model

model = Sequential([
    #first convolution
    Conv2D(32, (3,3), padding='same', activation='relu',kernel_initializer='he_normal', input_shape=(28,28, 1)),
    MaxPooling2D(2,2),
    
    #second convolution
    Conv2D(64, (3,3),padding='same', activation='relu'),
    MaxPooling2D(2,2),
    Dropout(0.2),
    
    #third convolution
    Conv2D(128, (3,3),padding='same', activation='relu'),
    MaxPooling2D(2,2),
    
    #fourth convolution
    Conv2D(128, (3,3),padding='same', activation='relu'),
    MaxPooling2D(2,2),
    Dropout(0.2),
    
    Flatten(),
    BatchNormalization(),
    Dense(512,activation='relu'),
    Dense(10,activation='softmax')

])