How do you test a GIZA++ model?

I'm using GIZA++ to align Korean and English sentences. I know how to train a model with GIZA++ with my parallel corpus, but I can't find a way to test my model and use it for future predictions. I see the -tc option that is supposed to be the test corpus, but not sure how to use the option to test the model (what format the test corpus should be in...) or make predictions. So any NLP gurus out there please help!!