Final.py

# -*- coding: utf-8 -*-
"""CNN.ipynb

Automatically generated by Colaboratory.

Original file is located at
    https://colab.research.google.com/drive/1z-MzR5uN73-ek3jLjZoHPSUS5X1lgPsI
"""

from tensorflow.keras.datasets import mnist
from tensorflow.keras.models  import Sequential
from tensorflow.keras.layers import Conv2D, Activation, Dropout, Dense, MaxPooling2D, Flatten
from tensorflow.keras.optimizers import SGD, Adam, Adadelta
import numpy as np
import tensorflow as tf


# Loading Dataset
(x_train, y_train), (x_test, y_test) = mnist.load_data()
x_train, x_test = x_train/255.0, x_test/255.0
x_train, x_test = np.expand_dims(x_train, axis=-1), np.expand_dims(x_test, axis=-1)

# CNN Layers
model = Sequential()
model.add(Conv2D(16, (3, 3),(2,2), input_shape=x_train.shape[1:], padding='same'))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2), padding='same'))

model.add(Conv2D(16, (2, 2),(2,2),padding='same'))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2), padding='same'))

model.add(Conv2D(16, (2, 2),(2,2),padding='same'))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2),padding='same'))

model.add(Flatten())

#FC Layers
model.add(Dense(16,activation='relu'))
model.add(Dense(10,activation='softmax'))

model.summary()
print("\nTraining...")
opt = SGD(learning_rate=0.05)
model.compile(optimizer=opt, loss="sparse_categorical_crossentropy", metrics=["accuracy"])
model.fit(x_train, y_train, epochs=20,batch_size=96)

# Testing
print("\nTesting...")
test_loss, test_acc = model.evaluate(x_test, y_test)
print("Test Loss: {0} - Test Acc: {1}".format(test_loss, test_acc))