Okay, so you've built your predictive model and it seems to be running pretty well. Maybe it's reasonably fast and it predicts the right word most of the time. So, you're done right? No. Put it this way, if you were playing Mahler's third symphony, you'd be like in the fourth movement right now. So you'd still have like two more movements to go, okay? So the goal of this task is to make things better, okay? So remember in the last task, you were supposed to come up with a list of ideas that you thought might improve the performance of your model in various ways. So, what, what, what we want to do here is go down this list of ideas, try different things out, try to think outside the box and see what works and what doesn't, okay? As you're going down this list of ideas you're going to have to evaluate the performance of your model as you go. So remember to hold out a test data set so that you can evaluate these new creative models that you're building, okay? So, try to have fun doing this task, because you know, if you're not having fun you're not being a data scientist.