Week 1-Introduction to LLMs and the generative AI project lifecycle
Course Introduction
()
Contributor Acknowledgments
Introduction - Week 1
()
Generative AI & LLMs
()
LLM use cases and tasks
()
Text generation before transformers
()
Transformers architecture
()
Generating text with transformers
()
Transformers: Attention is all you need
Prompting and prompt engineering
()
Generative configuration
()
Generative AI project lifecycle
()
Introduction to AWS labs
()
Lab 1 walkthrough
()
Week 1-LLM pre-training and scaling laws
Pre-training large language models
()
Computational challenges of training LLMs
()
Optional video: Efficient multi-GPU compute strategies
()
Scaling laws and compute-optimal models
()
Pre-training for domain adaptation
()
Domain-specific training: BloombergGPT
Week 1 resources
Week 1-Lecture Notes (Optional)
Lecture Notes Week 1
Week 2-Fine-tuning LLMs with instruction
Introduction - Week 2
()
Instruction fine-tuning
()
Fine-tuning on a single task
()
Multi-task instruction fine-tuning
()
Scaling instruct models
Model evaluation
()
Benchmarks
()
Week 2-Parameter efficient fine-tuning
Parameter efficient fine-tuning (PEFT)
()
PEFT techniques 1: LoRA
()
PEFT techniques 2: Soft prompts
()
Lab 2 walkthrough
()
Week 2 Resources
Week 2-Lecture Notes (Optional)
Lecture Notes Week 2
Week 3-Reinforcement learning from human feedback
Introduction - Week 3
()
Aligning models with human values
()
Reinforcement learning from human feedback (RLHF)
()
RLHF: Obtaining feedback from humans
()
RLHF: Reward model
()
RLHF: Fine-tuning with reinforcement learning
()
Optional video: Proximal policy optimization
()
RLHF: Reward hacking
()
KL divergence
Scaling human feedback
()
Lab 3 walkthrough
()
[IMPORTANT] Reminder about end of access to Lab Notebooks
Week 3-LLM-powered applications
Model optimizations for deployment
()
Generative AI Project Lifecycle Cheat Sheet
()
Using the LLM in applications
()
Interacting with external applications
()
Helping LLMs reason and plan with chain-of-thought
()
Program-aided language models (PAL)
()
ReAct: Combining reasoning and action
()
ReAct: Reasoning and action
LLM application architectures
()
Optional video: AWS Sagemaker JumpStart
()
Week 3 resources
Week 3-Course conclusion and ongoing research
Responsible AI
()
Course conclusion
()
Week 3-Lecture Notes (Optional)
Lecture Notes Week 3
Week 3-Acknowledgments
Acknowledgments
(Optional) Opportunity to Mentor Other Learners