# Load a pre-trained model model = VGG16(weights='imagenet', include_top=False, input_shape=(224, 224, 3))