PyTorch Ultimate 2024 - From Basics to Cutting-Edge

PyTorch Ultimate 2024 - From Basics to Cutting-Edge

by Bert Gollnick
Released September 2023
Publisher(s): Packt Publishing
ISBN: 9781801070089

Read it now on the O’Reilly learning platform with a 10-day free trial.

O’Reilly members get unlimited access to books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.

Start your free trial

Video description

PyTorch is a Python framework developed by Facebook to develop and deploy deep learning models. It is one of the most popular deep-learning frameworks nowadays.

You will begin with learning the deep learning concept. Dive deeper into tensor handling, acquiring the finesse to create and manipulate tensors while leveraging PyTorch's automatic gradient calculation through Autograd. Then transition to modeling by constructing linear regression models from scratch. After that, you will dive deep into classification models, mastering both multilabel and multiclass. You will then see the theory behind object detection and acquire the prowess to build object detection models. Embrace the cutting edge with YOLO v7, YOLO v8, and faster RCNN, and unleash the potential of pre-trained models and transfer learning.

Delve into RNNs and look at recommender systems, unlocking matrix factorization techniques to provide personalized recommendations. Refine your skills in model debugging and deployment, where you will debug models using hooks, and navigate the strategies for both on-premise and cloud deployment. Finally, you will explore ChatGPT, ResNet, and Extreme Learning Machines.

By the end of this course, you will have learned the key concepts, models, and techniques, and have the confidence to craft and deploy robust deep-learning solutions.

What you will learn

  • Grasp deep learning concepts and install tools/packages/IDE/libraries
  • Master CNN theory, image classification, layer dimensions, and transformations
  • Dive into audio classification using torchaudio and spectrograms
  • Do object detection with the help of YOLO v7, YOLO v8, and Faster RCNN
  • Learn word embeddings, sentiment analysis, and pre-trained NLP models
  • Deploy models using Google Cloud and other strategies

Audience

This course is ideal for Python developers and data enthusiasts seeking to expand their skills. This will also benefit aspiring data scientists, machine learning engineers, AI enthusiasts, and anyone intrigued by the transformative potential of deep learning. Whether you are a beginner or possess some prior knowledge, this course offers a smooth progression that will empower you to develop, deploy, and innovate with deep learning models using PyTorch. Basic Python knowledge is required to fully engage with the material.

About the Author

Bert Gollnick: Bert Gollnick is a proficient data scientist with substantial domain knowledge in renewable energies, particularly wind energy. With a rich background in aeronautics and economics, Bert brings a unique perspective to the field. Currently, Bert holds a significant role at a leading wind turbine manufacturer, leveraging his expertise to contribute to innovative solutions.

For several years, Bert has been a dedicated instructor, offering comprehensive training in data science and machine learning using R and Python. The core interests of Bert lie at the crossroads of machine learning and data science, reflecting a commitment to advancing these disciplines.

Table of contents

  1. Chapter 1 : Course Overview and System Setup
    1. Course Overview
    2. PyTorch Introduction
    3. System Setup
    4. How to Get the Course Material
    5. Setting Up the conda Environment
    6. How to Work with the Course
  2. Chapter 2 : Machine Learning
    1. Artificial Intelligence (101)
    2. Machine Learning (101)
    3. Machine Learning Models (101)
  3. Chapter 3 : Deep Learning Introduction
    1. Deep Learning General Overview
    2. Deep Learning Modeling 101
    3. Performance
    4. From Perceptron to Neural Network
    5. Layer Types
    6. Activation Functions
    7. Loss Functions
    8. Optimizers
    9. Deep Learning Framework
  4. Chapter 4 : Model Evaluation
    1. Underfitting Overfitting (101)
    2. Train Test Split (101)
    3. Resampling Techniques (101)
  5. Chapter 5 : Neural Network from Scratch
    1. Section Overview
    2. Neural Network from Scratch (101)
    3. Calculating the dot-product (Coding)
    4. Neural Network from Scratch (Data Prep)
    5. Neural Network from Scratch Modeling __init__ Function
    6. Neural Network from Scratch Modeling Helper Functions
    7. Neural Network from Scratch Modeling Forward Function
    8. Neural Network from Scratch Modeling Backward Function
    9. Neural Network from Scratch Modeling Optimizer Function
    10. Neural Network from Scratch Modeling Train Function
    11. Neural Network from Scratch Model Training
    12. Neural Network from Scratch Model Evaluation
  6. Chapter 6 : Tensors
    1. Section Overview
    2. From Tensors to Computational Graphs (101)
    3. Tensor (Coding)
  7. Chapter 7 : PyTorch Modeling Introduction
    1. Section Overview
    2. Linear Regression from Scratch (Coding, Model Training)
    3. Linear Regression from Scratch (Coding, Model Evaluation)
    4. Model Class (Coding)
    5. Exercise: Learning Rate and Number of Epochs
    6. Solution: Learning Rate and Number of Epochs
    7. Batches (101)
    8. Batches (Coding)
    9. Datasets and Dataloaders (101)
    10. Datasets and Dataloaders (Coding)
    11. Saving and Loading Models (101)
    12. Saving and Loading Models (Coding)
    13. Model Training (101)
    14. Hyperparameter Tuning (101)
    15. Hyperparameter Tuning (Coding)
  8. Chapter 8 : Classification Models
    1. Section Overview
    2. Classification Types (101)
    3. Confusion Matrix (101)
    4. ROC Curve (101)
    5. Multi-Class 1: Data Prep
    6. Multi-Class 2: Dataset Class (Exercise)
    7. Multi-Class 3: Dataset Class (Solution)
    8. Multi-Class 4: Network Class (Exercise)
    9. Multi-Class 5: Network Class (Solution)
    10. Multi-Class 6: Loss, Optimizer, and Hyperparameters
    11. Multi-Class 7: Training Loop
    12. Multi-Class 8: Model Evaluation
    13. Multi-Class 9: Naive Classifier
    14. Multi-Class 10: Summary
    15. Multi-Label (Exercise)
    16. Multi-Label (Solution)
  9. Chapter 9 : CNN: Image Classification
    1. Section Overview
    2. CNNs (101)
    3. CNN (Interactive)
    4. Image Preprocessing (101)
    5. Image Preprocessing (Coding)
    6. Binary Image Classification (101)
    7. Binary Image Classification (Coding)
    8. Multi-Class Image Classification (Exercise)
    9. Multi-Class Image Classification (Solution)
    10. Layer Calculations (101)
    11. Layer Calculations (Coding)
  10. Chapter 10 : CNN: Audio Classification
    1. Audio Classification (101)
    2. Audio Classification (Exercise)
    3. Audio Classification (Exploratory Data Analysis)
    4. Audio Classification (Data Prep-Solution)
    5. Audio Classification (Model-Solution)
  11. Chapter 11 : CNN: Object Detection
    1. Section Overview
    2. Accuracy Metrics (101)
    3. Object Detection (101)
    4. Object Detection with detecto (Coding)
    5. Training a Model on GPU for Free (Coding)
    6. YOLO (101)
    7. Labeling Formats
    8. YOLOv7 Project (101)
    9. YOLOv7 Coding: Setup
    10. YOLOv7 Coding: Data Prep
    11. YOLOv7 Coding: Model Training
    12. YOLOv7 Coding: Model Inference
    13. YOLOv8 Coding: Model Training and Inference
  12. Chapter 12 : Style Transfer
    1. Section Overview
    2. Style Transfer (101)
    3. Style Transfer (Coding)
  13. Chapter 13 : Pre-Trained Networks and Transfer Learning
    1. Section Overview
    2. Transfer Learning and Pre-Trained Networks (101)
    3. Transfer Learning (Coding)
  14. Chapter 14 : Recurrent Neural Networks
    1. Section Overview
    2. RNN (101)
    3. LSTM (Coding)
    4. LSTM (Exercise)
  15. Chapter 15 : Recommender Systems
    1. Recommender Systems (101)
    2. RecSys (Coding 1/4) - Dataset and Model Class
    3. RecSys (Coding 2/4) - Model Training and Evaluation
    4. RecSys (Coding 3/4) - Users and Items
    5. RecSys (Coding 4/4) - Precision@k and Recall@k
  16. Chapter 16 : Autoencoders
    1. Section Overview
    2. Autoencoders (101)
    3. Autoencoders (Coding)
  17. Chapter 17 : Generative Adversarial Networks
    1. Section Overview
    2. GANs (101)
    3. GANs (Coding)
    4. GANs (Exercise)
  18. Chapter 18 : Graph Neural Networks
    1. Graph Neural Networks (101)
    2. Graph Introduction (Coding)
    3. Node Classification (Coding: Data Prep)
    4. Node Classification (Coding: Model Train)
    5. Node Classification (Coding: Model Eval)
  19. Chapter 19 : Transformers
    1. Transformers 101
    2. Vision Transformers (ViT)
    3. Train ViT on Custom Dataset (Coding)
  20. Chapter 20 : PyTorch Lightning
    1. PyTorch Lightning (101)
    2. PyTorch Lightning (Coding)
    3. Early Stopping (101)
    4. Early Stopping (Coding)
  21. Chapter 21 : Semi-Supervised Learning
    1. Semi-Supervised Learning (101)
    2. Supervised Learning (Reference Model, Coding)
    3. Semi-Supervised Learning (1/2: Dataset and Dataloader)
    4. Semi-Supervised Learning (2/2 Modeling)
  22. Chapter 22 : Natural Language Processing (NLP)
    1. Natural Language Processing (101)
    2. Word Embeddings Intro (101)
    3. Sentiment OHE Coding Introduction
    4. Sentiment OHE (Coding)
    5. Word Embeddings with Neural Network (101)
    6. GloVe: Get Word Embedding (Coding)
    7. Glove: Find Closest Words (Coding)
    8. GloVe: Word Analogy (Coding)
    9. GloVe Word Cluster (101)
    10. GloVe Word (Coding)
    11. Sentiment with Embedding (101)
    12. Sentiment with Embedding (Coding)
    13. Apply Pre-Trained Natural Language Processing Models (101)
    14. Apply Pre-Trained Natural Language Processing Models (Coding)
    15. Vector Databases (101)
    16. Retrieval Augmented Generation (101)
    17. Claude 3 (101)
    18. Claude 3 (Coding)
    19. Zero-Shot Classification (101)
    20. Zero-Shot Classification (Coding)
  23. Chapter 23 : Miscellaneous Topics
    1. OpenAI ChatGPT (101)
    2. ResNet (101)
    3. Inception (101)
    4. Inception Module (Coding)
    5. Extreme Learning (101)
    6. Extreme Learning (Coding)
  24. Chapter 24 : Model Debugging
    1. Hooks (101)
    2. Hooks (Coding)
  25. Chapter 25 : Model Deployment
    1. Model Deployment (101)
    2. Flask On-Premise, Hello World (Coding)
    3. API On-Premise with Deep Learning Model (Coding)
    4. API On-Premise: How to Consume the Data (Coding)
    5. Google Cloud: Deploy Model Weights (Coding)
    6. Google Cloud: Deploy REST API (Coding)
  26. Chapter 26 : Final Section
    1. Thank You and Further Resources

Product information

  • Title: PyTorch Ultimate 2024 - From Basics to Cutting-Edge
  • Author(s): Bert Gollnick
  • Release date: September 2023
  • Publisher(s): Packt Publishing
  • ISBN: 9781801070089