Tensorflow autoencoder loss not converging

Question

I was going through keras blog and found one simple autoencoderes. It is written using keras and it is working as expected.

I have done some changes in the code to use tensorflow 2 keras functional API. Now problem is code is not throwing any error but it is not working as expected (val loss is more than 0.6).

I am not able to find any mistake in the code. Here is the modified code:

from tensorflow.keras.layers import Dense, Input
from tensorflow import keras
from tensorflow.keras.datasets import mnist
import numpy as np

encoding_dim = 32


input_img = Input(shape=(784,))
encoded = Dense(encoding_dim, activation='relu')(input_img)
decoded = Dense(784, activation='sigmoid')(encoded)

autoencoder = keras.Model(input_img, decoded)

encoder = keras.Model(input_img, encoded)

encoded_input = Input(shape=(encoding_dim,))
decoder_layer = autoencoder.layers[-1]
decoder = keras.Model(encoded_input, decoder_layer(encoded_input))

autoencoder.compile(optimizer='adadelta', loss='binary_crossentropy')


(x_train, _), (x_test, _) = mnist.load_data()

x_train = x_train.astype('float32') / 255.
x_test = x_test.astype('float32') / 255.
x_train = x_train.reshape((len(x_train), np.prod(x_train.shape[1:])))
x_test = x_test.reshape((len(x_test), np.prod(x_test.shape[1:])))
print(x_train.shape)
print(x_test.shape)

autoencoder.fit(x_train, x_train,
               epochs=50,
               batch_size=256,
               shuffle=True,
               validation_data=(x_test, x_test))

encoded_imgs = encoder.predict(x_test)
decoded_imgs = decoder.predict(encoded_imgs)


# use Matplotlib (don't ask)
import matplotlib.pyplot as plt

n = 10  # how many digits we will display
plt.figure(figsize=(20, 4))
for i in range(n):
    # display original
    ax = plt.subplot(2, n, i + 1)
    plt.imshow(x_test[i].reshape(28, 28))
    plt.gray()
    ax.get_xaxis().set_visible(False)
    ax.get_yaxis().set_visible(False)

    # display reconstruction
    ax = plt.subplot(2, n, i + 1 + n)
    plt.imshow(decoded_imgs[i].reshape(28, 28))
    plt.gray()
    ax.get_xaxis().set_visible(False)
    ax.get_yaxis().set_visible(False)
plt.show()

Answer 1

If you change the optimizer to adam the loss function converges. Also check this question:

• Ada-Delta method doesn't converge when used in Denoising AutoEncoder with MSE loss & ReLU activation?