Deep Learning in Modern C++ Tarnowskie Góry

Description Deep learning is revolutionizing how we approach complex problems, and harnessing its power directly within C++ provides unparalleled control and efficiency. This book bridges the gap between cutting-edge deep learning techniques and the robust, high-performance capabilities of modern C++, empowering developers to build sophisticated AI applications from the ground up. This book guides you through the entire development lifecycle, starting with a solid foundation in the modern features and essential libraries, like Eigen, for C++. You will master core deep learning concepts by implementing convolutions, fully connected layers, and activation functions, while learning to optimize models using gradient descent, backpropagation, and advanced optimizers like SGD, Momentum, RMSProp, and Adam. Crucial topics like cross-validation, regularization, and performance evaluation are covered, ensuring robust and reliable applications. Finally, you will dive into computer vision, building image classifiers and object localization systems, leveraging transfer learning for optimal performance. By the end of this book, you will be proficient in developing and deploying deep learning models within C++, equipped with the tools and knowledge to tackle real-world AI challenges with confidence and precision. What you will learn Implement core deep learning models in modern C++. Code CNNs, RNNs, GANs, and optimization techniques. Build and test robust deep learning C++ applications. Apply transfer learning in C++ computer vision tasks. Master backpropagation and gradient descent in C++. Develop image classifiers and object detectors in C++. Who this book is for This book is tailored for C++ developers, data scientists, and machine learning engineers seeking to implement deep learning models using modern C++. A foundational understanding of C++ programming and basic linear algebra is recommended. Table of Contents 1. Introduction to Deep Learning Programming 2. Coding Deep Learning with Modern C++ 3. Testing Deep Learning Code 4. Implementing Convolutions 5. Coding the Fully Connected Layer 6. Learning by Minimizing Cost Functions 7. Defining Activation Functions 8. Using Pooling Layers 9. Coding the Gradient Descent Algorithm 10. Coding the Backpropagation Algorithm 11. Underfitting, Overfitting, and Regularization 12. Implementing Cross-validation, Mini Batching, and Model Performance Metrics 13. Implementing Optimizers 14. Introducing Computer Vision Models 15. Developing an Image Classifier 16. Leveraging Training Performance with Transfer Learning 17. Developing an Object Localization System