Build a Reasoning Model, From Scratch, Version 7, MEAP, Raschka S., 2026

Build a Reasoning Model, From Scratch, Version 7, MEAP, Raschka S., 2026.
 
   Reasoning is one of the most exciting and important recent advances in improving LLMs, but it’s also one of the easiest to misunderstand if you only hear the term reasoning and read about it in theory. That’s why this book takes a hands-on approach. We’ll start with a pre-trained base LLM and then add reasoning capabilities ourselves, step by step in code, so you can see exactly how it works.
This book isn’t a “production deployment” manual, and we won’t use any third-party LLM libraries. Instead, think of it as a behind-the-scenes tour where you get to develop the machinery yourself.
By the end, you will not only understand what reasoning is and how it works, but you will also have built it from scratch. That’s a perspective that will serve you well whether you are using, developing, or planning to deploy LLMs in the future.




A roadmap to reasoning models from scratch.
Now that we have discussed reasoning in LLMs from a bird's-eye view, the subsequent chapters will guide you through the process of coding and applying reasoning methods from scratch. We will tackle this in multiple stages, as outlined in figure 1.9.

As shown in figure 1.9, we cover the reasoning model development in several stages. In stage 1 (next chapter), we load a conventional LLM that has already undergone the basic pre-training and instruction fine-tuning stages. Then, in stage 2, we cover common methods for evaluating LLMs and reasoning capabilities, so that we can measure the improvements we make when we apply reasoning-enhancing methods in stages 3 and 4.

Stage 3 covers inference techniques that can improve the response quality and reasoning behavior of LLMs. Note that these techniques can be applied to improve any LLM, conventional LLMs and LLMs that have been trained as reasoning models. Stage 4 will introduce training methods to develop reasoning models.

Contents.
1 Understanding reasoning models.
2 Generating text with a pre-trained LLM.
3 Evaluating reasoning models.
4 Improving reasoning with inference-time scaling.
5 Inference-time scaling via self-refinement.
6 Training reasoning models with reinforcement learning.
7 Improving GRPO for reinforcement learning.
8 Distilling reasoning models for efficient reasoning.
Appendix A. References and further reading.
Appendix B. Exercise solutions.
Appendix C. Qwen3 LLM source code.
Appendix D. Using larger LLMs.
Appendix E. Batched inference.
Appendix F. Common approaches to model evaluation.



Бесплатно скачать электронную книгу в удобном формате, смотреть и читать:
Скачать книгу Build a Reasoning Model, From Scratch, Version 7, MEAP, Raschka S., 2026 - fileskachat.com, быстрое и бесплатное скачивание.

Скачать файл № 1 - pdf
Скачать файл № 2 - epub
Ниже можно купить эту книгу, если она есть в продаже, и похожие книги по лучшей цене со скидкой с доставкой по всей России.Купить книги


Скачать - epub - Яндекс.Диск.

Скачать - pdf - Яндекс.Диск.




Теги: учебник по программированию :: программирование :: Raschka