{ "item_title" : "Deep Learning on Accelerated Computing", "item_author" : [" Ajit Singh "], "item_description" : "Deep Learning on Accelerated Computing is a definitive, heavily practical guide engineered to teach the end-to-end lifecycle of deep learning applications using modern hardware accelerators. Designed to meet the stringent requirements of international university syllabi projected for 2030, this book anticipates the future needs of the global technology industry. It strips away excessive theoretical bloat and focuses relentlessly on implementation, application development, and production-grade engineering. Philosophy The core philosophy of this book is Implementation Over Abstraction. In the modern tech industry, understanding the mathematical proof of backpropagation is far less valuable than knowing how to implement, optimize, and deploy a neural network onto a cluster of GPUs or a specialized NPU. This book operates on the belief that true mastery of deep learning comes from building. Therefore, the text is strictly oriented toward practical application. Key Features 1. Comprehensive Chapter 1: An exhaustive introductory chapter that establishes a rock-solid foundation, uniquely covering the definition, history, evolution, version comparisons, classifications, necessity, features, services, components, advantages, and applications of accelerated deep learning. 2. Hardware-First Approach: Unlike traditional books that treat hardware as an afterthought, this text integrates GPU/TPU/NPU system design, CUDA memory management, and hardware interfacing directly into the model building process. 3. Production-Oriented: Covers crucial, modern industry skills such as setting up Docker containers for AI, writing APIs for model inference, utilizing Triton Inference Server, and edge deployment. 4. Simple Examples: Uses the simplest, most relatable real-world examples to explain complex topics like tensor operations and memory bandwidth. 5. Complete DIY Capstone Project (Chapter 10): The book culminates in a full, live, working project. It provides the complete working code, step-by-step explanation, and deployment strategy, proving the practical viability of everything learned in chapters 1 through 9. 6. Future-Proof Content: Tailored to the 2030 academic syllabus, ensuring that the skills learned are highly relevant to future industry demands, including sustainable AI and highly distributed computing. Key Takeaways By the conclusion of this book, readers will have achieved the following: 1. A profound understanding of how hardware accelerators (GPUs, TPUs) function at a component level and how they drastically reduce training and inference times. 2. The ability to design, setup, and configure industry-standard accelerated computing environments from scratch using modern frameworks and containerization. 3. The skill to write simple, efficient code to build Deep Neural Networks, Convolutional Neural Networks, and Transformers tailored for hardware optimization. 4. Mastery over model deployment techniques, including quantization, pruning, and optimizing models for edge devices and cloud servers. 5. The practical experience of building a complete, live AI solution, complete with frontend interfacing, backend API deployment, and hardware-accelerated inference. Disclaimer: Earnest request from the Author. Kindly go through the table of contents and refer kindle edition for a glance on the related contents. Thank you for your kind consideration!", "item_img_path" : "https://covers4.booksamillion.com/covers/bam/9/79/816/503/9798165037771_b.jpg", "price_data" : { "retail_price" : "31.23", "online_price" : "31.23", "our_price" : "31.23", "club_price" : "31.23", "savings_pct" : "0", "savings_amt" : "0.00", "club_savings_pct" : "0", "club_savings_amt" : "0.00", "discount_pct" : "10", "store_price" : "" } }

Overview

Customers Also Bought

Details