VGG16 training with colab reutrns WARNING:root:kernel - numpy

Question

im still beginner with DL, here im trying to train VGG16 on a list of images to return 2048 features, but my issue was it returns 4094 features instead of 2048, so what i did to solve the issue is to create sequential model and remove one layer to be 2048 but will during the training on google cloab it returns kernel issue as below :
WARNING:root:kernel d3b89e05-3e27-4035-afcd-79a2c35a319a restarted
i have restarted the runtime and make a factorial restart, but still issue exists maybe because I did something wrong I'm my code, this issue happened only in the last code block
def read_images(folder_path, classlbl):
# load all images into a list
images = []
img_width, img_height = 224, 224
class1=[]
for img in os.listdir(folder_path):
img = os.path.join(folder_path, img)
img = load_img(img, target_size=(img_width, img_height))
class1.append(classlbl)# class one.
images.append(img)
return images, class1
#compute features for each image.
def computefeatures(model,image):
# convert the image pixels to a numpy array
image = img_to_array(image)
# reshape data for the model
image = image.reshape((1, image.shape[0], image.shape[1], image.shape[2]))
# prepare the image for the VGG model
image = preprocess_input(image)
# get extracted features
features = model.predict(image)
return features
model = Sequential()
# second convolutional block
model.add(Conv2D(128, (3,3), strides=(1,1), padding="same", activation="relu"))
model.add(Conv2D(128,(2,2), strides=(1,1), padding="same",activation="relu"))
model.add(MaxPooling2D((2, 2), strides=(2,2)))
# third convolutional block
model.add(Conv2D(256, (3,3), strides=(1,1), padding="same", activation="relu"))
model.add(Conv2D(256,(3,3), strides=(1,1), padding="same",activation="relu"))
model.add(Conv2D(256,(3,3), strides=(1,1), padding="same",activation="relu"))
model.add(MaxPooling2D((2, 2), strides=(2,2)))
# third convolutional block
model.add(Conv2D(512, (3,3), strides=(1,1), padding="same", activation="relu"))
model.add(Conv2D(512,(3,3), strides=(1,1), padding="same",activation="relu"))
model.add(Conv2D(512,(3,3), strides=(1,1), padding="same",activation="relu"))
model.add(MaxPooling2D((2, 2), strides=(2,2)))
# third convolutional block
model.add(Conv2D(512, (3,3), strides=(1,1), padding="same", activation="relu"))
model.add(Conv2D(512,(3,3), strides=(1,1), padding="same",activation="relu"))
model.add(Conv2D(512,(3,3), strides=(1,1), padding="same",activation="relu"))
#DNN Backend
model.add(Flatten())
model.add(Dense(4096, activation='relu'))
model.add(Dense(4096, activation='relu'))
#Output layer for classification (1000 classes)
model.add(Dense(1000, activation='softmax'))
#use the categorical cross entropy loss function for a multi-class classifier.
model.compile(loss='categorical_crossentropy',
optimizer='adam',
metrics=['accuracy'])
# call the image read and
folder_path = '/content/imageDir/101_ObjectCategories/windsor_chair'
classlbl=0
images, class1 =read_images(folder_path, classlbl)
# call the fucntion to compute the features for each image.
list_features1=[]
list_features1 = np.empty((0,4096), float)# create an empty array with 0 row and 4096 columns this number from fature
# extraction from vg16
for img in range(len(images)):
f2=computefeatures(model,images[img]) # compute features forea each image
list_features1 = np.append(list_features1, f2, axis=0)

Answer 1

One of the most popular reasons leads to restart kernel itself is out of memory (CPU RAM). It seems like because of the read_images function, you were trying to read and store too many image files at the same time.
You may want to try ImageDataGenerator API for reading images in small batches.
OR
You can simply change your read_images function as below:
def read_images(folder_path, classlbl):
# load all images into a list
images = []
img_width, img_height = 224, 224
class1=[]
for img in os.listdir(folder_path):
img = os.path.join(folder_path, img)
img = load_img(img, target_size=(img_width, img_height))
yield img, classlbl ## <--- this line
And use it like this
result =read_images(folder_path, classlbl)
# call the fucntion to compute the features for each image.
list_features1=[]
list_features1 = np.empty((0,4096), float)# create an empty array with 0 row and 4096 columns this number from fature
# extraction from vg16
for img, _ in result: # <--- change this line
f2=computefeatures(model, img) # compute features forea each image
list_features1 = np.append(list_features1, f2, axis=0)