import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt
# Load training dataset of 60000 images with greyscale values in 28 x 28
# and labels 
(x_train, y_train), (x_test, y_test) = tf.keras.datasets.mnist.load_data()

# print the shape of the numpy arrays
print ('Print shape of pixel data')
print(x_train.shape)
print ('Print shape of labels')
print(y_train.shape)

# normalize pixel to 0-1
x_train = x_train / 255
x_test = x_test / 255

# choose an image num  to display and print
num = 20

image = x_train[num]
label = y_train[num]

print ('Print normailzed pixel data of image ',num, ' :')
print(x_train[num])
print ('Print label of image ',num ,  ' :' )
print(y_train[num])

# plot the image using imshow
plt.imshow(image, cmap='gray')

# set the title
plt.title("Label: %d" % label )

# remove the axis labels and ticks
plt.axis('off')

# show the plot
plt.show()

# Plot 16 examples from the numpy array which was read in above
# and display it

fig, axes = plt.subplots(4, 4, figsize=(10, 10))
for i , ax in enumerate(axes.ravel()):
    ax.imshow(x_train[num+i], cmap='gray')
    ax.set_title("Label: %d" % y_train[num+i])
    ax.axis('off')
plt.suptitle("Examples of training set images")
plt.show()