A person looks at the map and the map provides them with information that tells them what will happen if they move in a certain direction or whatever. A map can tell someone standing on a road whether if they take the second left hand turn they will come across a church or an open field or a roundabout or another street. Its giving them information about something they cannot immediately access and don't know about. That is a form of prediction, made by the person using the information from the map which is a model of the topographic features of some landscape. If I have never been somewhere before and have no knowledge of its terrain, then I can think of the map as allowing me to make a prediction of the kind of terrain I might expect to see. Its my personal prediction. Maybe you will see it more easily if I use words like knowledge or expectation instead of prediction, but I would be meaning the exact same thing. I don't necessarily mean predicting something no one has ever seen before. This is about the personal knowledge of whoever is using the map. They get knowledge from the map and they use that knowledge to act. That implies prediction. I am not going to embark on a route unless I know whats at the end of it which means I can predict what will happen if I were to go down that route, which is essentially just equivalent to making factual statements about this route and its endpoint which I cannot access immediately from my current position. When I say prediction, I basically just mean the utilization of knowledge, knowing what will happen or what is the case beyond my immediate experience. A map trivially allows this to occur. Even the photo example too: if you have never seen someone before and you have seen their photo, then you suddenly have information about them which you can use in novel contexts, you might be able to recognize them walking down the street.

>map itself cannot predict because it is an inanimate objects

Well so are models. no model is useful unless someone is there to initialize it and put in the parameters, the variables, the initial conditions that need to be used to predict something.