It would be more interesting to discuss the model itself, rather than its predictions. It's pretty evident that the output is quite useless, so why keep posting it?