import tensorflow as tf
 from tensorflow import keras
 from tensorflow.keras.models import Sequential
 from tensorflow.keras.layers import Dense
 import numpy as np
 import logging
 tf.get_logger().setLevel(logging.ERROR)

 EPOCHS = 500
 BATCH_SIZE = 16
 from tensorflow import keras
 # Load the dataset
 (raw_x_train, y_train), (raw_x_test, y_test) = keras.datasets.boston_housing.load_data()

x_mean = np.mean(raw_x_train, axis=0)
 x_stddev = np.std(raw_x_train, axis=0)
 x_train =(raw_x_train - x_mean) / x_stddev
 x_test =(raw_x_test - x_mean) / x_stddev
 # Create and train model.
 model = Sequential()
 model.add(Dense(64, activation='relu', input_shape=[13]))
 model.add(Dense(64, activation='relu')) # We are doing DL!
 model.add(Dense(1, activation='linear'))
 model.compile(loss='mean_squared_error', optimizer='adam',
 metrics =['mean_absolute_error'])
 model.summary()

  history = model.fit(x_train, y_train, validation_data=(x_test, y_test), epochs=EPOCHS, batch_size=BATCH_SIZE,verbose=2, shuffle=True)


 predictions = model.predict(x_test)
 for i in range(0, 4):
 print('Prediction: ', predictions[i],', true value: ', y_test[i])