.replicate_for_training)

    #Train the model
    autoencoder, encoder = train_autoencoder(AnomalyData, data, labels,save_model = True)

    # Delete the data to release the space
    del data
    del labels
    del data_train
    del data_test
    del labels_train
    del labels_test
    dir()

# Update models according to the list
read_and_train(read_mnist_data)
read_and_train(get_yale_faces_data)
read_and_train(read_synthetic_data,parm='Synthetic/')
read_and_train(read_synthetic_data,parm='Synthetic_2/')
read_and_train(read_synthetic_data,parm='Synthetic_3/')
read_and_train(read_synthetic_data,parm='Synthetic_4/')