Welcome back. In this week, you'll learn how to take a model that's been trained in TensorFlow in Python and run that in a web browser, in TensorFlow.js. Even though you've learned how to train models in the web browser, today, a lot of models are still trained in a data center using TensorFlow. Yes. So how do you get that to run in a web browser? There's actually a suite of tools in TensorFlow called TensorFlow.js in Python that you can use to then convert your TensorFlow saved models into JSON notation. If you look on GitHub, there's a library of models that have been pre-converted. This week, we're going to look at two of those models. One of them is going to be toxicity, that you can pass it some texts then it will classify that for various items of toxicity. One of them is MobileNet for image classification. I know you're familiar with MobileNet. Yes. MobileNet is actually one of the coolest recent neural network architectures that was designed to run on mobile devices and on web browsers and low compute resource environments. So is one the architectures that I have used a few times, and it does help to train it in a beefy data center. So one of the cool things about the TensorFlow, TensorFlow.js conversion is you can take them all like MobileNet, train on the Big Data Center and convert it and then run it on a web browser. Yeah, exactly. So we'll show using that model and uploading some images to it and all of the inferences then done locally, which is really cool for privacy and saving round tripping and all that. Then, of course, the natural question students will have and learners will have is, how do I do this for myself? We'll take a look at that. We'll take a look at building a model in Python, saving it out as a saved model, and then converting it into JSON so that you can do inference in the browser. In this week, you'll learn how to take a model that's been trained in TensorFlow and run that in a web browser using TensorFlow.js. This includes both a model that you could train yourself and convert, or taking a model that someone else has trained and just downloading and using that. Please go onto the next video.