mnist2.

ipynb - Colaboratory 27/06/23, 6:40 PM

from numpy import unique, argmax

from tensorflow.keras.datasets.mnist import load_data
from tensorflow.keras import Sequential
from tensorflow.keras.layers import Conv2D
from tensorflow.keras. layers import MaxPool2D
from tensorflow.keras.layers import Dense
from tensorflow.keras.layers import Flatten
from tensorflow.keras.layers import Dropout
from tensorflow.keras.utils import plot_model
from matplotlib import pyplot
import matplotlib.pyplot as plt
import numpy as np

# loading the MNIST dataset

(x_train, y_train), (x_test, y_test) = load_data ()
# reshaping the training and testing data
x_train = x_train.reshape ((x_train.shape[0], x_train.shape [1], x_train.shape[
x_test = x_test.reshape ((x_test.shape[0], x_test.shape[1], x_test.shape[2], 1)

x_train==x_test.astype('float32')/255.0
x_test x_train.astype('float32')/255.0
# #
changes
255 = in
pixel
32 bit
values
memory
are integer values
x_test = x_test.astype('float32')/255.0 # changes in 32 bit memory

fig =
ax.set_title
plt.figure(figsize=
(y_train (5,3))
[i])
for i in range (15):
ax = fig.add_subplot(2,10, i+1, xticks=[], yticks=[])
ax.imshow(np.squeeze(x_train[i]), cmap='gray')
ax.set_title (y_train [i])

5 4 1 9 2 1 3 1 4

5047 21314

3 5 3 6 1
35360

https://colab.research.google.com/drive/11c1QESByXTYVJLqxlDd1l5aSeY90ptfS#scrollTo=h0QAcvWmndrb Page 1 of 6
mnist2.ipynb - Colaboratory 27/06/23, 6:40 PM

img_shape = x_train.shape [1:]

# determine the shape of the input images
print (img_shape)
img_shape = x_train.shape [1:]
print (img_shape)

(28, 28, 1)

model = Sequential()
model.add(Conv2D(32, (3,3), activation='relu', input_shape=img_shape))
model.add(MaxPool2D((2, 2)))
model.add(Conv2D (48, (3,3), activation='relu'))
model.add(MaxPool2D ((2, 2)))
model.add(Dropout(0.5))
model.add(Flatten())
model.add(Dense(500, activation='relu'))
model.add(Dense(10, activation='softmax'))
model.summary()

Model: "sequential_5"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
conv2d_4 (Conv2D) (None, 26, 26, 32) 320

max_pooling2d_2 (MaxPooling (None, 13, 13, 32) 0

2D)

conv2d_5 (Conv2D) (None, 11, 11, 48) 13872

max_pooling2d_3 (MaxPooling (None, 5, 5, 48) 0

2D)

dropout_1 (Dropout) (None, 5, 5, 48) 0

flatten_1 (Flatten) (None, 1200) 0

dense_2 (Dense) (None, 500) 600500

dense_3 (Dense) (None, 10) 5010

=================================================================
Total params: 619,702
Trainable params: 619,702
Non-trainable params: 0
_________________________________________________________________

plot_model(model, 'model.jpg', show_shapes=True)

conv2d4input input: [(None,28,28,1)]

InputLaver output:[(None,28,28,1)]
https://colab.research.google.com/drive/11c1QESByXTYVJLqxlDd1l5aSeY90ptfS#scrollTo=h0QAcvWmndrb Page 2 of 6
mnist2.ipynb - Colaboratory 27/06/23, 6:40 PM

InputLaver output:[(None,28,28,1)]

conv2d4 input: (None,28,28,1)

Conv2D output:(None,26,26,32)

maxpooling2d2 input: (None,26,26,32)

MaxPooling2D output:(None,13,13,32)

conv2d5 input: (None,13,13,32)

Conv2D output:(None,11,11,48)

maxpooling2d3 input: (None,11,11,48)

MaxPooling2D output:(None,5,5,48)

dropout1 input: (None,5,5,48)

Dropout output:(None,5,5,48)

flatten1 input: (None,5,5,48)

Flatten output: (None,1200)

dense2 input: (None,1200)

Dense output: (None,500)

dense3 input: (None,500)

Dense output: (None,10)

https://colab.research.google.com/drive/11c1QESByXTYVJLqxlDd1l5aSeY90ptfS#scrollTo=h0QAcvWmndrb Page 3 of 6
mnist2.ipynb - Colaboratory 27/06/23, 6:40 PM

model.compile (optimizer='adam', loss='sparse_categorical_crossentropy', metrics=

x = model.fit(x_train, y_train, epochs=10, batch_size=128, verbose=2, validation_sp

Epoch 1/10
422/422 - 50s - loss: 0.2365 - accuracy: 0.9278 - val_loss: 0.0561 - val_accur
Epoch 2/10
422/422 - 43s - loss: 0.0813 - accuracy: 0.9744 - val_loss: 0.0398 - val_accur
Epoch 3/10
422/422 - 43s - loss: 0.0599 - accuracy: 0.9809 - val_loss: 0.0359 - val_accur
Epoch 4/10
422/422 - 44s - loss: 0.0472 - accuracy: 0.9844 - val_loss: 0.0305 - val_accur
Epoch 5/10
422/422 - 43s - loss: 0.0409 - accuracy: 0.9872 - val_loss: 0.0310 - val_accur
Epoch 6/10
422/422 - 44s - loss: 0.0356 - accuracy: 0.9881 - val_loss: 0.0309 - val_accur
Epoch 7/10
422/422 - 43s - loss: 0.0324 - accuracy: 0.9895 - val_loss: 0.0308 - val_accur
Epoch 8/10
422/422 - 44s - loss: 0.0278 - accuracy: 0.9909 - val_loss: 0.0261 - val_accur
Epoch 9/10
422/422 - 44s - loss: 0.0261 - accuracy: 0.9914 - val_loss: 0.0256 - val_accur
Epoch 10/10
422/422 - 43s - loss: 0.0239 - accuracy: 0.9920 - val_loss: 0.0211 - val_accur

loss, accuracy = model.evaluate(x_test, y_test, verbose=0)

print (f'Accuracy: {accuracy*100}')

Accuracy: 99.27999973297119

https://colab.research.google.com/drive/11c1QESByXTYVJLqxlDd1l5aSeY90ptfS#scrollTo=h0QAcvWmndrb Page 4 of 6
mnist2.ipynb - Colaboratory 27/06/23, 6:40 PM

image = x_train[5]
# lets display the image which we want to predict
plt.imshow(np.squeeze(image), cmap='gray')
plt.show()

OL

0
§ SZ

image= image.reshape (1, image.shape [0], image.shape [1], image.shape [2])

p = model.predict ([image])
print ('Predicted: {}'.format(argmax (p)))

1/1 [==============================] - 0s 109ms/step

Predicted: 2

https://colab.research.google.com/drive/11c1QESByXTYVJLqxlDd1l5aSeY90ptfS#scrollTo=h0QAcvWmndrb Page 5 of 6
mnist2.ipynb - Colaboratory 27/06/23, 6:40 PM

Colab paid products - Cancel contracts here

 0s completed at 6:06 PM

https://colab.research.google.com/drive/11c1QESByXTYVJLqxlDd1l5aSeY90ptfS#scrollTo=h0QAcvWmndrb Page 6 of 6