Ques1-Implement CNN for MNIST image classification.

Upload screenshot of your

implementation: -code -output screen

Code-:

import keras
mnist = keras.datasets.mnist

(x_train, y_train), (x_test, y_test) = mnist.load_data()
import matplotlib.pyplot as plt

plt.imshow(x_train[0]) #training image
plt.show()
print(y_train[0]) #what the number is supposed to be

#Reshape the data
x_train = x_train.reshape(x_train.shape[0], 28, 28, 1)
x_test = x_test.reshape(x_test.shape[0], 28, 28, 1)
input_shape = (28, 28, 1)

#Normalize the pixel values from a scale out of 255 to a scale out of 1
x_train = x_train.astype('float32')
x_test = x_test.astype('float32')
x_train /= 255
x_test /= 255
y_train = keras.utils.to_categorical(y_train, 10)
y_test = keras.utils.to_categorical(y_test, 10)

model = keras.models.Sequential()
model.add(keras.layers.Conv2D(32, kernel_size=(3, 3), activation='relu'
,
                              input_shape=input_shape))

model.add(keras.layers.Conv2D(64, (3, 3), activation='relu'))
model.add(keras.layers.MaxPooling2D(pool_size=(2, 2)))
model.add(keras.layers.Dropout(0.25))
model.add(keras.layers.Flatten())
model.add(keras.layers.Dense(128, activation='relu'))
model.add(keras.layers.Dropout(0.5))
model.add(keras.layers.Dense(10, activation='softmax'))

model.compile(“adam”,
              optimizer=keras.optimizers.Adadelta(),

NAUSHEEN FATIMA 1
              metrics=['accuracy'])

model.fit(x_train, y_train,
          batch_size=128,
          epochs=20,
          validation_data=(x_test, y_test))
Output:-

NAUSHEEN FATIMA 2