The Inference Versus Training Mode

In training mode, it's really not that different than serving mode if that makes sense. So basically you have this forward pass where you run through, you pass inputs into this neural network and then you get some output at the end. When you're trying to say use the model, like let's say you have a translation model, now you want to make predictions about what the translation should be, you just only do the forward part. Because you don't know what the answer should be.