vaibhavcodes

model = keras.Sequential()

# Adding pre-trained model layers
model.add(model_vgg)

# Flattening the image pixels
model.add(Flatten())

# Adding 2 dense layers with 4096 neurons
model.add(Dense(4096, activation='relu'))

vaibhavcodes

model_vgg.trainable = False

vaibhavcodes

# The entire network is present under VGG16 feature extractor
from keras.applications.vgg16 import VGG16
model_vgg = VGG16(weights='imagenet', include_top=False, input_shape= (224,224,3))

# include_top: Whether to include the 3 fully-connected layers at the top of the network.

# weights: 'imagenet' (pre-training on ImageNet), or the path to the weights file to be loaded.

# input_shape: optional shape tuple, only to be specified if include_top is False 
#             (otherwise the input shape has to be (224, 224, 3) (with channels_last data format) 

vaibhavcodes

# Predicting the loaded image
output = model.predict(test_img)

# Condition of checking the label of the loaded or tested image
if output[0][0] > output[0][1]:
  print("Cat")
else:
  print("Dog")

vaibhavcodes

# Converting image to array
test_img = np.asarray(test_img)

# Expanding the dimensions
test_img = np.expand_dims(test_img, axis=0)

vaibhavcodes

# Load the testing image
test_img = image.load_img(r"/content/cat.jpg", target_size=(224,224))
test_img

vaibhavcodes

# Fitting the model
result = model.fit_generator(steps_per_epoch=100, generator=training_data, verbose=1, validation_data=testing_data, epochs=50 )

vaibhavcodes

# Compiling the model
model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])

vaibhavcodes

# Getting the summary of the model
model.summary()

vaibhavcodes

# Initializing a Sequential model
model = Sequential()

# Creating first block- (2 Convolution + 1 Max pool)
model.add(Conv2D(filters= 64, kernel_size= (3,3), strides= (1,1), padding='same', input_shape= (224, 224, 3), activation= 'relu'))
model.add(Conv2D(filters= 64, kernel_size= (3,3), strides= (1,1), padding='same', activation= 'relu'))
model.add(MaxPool2D(pool_size= (2,2), strides=(2,2)))

# Creating second block- (2 Convolution + 1 Max pool)
model.add(Conv2D(filters= 128, kernel_size= (3,3), strides= (1,1), padding='same', activation= 'relu'))