S Oracle Cloud Infrastructure Generative AI Professional LVC

Module 1: Course Introduction

• For Whom This Course is Intended
• Course Outline
• Fundamentals of Large Language Models (LLMs)
• Dive Deep on OCI Generative AI Service
• Build an LLM App using OCI Generative AI Service
• Meet Your Instructors
• Measuring Your Progress: Skill Checks
• Getting Answers: "Ask Your Instructor" Form & OU Community
• Best Practices and Retention Tips
• Keep Progressing: Path to Success

Module 2: Introduction to Large Language Models

• What is a Large Language Model?
• Module Overview
• LLM Architectures
• Encoders
• Decoders
• Encoder-Decoder Architectures
• Model Ontology
• Architectures at a Glance

Module 3: Prompting and Prompt Engineering

• Affecting Distribution over Vocabulary
• Prompting and Prompt Engineering
• In-Context Learning and Few-Shot Prompting
• Example Prompts & Advanced Prompting Strategies
• Issues with Prompting:
• Prompt Injection
• Memorization
• Training LLMs
• Hardware Costs
• Decoding Techniques:
• Greedy and Non-Deterministic Decoding
• Temperature
• Hallucination: Groundedness and Attributability
• LLM Applications
• Retrieval Augmented Generation (RAG)
• Code Models
• Multi-Modal Models
• Language Agents

Module 4: OCI Generative AI Service

• OCI Generative AI Introduction & Service Overview
• How OCI Generative AI Service Works
• Pretrained Foundational Models & Fine-Tuning
• Dedicated AI Clusters
• Demo: Generative AI Service Walkthrough
• Chat Models

Tokens, Parameters, Preamble Override, Temperature, Top-k, Top-p, Frequency/Presence Penalties

• Demo: Chat Models & Inference API
• Embedding Models
• Word & Sentence Embeddings
• Semantic Similarity
• Embeddings Use Cases
• Demo: Embedding Models

Module 5: Advanced Prompting, Fine-Tuning, and RAG

• Prompt Engineering Refresher
• LLMs as Next-Word Predictors
• Aligning LLMs to Instructions
• Prompt Formats & Few-Shot Learning
• Customizing LLMs with Your Data
• Training from Scratch
• Fine-Tuning Pretrained Models (T-Few)
• Fine-Tuning Benefits
• Inference Workflow
• Reducing Inference Costs
• Dedicated AI Cluster Sizing and Pricing
• Example Pricing & Demo
• OCI Generative AI Security
• Dedicated GPU & RDMA Network
• Model Endpoints
• Customer Data and Model Isolation
• Retrieval Augmented Generation (RAG)
• Framework, Techniques, and Pipeline
• Application & Evaluation
• Vector Databases
• Embeddings, Distance, Similarity, Dense Retrieval, Hybrid Search
• Oracle 23ai: AI Vector Search

Module 6: Application Development with OCI Generative AI

• Chatbot Introduction & Demo
• Chatbot Architecture & Components
• OCI Generative AI + LangChain Integration
• Models, Prompts, Chains
• Prompt Templates
• Setting Up Development Environment
• Demo: Prompts, Chains, and LLMs
• Extending Chatbot
• Adding Memory (LangChain Memory)
• Demo: Memory & Streamlit Integration
• Adding RAG
• Indexing, Retrieval, Generation
• Combining RAG + Memory
• Chat History as Context
• Demo: Tracing with LangSmith
• Chatbot Recap & Technical Architecture

Module 7: Deploying LLM Applications

• Deploy Chatbot to OCI Compute Instance (VM)
• Demo: Deployment to VM
• Deploy Chatbot to OCI Data Science
• Deploy LangChain Application as Model