Build a machine learning library from scratch using only C++ | Part 2

Mastering Gradients & Backpropagation: Building a Neural Network Library in C++ | Part 2 In this tutorial, we delve into the essential components of creating a C++ machine learning library, focusing on gradients and backpropagation. Learn how to construct fundamental classes that handle memory allocation, data storage, and operation management. Follow along as we cover both forward and backward propagation, including detailed explanations on gradient calculation and topological sorting. We also demonstrate advanced C++ techniques like using shared pointers and custom hash functions. Join us for an in-depth guide that prepares you to build efficient and robust neural network algorithms. 00:00 Introduction to Building a Machine Learning Library 00:23 Setting Up the Class Skeleton 01:42 Understanding Shared Pointers 02:10 Defining Data Members 03:26 Creating and Managing Values 07:23 Implementing Arithmetic Operations 12:42 Forward Propagation and Testing 14:56 Storing and Managing Children Nodes 15:56 Forward Propagation Explained 16:59 Understanding Backward Propagation 18:10 Gradient Calculation Techniques 21:32 Topological Sorting for Graphs 22:59 Implementing the Backpropagation Function 23:47 Building the Topological Graph 25:34 Hash Functions and Gradient Calculation 27:42 Compiling and Testing the Code 30:29 Final Thoughts and Next Steps Github: https://github.com/ggsharma/code-with-me/blob/main/ML%20from%20basics/microgradpp.cpp