Mastering Transformers

Address NLP tasks as well as multi-modal tasks including both NLP and CV through the utilization of modern transformer architecture.

Key Features

• Understand the Complexity of Deep Learning Architectures and Transformers Architecture
• Learn how to create effective solutions to industrial NLP and CV problems
• Learn about the challenges in the preparation process, such as problem and language-specific data sets transformation

Book Description

The Transformer-based language models such as BERT, T5, GPT, DALL-E, ChatGPT have dominated natural language processing studies and become a new paradigm. Understand and be able to implement multimodal solutions including text-to-image (stable diffusions). Computer vision solutions that are based on Transformers are also explained in the book. Technical details and how to use them are parts of this book.

Thanks to their accurate and fast fine-tuning capabilities, Transformer-based language models outperformed traditional machine learning-based approaches for many challenging natural language understanding (NLU) problems.

Apart from NLP, recently a fast-growing area in multimodal learning and generative AI has been established which shows promising results. Dalle and Stable diffusions are examples of it. A developer can expect to learn more about them and how to tune them for specific purposes.

Developers working with The Transformers architecture will be able to put their knowledge to work with this practical guide to NLP. The book provides a hands-on approach to implementation and associated methodologies in the field of NLP that will have you up-and-running, and productive in no time. Also, developers that want to learn more about multimodal models and generative AI in the field of computer vision can use this book as a source.

What you will learn

• How NLP technologies have evolved over the past years
• How to solve simple/complex NLP problems with Python programming language
• How to solve classification/regression problems with traditional NLP approaches
• Training a language model and further exploring how to fine-tune the models to the downstream tasks
• How to use Transformers for generative AI and computer vision tasks
• How to build Transformers-based NLP applications with the Python Transformers library
• How to build language generation such as Machine Translation, Conversational AI in any language
• How to speed up transformer model inference to reduce latency

Who This Book Is For

The book is for deep learning researchers, hands-on practitioners, ML/NLP researchers, educators and their students who have a good command of programming subjects, have knowledge in the field of machine learning and artificial intelligence, and want to develop applications in the field of cutting-edge natural language processing as well as multimodal tasks. The readers will have to know at least python or any programming language, know machine learning literature, have some basic understanding of computer science, as this book is going to cover the practical aspects of natural language processing and multimodal deep learning.

Table of Contents

1. From bag-of-words to the Transformers
2. A hands-on Introduction to the Subject
3. Autoencoding Language Models
4. Autoregressive Language Models
5. Fine-tuning Language Model for Text Classification
6. Fine-tuning Language Model for Token Classification
7. Text Representation
8. Boosting your model performance
9. Parameter Efficient Fine-tuning
10. Zero-shot and Few-shot learning in NLP
11. Explainable AI (XAI) for NLP
12. Working with Efficient Transformers
13. Cross-Lingual Language Modeling
14. Serving Transformer Models
15. Model Tracking and Monitoring
16. Vision Transformers
17. Tabular Transformers
18. Multi-model Transformers
19. Graph Transformers