You have the code...you should know it. Straight away getting into the problem without studying any basics is not advisable. I told this to you multiple times.
Read about deepNetworkDesigner. This is a gui, you import the defined layer into the GUI, you will get your required information.
By the way what do you mean by input layer nodes and output layer nodes? It should be dimensions/ features of input and output. It is already said in the code the number of features are 12 and number of classes are 9. So what you have code for is a classification problem. You will give a input of size 1*12 and your arcitecture classifies it into any one of the nine classes it falls in.