Book Image

Python Deep Learning - Third Edition

By : Ivan Vasilev
4 (1)
Book Image

Python Deep Learning - Third Edition

4 (1)
By: Ivan Vasilev

Overview of this book

The field of deep learning has developed rapidly recently and today covers a broad range of applications. This makes it challenging to navigate and hard to understand without solid foundations. This book will guide you from the basics of neural networks to the state-of-the-art large language models in use today. The first part of the book introduces the main machine learning concepts and paradigms. It covers the mathematical foundations, the structure, and the training algorithms of neural networks and dives into the essence of deep learning. The second part of the book introduces convolutional networks for computer vision. We’ll learn how to solve image classification, object detection, instance segmentation, and image generation tasks. The third part focuses on the attention mechanism and transformers – the core network architecture of large language models. We’ll discuss new types of advanced tasks they can solve, such as chatbots and text-to-image generation. By the end of this book, you’ll have a thorough understanding of the inner workings of deep neural networks. You'll have the ability to develop new models and adapt existing ones to solve your tasks. You’ll also have sufficient understanding to continue your research and stay up to date with the latest advancements in the field.
Table of Contents (17 chapters)
1
Part 1:Introduction to Neural Networks
5
Part 2: Deep Neural Networks for Computer Vision
8
Part 3: Natural Language Processing and Transformers
13
Part 4: Developing and Deploying Deep Neural Networks

Training LLMs

Since most LLMs are decoder-only, the most common LLM pre-training task is NWP. The large number of model parameters (up to hundreds of billions) requires comparatively large training datasets to prevent overfitting and realize the full capabilities of the models. This requirement poses two significant challenges: ensuring training data quality and the ability to process large volumes of data. In the following sections, we’ll discuss various aspects of the LLM training pipeline, starting from the training datasets.

Training datasets

We can categorize the training data into two broad categories:

  • General: Examples include web pages, books, or conversational text. LLMs almost always train on general data because it’s widely available and diverse, improving the language modeling and generalization capabilities of LLMs.
  • Specialized: Code, scientific articles, textbooks, or multilingual data for providing LLMs with task-specific capabilities...