Welcome to PyTorch: Techniques and Ecosystem Tools, the second course in the PyTorch Professional Certificate. I'm joined once again by Laurence Moroney. In this course, you focus on optimizing model performance and building more advanced architectures. So you see how to take basic building blocks and quickly combine them into much more complex, much more advanced neural network architectures. In the first part of the course, we're actually going to tackle one of the biggest challenges in model optimization with hyperparameter tuning. There's so many ways that you can do it, including writing your own code for rough searches or using tools like Optuna from the Python ecosystem. Once I've designed an architecture that works, it's rarely optimal. There's all these choices that I have to make. Do I build deeper? Do I build wider? What kind of optimizer will I use? We'll actually start with giving you the tools for hyperparameter searching and neural architecture search. And then we'll teach you how to build these models from scratch, computer vision models and natural language models, before overlaying these with transfer learning from existing models and then doing fine tuning on them. You will learn a lot of those key building blocks, things like tokenization, pre-training, fine tuning that have been key to really everything happening in AI around us and that maybe even more importantly, will give you key skills to be able to navigate all of these very large and important portfolios of technologies. And then to bring it full circle, we'll go back to optimization to take a look at, are we efficiently loading data from the drive to the CPU, to the GPU? Are we having things hanging around waiting when they could be working? And like to just look at the various tools and techniques that we have to be able to squeeze the most optimization that we have out of that, because the faster we can train accurately, the quicker we can actually iterate on this. I think you'll enjoy this course, where you go from the basics to understanding a lot more sophisticated concepts in how to build deep learning models. Let's go on to the next video, where we'll start with hyperparameter search.