Generative AI with Large Language Models - Deeplearning.AI and Amazon Web Services

This repository contains all the resources I used and created for the Generative AI with LLMs course on Coursera

Course Syllabus

Week 1

• Transformer Architecture
• Prompting and Prompt Engineering
• Inference Configuration Parameters
• Generative AI Project Lifecycle
• Pre-training LLMs
• Challenges - Quantisation & Computational Memory
• Pre-training vs Fine-tuning LLMs
• Efficient multi-GPU Strategies - DDP, FSDP
• Scaling laws and optimal compute models
• Pre-training for domain specific adaption

Week 2

• Instruction Fine-Tuning
• Catastrophic forgetting
• Multi Task Instruction Fine tuning
• Model Evaluation and Benchmarks
• Parameter Efficient Fine-Tuning (PEFT)
• Selective, Reparameterization, Additive methods
• LoRA and Soft Prompts

Week 3

• Reinforcement Learning from Human Feedback (RLHF)
• Reward model
• Proximal Policy Optimization
• Reward Hacking and KL Divergence
• Reinforcement Learning with AI Feedback (RLAIF)
• Model optimization for deployment - Distillation, Quantisation, Pruning
• LLM application architecture - Knowledge Cutoff & Hallucinations
• Retrieval Agumented Generation
• Chain-of-thought prompting
• Program aided language models (PAL)
• ReAct - Reasoning and action
• Langchain
• Responsible AI