Mastering Computer Vision with TensorFlow 2.x

Developing the model

Once the images have been prepared, we can start building the model. The Keras sequential model allows us to do just that. It is a list of model layers stacked on top of each other. Next, we build a sequential model by stacking convolution, activation, max pooling, dropout, and padding, as shown in the following code:

model = Sequential([Conv2D(96, 11, padding='valid', activation='relu',input_shape=(img_height, img_width,3)), MaxPooling2D(),Dropout(0.2),Conv2D(256, 5, padding='same', activation='relu'),MaxPooling2D(),Conv2D(384, 3, padding='same', activation='relu'),Conv2D(384, 3, padding='same', activation='relu'), Conv2D(256, 3, padding='same', activation='relu'),MaxPooling2D(), Dropout(0.2),Conv2D(1024, 3, padding='same', ...