Deep Learning for Modern AI
Published by Pearson
Start building and training generative AI and multimodal models for modern tasks
- Real-world deep learning examples with practical use-cases for modern Generative AI
- Fully fleshed-out code for repeatable use after the session is over
- A focus on experimentation to encourage continued learning and iteration on topics discussed during the training
This training provides the theory and practical concepts for a comprehensive introduction to machine learning and deep learning with PyTorch —foundational knowledge needed to successfully build and train GenAI and multimodal models. By making our way through several real-world case studies including object recognition and text classification this session is an excellent crash course in deep learning with PyTorch.
We use tools including large pre-trained models and model training dashboards to set up reproducible deep learning experiments and build machine learning models optimized for performance. There are several code examples throughout the training to help solidify the theoretical concepts that will be introduced. Models like Stable Diffusion, Llama 3, GPT, and BERT are highlighted as we uncover the training and optimization strategies to get the most of our models' performance, speed, and memory usage.
What you’ll learn and how you can apply it
- How to train a neural network using gradient descent on most datasets
- How to load pre-trained models to achieve state-of-the-art deep learning performance for modern applications
- How modern Generative AI can produce text, images, and video with relative ease
This live event is for you because...
- You’re a software engineer with an interest in learning and practicing deep learning
- You are comfortable using Python and scikit-learn
- You want to learn a modern deep learning framework
Prerequisites
- Python 3 proficiency with some familiarity with working in interactive Python environments including Notebooks (Jupyter / Google Colab / Kaggle Kernels)
- Working knowledge of some machine learning library like scikit-learn
Course Set-up
- A GitHub repository with the slides / code / links will be provided upon completion
- Attendees will need to have access to the notebooks in the GitHub: https://github.com/sinanuozdemir/oreilly-pytorch-dl
Recommended Preparation
- Watch: Machine Learning with scikit-learn LiveLessons by David Mertz
- Watch: Real-World Machine Learning video edition by Henrik Brink, Joseph W. Richards, Mark Fetherolf
- Read: Deep Learning Illustrated: A Visual, Interactive Guide to Artificial Intelligence by Jon Krohn, Grant Beyleveld and Aglaé Bassens
Recommended Follow-up
- Read: Applied Deep Learning with TensorFlow 2: Learn to Implement Advanced Deep Learning Techniques with Python by Umberto Michelucci
- Watch: Deep Learning with Python, Second Edition, Video Edition by Francois Chollet
- Explore: Getting Started with Data, LLMs and ChatGPT by Sinan Ozdemir
Schedule
The time frames are only estimates and may vary according to how the class is progressing.
Segment 1: Introduction to deep learning and PyTorch (30 min)
- Learn the basic concepts of deep learning including neural networks and gradient descent
- See how PyTorch and Python packages like transformers, peft, and diffusers make training and deploying deep learning models simpler and reproducible
- Exercise - Build a tiny neural network and train it using gradient descent
Segment 2: Experiment design and introduction to loss functions (30 min)
- Exercise - Set up experiments and experiment tracking tools for a reproducible environment
- Learn the different loss functions and metrics used to quantify deep learning performance
Break / Q&A 5 min
Segment 3: Training text based LLMs (60 min)
- See how the attention mechanism changed the way we think about modern Natural Language Processing via the invention of the Transformer in 2017
- Exercise - Train LLMs like BERT to be classifiers
- Exercise - Use techniques like QLoRA to optimize for speed and memory while training larger models like Llama 3 for conversational purposes
Break / Q&A 5 min
Segment 4: Multimodal Deep Learning (50 min)
- See how autoencoders, U-nets and a forward/reverse diffusion process make generating videos and images possible
- Exercise - Train a Visual Q/A system from scratch using off the shelf open source tools
Break / Q&A 5 min
Segment 5: Deployment with PyTorch (30 min)
- See how techniques like pruning and quantization transform models from experiment-grade to production-grade
- Exercise - Quantize a Llama 3 model to measure speed/memory boosts while interrogating the delta in accuracy.
Segment 6: Course wrap-up and next steps (10 min)
- Next steps / further resources
Final Q/A 10 min
Your Instructor
Sinan Ozdemir
Sinan Ozdemir is founder and CTO of LoopGenius, where he uses state-of-the-art AI to help people create and run their businesses. He has lectured in data science at Johns Hopkins University and authored multiple books, videos and numerous online courses on data science, machine learning, and generative AI. He also founded the recently acquired Kylie.ai, an enterprise-grade conversational AI platform with RPA capabilities. Sinan most recently published Quick Guide to Large Language Models, and launched a podcast audio series, AI Unveiled. Ozdemir holds a master’s degree in pure mathematics from Johns Hopkins University.