The last few you train horse versus humans, cats versus dogs, would bind the classification problems with just two output. For a lot of applications, you might want more than two output courses such as Rock Paper Scissors where you would have three possible courses. So in this week, you learn the details of how to implement this. And in fact, you've actually seen one version of this already in the first course where you did the ten course classification problem of fashion. But in this week, you'll learn the important coding difference of how to actually build these multi-class classifiers. >> Yes, exactly, and it's one of the things that's really exciting about the image data generator is how your images get automatically labeled based on the directories that they're in. So we'll take a look at having with rock, paper, scissors, for example, we'll have three sub-directories of rock, paper, and scissors with the images in them. And we'll learn how the image data generator then works in outside of binary mode and in multi-class mode. We'll also look at the coding changes that you would have within the DNN part of your convolutional neural network so that you can work more efficiently with multi-class. >> And one of the really cool things we'll see this week is the rock, paper, scissors data set that you generated using computer graphics. >> Yeah, so I wanted to explore that one of the bottlenecks I find in creating data sets is in getting access to data. And if I wanted to build a rock, paper, scissors, I wanted to have people of different races hands. So Asian hands, black hands, white hands. I wanted to have male and female hands. I wanted to have nail polish and pointy nails on some and I wanted to have big hands and little hands. And if I were to do that with real images, I'd have to hire a lot of models and I didn't have the budget for that. And so, I thought well, what if I do it with CGI? And if I generate CGI hands for all of these, and then to see. Because convolutions are really all about extracting features. What does a finger look like? What does a finger nail look like? That type of thing. And with photo real CGI, the convolutions that make for finger might look exactly the same that it would be of a photograph of a hand. So I wanted to explore that and do some training on that. And it works surprisingly well. >> Yeah, one of the exciting developments in deep learning is that computer graphics is advanced to the point where more and more groups are considering using computer graphics to generate images to feed to these incredibly data hungry deep learning algorithms. Several years ago, when computer graphics was less advanced, it was like a slightly fishy thing to do to use computer graphics to synthesize data. And even though today, using graphics, not a slam dunk, it does not always work, and there's still problems quite often. I thought it was really cool that you were able to generate a pretty decent data set, certainly good enough for a learning tool. And in other context, I see companies starting to explore these ideas for more and more settings as well. >> Yeah, I think as an industry, we're only getting started with it. And that's really, really good for us to experiment with these things. And hopefully, the students in this course being able to experiment using these CGI based data sets will be a nice introduction for them. >> Yeah, so more data to feed to this giant hungry maw of our data hungry deep learning algorithms. >> Exactly. >> So in this week, you get to play with this cool new data set and use it to practice implementing building multi-class classifiers in TensorFlow. Please dive in.