Hello, i am new to pytorch and wanted to know how exactly the trained model could be used for prediction with images? thank you