Transformers for Natural Language Processing

Denis Rothman, Antonio Gulli

• Publisher: Packt Publishing
• Year: 2022
• eBook ISBN: 9781803243481
• Topic: Informatica
• Subtopic: Intelligenza artificiale (IA) e semantica

OpenAI's GPT-3, ChatGPT, GPT-4 and Hugging Face transformers for language tasks in one book. Get a taste of the future of transformers, including computer vision tasks and code writing and assistance.Purchase of the print or Kindle book includes a free eBook in PDF format

Key Features

• Improve your productivity with OpenAI's ChatGPT and GPT-4 from prompt engineering to creating and analyzing machine learning models
• Pretrain a BERT-based model from scratch using Hugging Face
• Fine-tune powerful transformer models, including OpenAI's GPT-3, to learn the logic of your data

Book Description

Transformers are...well...transforming the world of AI. There are many platforms and models out there, but which ones best suit your needs?Transformers for Natural Language Processing, 2nd Edition, guides you through the world of transformers, highlighting the strengths of different models and platforms, while teaching you the problem-solving skills you need to tackle model weaknesses.You'll use Hugging Face to pretrain a RoBERTa model from scratch, from building the dataset to defining the data collator to training the model.If you're looking to fine-tune a pretrained model, including GPT-3, then Transformers for Natural Language Processing, 2nd Edition, shows you how with step-by-step guides.The book investigates machine translations, speech-to-text, text-to-speech, question-answering, and many more NLP tasks. It provides techniques to solve hard language problems and may even help with fake news anxiety (read chapter 13 for more details).You'll see how cutting-edge platforms, such as OpenAI, have taken transformers beyond language into computer vision tasks and code creation using DALL-E 2, ChatGPT, and GPT-4.By the end of this book, you'll know how transformers work and how to implement them and resolve issues like an AI detective.

What you will learn

• Discover new techniques to investigate complex language problems
• Compare and contrast the results of GPT-3 against T5, GPT-2, and BERT-based transformers
• Carry out sentiment analysis, text summarization, casual speech analysis, machine translations, and more using TensorFlow, PyTorch, and GPT-3
• Find out how ViT and CLIP label images (including blurry ones!) and create images from a sentence using DALL-E
• Learn the mechanics of advanced prompt engineering for ChatGPT and GPT-4

Who this book is for

If you want to learn about and apply transformers to your natural language (and image) data, this book is for you.You'll need a good understanding of Python and deep learning and a basic understanding of NLP to benefit most from this book. Many platforms covered in this book provide interactive user interfaces, which allow readers with a general interest in NLP and AI to follow several chapters. And don't worry if you get stuck or have questions; this book gives you direct access to our AI/ML community to help guide you on your transformers journey!

]]>

Index

Symbols

345M-parameter GPT-2 model

downloading 475, 476

A

accuracy score 125

Allen Institute for AI

reference link 257

AllenNLP 343

URL 343

Amazon Web Services (AWS) 1, 12, 392

artificial intelligence, properties

computing power 6

data 5

model architecture 5

prompt engineering 6

attention heads 459

attention masks

creating 76

Automated Machine Learning (AutoML) 120

automatic question generation 303, 304

B

BERT-based transformer

architecture 258

basic samples 261-267

difficult samples 267-273

running 258

SRL experiments 259, 260

BERT-base multilingual model 324, 325

BERT model

architecture 62

attention masks, creating 76

batch size, selecting 77

BERT tokenizer, activating 75

BERT tokens, adding 75

configuration 78, 80

CUDA, specifying as device for torch 72

data, converting into torch tensors 77

data, processing 76

dataset, loading 73-75

data, splitting into training set 76

data, splitting into validation set 76

encoder stack 62-65

fine-tuning 68-70

hardware constraints 71

holdout dataset, used for evaluating 86, 87

holdout dataset, used for predicting 86, 87

Hugging Face BERT uncased base model, loading 80-82

Hugging Face PyTorch interface, installing 71

hyperparameters for training loop 83

iterator, creating 77

key features 68

label lists, creating 75

Matthews Correlation...

Table of contents

• Preface
• What are Transformers?
• Getting Started with the Architecture of the Transformer Model
• Fine-Tuning BERT Models
• Pretraining a RoBERTa Model from Scratch
• Downstream NLP Tasks with Transformers
• Machine Translation with the Transformer
• The Rise of Suprahuman Transformers with GPT-3 Engines
• Applying Transformers to Legal and Financial Documents for AI Text Summarization
• Matching Tokenizers and Datasets
• Semantic Role Labeling with BERT-Based Transformers
• Let Your Data Do the Talking: Story, Questions, and Answers
• Detecting Customer Emotions to Make Predictions
• Analyzing Fake News with Transformers
• Interpreting Black Box Transformer Models
• From NLP to Task-Agnostic Transformer Models
• The Emergence of Transformer-Driven Copilots
• Appendix I — Terminology of Transformer Models
• Appendix II — Hardware Constraints for Transformer Models
• Appendix III — Generic Text Completion with GPT-2
• Appendix IV — Custom Text Completion with GPT-2
• Appendix V — Answers to the Questions
• Other Books You May Enjoy
• Index