A simple Deep Autoencoder using keras

deep_autoencoder.py

import tensorflow as tf
from tensorflow import keras
from matplotlib import pyplot as plt

from keras.datasets import mnist
from keras.layers import Input, Dense
from keras.models import Model

(x_train, y_train), (x_test, y_test) = mnist.load_data()
x_train = x_train.astype('float32')/255.0
x_test = x_test.astype('float32')/255.0

x_train = x_train.reshape(len(x_train), (x_train.shape[1]*x_train.shape[2]))
x_test = x_test.reshape(len(x_test), (x_test.shape[1]*x_test.shape[2]))

encoding_dim = 32

input_img = Input(shape = (784, ))
encoded = Dense(encodin_dim, activation='relu')(input_img)
decoded = Dense(784, activation='sigmoid')(encoded)
autoencoder = Model(input_img, decoded)

encoder = Model(input_img, encoded)
enncoded_input = Input(shape=(encoding_dim, ))
decode_layer = autoencoder.layers[-1]
decoder = Model(encoded_input, decode_layer(encoded_input))

autoencoder.compile(optimizer='adadelta', loss='binary_crossentropy', metrics=['accuracy'])

autoencoder.fit(x_train, x_train, epochs=20, validation_data=(x_test, x_test))

encoded_images = encoder.predict(x_test)
encoded_images.shape
predicted = decoder.predict(encoded_images)
plt.imshow(predicted[0].reshape(28, 28))

def plot_imgs(index=0):
  f = plt.figure()
  f.add_subplot(1,3, 1)
  plt.imshow(x_test[index].reshape(28, 28))
  f.add_subplot(1,3, 2)
  plt.imshow(encoded_images[index].reshape(-1, 1))
  f.add_subplot(1,3, 3)
  plt.imshow(predicted[index].reshape(28, 28))
  plt.show(block=True)
  
plot_imgs(2)