Deep Learning with PyTorch, Second Edition: Training and Applying Deep Learning and Generative Ai Models

Get a free eBook (PDF or ePub) from Manning as well as access to the online liveBook format (and its AI assistant that will answer your questions in any language) when you purchase the print book. PyTorch core developer Howard Huang updates the bestselling original Deep Learning with PyTorch with new insights into the transformers architecture and generative AI models. Instantly familiar to anyone who knows PyData tools like NumPy, PyTorch simplifies deep learning without sacrificing advanced features. In this book you’ll learn how to create your own neural network and deep learning systems and take full advantage of PyTorch’s built-in tools for automatic differentiation, hardware acceleration, distributed training, and more. You’ll discover how easy PyTorch makes it to build your entire DL pipeline, including using the PyTorch Tensor API, loading data in Python, monitoring training, and visualizing results. Each new technique you learn is put into action with practical code examples in each chapter, culminating into you building your own convolution neural networks, transformers, and even a real-world medical image classifier. In Deep Learning with PyTorch, Second Edition you’ll find: • Deep learning fundamentals reinforced with hands-on projects • Mastering PyTorch's flexible APIs for neural network development • Implementing CNNs, transformers, and diffusion models • Optimizing models for training and deployment • Generative AI models to create images and text About the technology The powerful PyTorch library makes deep learning simple—without sacrificing the features you need to create efficient neural networks, LLMs, and other ML models. Pythonic by design, it’s instantly familiar to users of NumPy, Scikit-learn, and other ML frameworks. This thoroughly-revised second edition covers the latest PyTorch innovations, including how to create and refine generative AI models. About the book Deep Learning with PyTorch, Second Edition shows you how to build neural network models using the latest version of PyTorch. Clear explanations and practical projects help you master the fundamentals and explore advanced architectures including transformers and LLMs. Along the way you’ll learn techniques for training using augmented data, improving model architecture, and fine tuning. What's inside • PyTorch APIs for neural network development • LLMs, transformers, and diffusion models • Model training and deployment About the reader For Python programmers with a background in machine learning. About the author Howard Huang is a software engineer and developer on the PyTorch library focusing on large scale, distributed training. Eli Stevens, Luca Antiga, and Thomas Viehmann authored the first edition of Deep Learning with PyTorch. Table of Contents Part 1 1 Introducing deep learning and the PyTorch library 2 Pretrained networks 3 It starts with a tensor 4 Real-world data representation using tensors 5 The mechanics of learning 6 Using a neural network to fit the data 7 Telling birds from airplanes: Learning from images 8 Using convolutions to generalize Part 2 9 How transformers work 10 Diffusion models for images 11 Using PyTorch to fight cancer 12 Combining data sources into a unified dataset 13 Training a classification model to detect suspected tumors 14 Improving training with metrics and augmentation 15 Using segmentation to find suspected nodules 16 Training models on multiple GPU 17 Deploying to production