by Anastasia Yanina | Released January 2019 | ISBN: 9781789610314

Train your networks faster with PyTorchAbout This VideoBuild computational graphs on-the-fly using strong PyTorch skills and develop a solid foundation in neural network structures.The course is embedded with easy-to-follow instructions that will help you build your first dynamic graph.You will apply dynamic neural networks to solve various real-world problems using dynamic memory and dynamic computations.In DetailDeep learning influences key aspects of core sectors such as IT, finance, science, and many more. Problems arise when it comes to getting computational resources for your network. You need to have a powerful GPU and plenty of time to train a network for solving a real-world task. Dynamic neural networks help save training time on your networks. They also reduce the amount of computational resources required. In this course, you'll learn to combine various techniques into a common framework. Then you will use dynamic graph computations to reduce the time spent training a network. By the end, you'll be ready to use the power of PyTorch to easily train neural networks of varying complexities.All the related code files are placed on GitHub repository at https://github.com/PacktPublishing/-Dynamic-Neural-Network-Programming-with-PyTorchDownloading the example code for this course: You can download the example code files for all Packt video courses you have purchased from your account at http://www.PacktPub.com. If you purchased this course elsewhere, you can visit http://www.PacktPub.com/support and register to have the files e-mailed directly to you. Show and hide more

1. Chapter 1 : Getting Started with PyTorch
   • The Course Overview 00:03:10
   • Installation Checklist 00:03:33
   • Tensors, Autograd, and Backprop 00:03:47
   • Backprop, Loss Functions, and Neural Networks 00:06:24
   • PyTorch on GPU: First Steps 00:03:13
2. Chapter 2 : Imperative Side of PyTorch
   • Imperative Programming Architectures 00:02:46
   • Static Graphs versus Dynamic Graphs 00:04:42
   • Neural Network Debugging: Why Imperative Philosophy Helps 00:02:02
3. Chapter 3 : Dynamic Computational Graphs: Intuition and Examples
   • Feedforward and Recurrent Neural Networks 00:13:07
   • Convolutional Neural Networks 00:19:36
   • Autoencoders 00:11:47
4. Chapter 4 : Creating Extensions with PyTorch
   • Extensions with Numpy – Part 1 00:05:11
   • Extensions with Numpy – Part 2 00:05:19
   • Custom C++ and CUDA Extensions: Motivation 00:04:17
   • Custom C++ and CUDA Extensions: Setuptools 00:04:31
   • Custom C++ and CUDA Extensions: Binding to Python 00:03:21
   • Custom C++ and CUDA Extensions: JIT Compilation 00:03:22
5. Chapter 5 : Image Captioning: Why Dynamic Graph Is a Good Choice?
   • Image Captioning: First Steps 00:02:18
   • PyTorch DataLoaders 00:09:06
   • Image Captioning: Theory 00:09:48
   • Image Captioning: Practice 00:11:12
   • Honor Track: Image Captioning Datasets 00:02:57
6. Chapter 6 : Natural Language Processing: Intuition for Dynamic Programming
   • Motivation and Section Overview 00:01:52
   • Word Embeddings 00:12:49
   • Sentiment Analysis with PyTorch 00:15:49
   • Char-Level RNN for Text Generation 00:20:34
Show and hide more