Скачать или смотреть Deep Dive | C++ Machine Learning Tutorial Part 1 - Extract, Transform, Load (ETL)

This YouTube transcript details the first part of a C++ machine learning tutorial focused on building a data handling library for the MNIST handwritten digit dataset. The tutorial eschews external machine learning frameworks, instead implementing core functionalities from scratch to illustrate underlying principles. The initial focus is on creating a Data container class and a DataHandler class. The DataHandler class is designed to read the binary MNIST data, convert it from big-endian to little-endian format, split it into training, testing, and validation sets, and provide methods for accessing this data. The tutorial concludes with the successful compilation of the data handling library into a reusable shared library (libdata), setting the stage for future algorithm implementations within the tutorial series.