Generative A IBootcamp-Presentation echnologies and how they connect Using tools available to building key Generative AI use cases

Generative AI Bootcamp

1. Generative AI Bootcamp Leveling Up Your Generative AI in 3 hours or Less!

2. Learning Goals For Today’s Bootcamp Build A Foundation of Generative AI Knowledge Understanding new terms and technologies and how they connect Exploring Large Language Models Comparing Different Model Types Improving the Results of a Model Understand the tools available to build with Using tools available to building key Generative AI use cases Common Use Cases Prompt Engineering Foundations Advanced Prompt Engineering Build a Generative AI application Applying the concepts into an application build Deploying a model Integrating the model into an application Handling model responses

3. 12 Lessons - Everything You Need to Know to Build Generative AI Applications Find the Repo Here: aka.ms / genai -beginners

4. Apply Here: https://azure.microsoft.com/products/ai-services/openai-service Request Access to Azure OpenAI Service

5. Who am I? Introduction to myself! /in/ koreypace / @ koreyspace / koreyspace

6. Who are you? Time to find out! https:// www.menti.com /alo59fd8wc6i

7. Generative AI Foundations Understanding the core terms and technologies

8. GPT-4 GPT-4-32k GPT-3.5-Turbo Text Embeddings Ada Davinci / Babbage Dalle3 OpenAI Models Open Source Models Llama 2 Falcon Dolly v2 Mistral Large Language Models (LLMs) Whisper

9. Understanding the Difference Criteria Open Source LLMs Proprietary LLMs  Availability Publicly Available and can be used by anyone Owned by an organization, access limited. Customization Allowed to be inspected and customized for different use cases than original foundation model Limited customization – fine tuning available in certain cases Performance May not be as performant as proprietary models Often optimized for production use Cost Free to use – may require hosting resources May require a subscription or payment based on use Maintenance May not be maintained in long term Often maintained and updated by model owner

10. Understanding the Difference Model Types Embeddings Models Text Generation Models

11. Understanding the Difference Model Types Chat Completions Image Generation

12. Exploring the Models Azure AI Studio - https:// ai.azure.com

13. Exploring the Models Azure AI Studio

14. Exploring the Models Azure AI Studio

15. Comparing Models Azure AI Studio

16. Improving Results Img source:  Four Ways that Enterprises Deploy LLMs | Fiddler AI Blog

17. Tools Available to Build with Generative AI What is out there and when to use it

18. Generative AI Applications Common Use Cases Text Image Search

19. Text Applications

20. LLMs sees prompt as a sequence of tokens.. Prompt Engineering How does a Prompt Work?

21. LLMs sees prompt as a sequence of tokens.. Prompt Engineering How does a Prompt Work? https:// platform.openai.com /tokenizer

22. Base LLMs will predict the next token Prompt Engineering How does a Prompt Work? https:// ai.azure.com /playground/gpt-35-turbo

23. Instruction-tuned LLM extends base behavior for task Prompt Engineering How does a Prompt Work? https:// ai.azure.com /playground/gpt-35-turbo

24. System Message Prompt Engineering How does a Prompt Work? Define the model’s profile, capabilities, and limitations for your scenario Define the model’s output format Provide example(s) to demonstrate the i ntended behavior of the model Provide additional behavioral guardrails

25. Prompt Engineering Building in Responsible AI with Metaprompting / System Message

26. Adding a Data Source Prompt Engineering Chat With Your Data

27. Providing Examples Prompt Engineering - Zero Shot Prompting - Few-Shot Prompting - Chain of Thought - Tree of Thought

28. Model Temperature Controlling Randomness Temperature 0 Repetitive / Deterministic Temperature .7 No Repetition / More Randomness

29. Break – 10 Mins

30. Search Applications

31. Understanding Embeddings How does a LLM find things? Embedding [1.76,0.33,] Vector Store / Database Some Data User Query Embeddings Model

32. Understanding Embeddings How does a LLM find things? Embedding [1.76,0.33,] Vector Store / Database YouTube Transcripts What is Azure ML? Embeddings Model

33. Understanding Chunking Fixed-Size Chunking Context Aware Chunking

34. Exploring Vector Databases

35. Image Generation Applications

36. Image Generation Use Cases Image Generation Editing Images Creating Variations Only Available on DALLE-2 Open AI Model Accepts: Prompt - Description of Image Size of Image - # of Pixels N - # of images generated Temperature – Randomness of output Model Accepts: Image – Original Image Image Mask – Area for edits Prompt - Description of edits N - # of images generated Temperature – Randomness of output Model Accepts: Image – Original Image N - # of variations Size of Image – # of Pixels

37. Exploring Vector Databases

38. Build a Generative AI Application Let’s build something together!

39. Deploying a Model

40. Connecting to a Model – Sample Code

41. Connecting to a Model – Making a Request

42. Observing the Response

43. Integrating Responses – Function Calling User: ”I am a beginner with Azure AI” LLM Skill: Beginner Product: Azure Function find_course(skill, product) API Call /course/{product}/{skill } LLM I recommend AI-900

44. Integrating Responses – Function Calling

45. lead the AI transformation The opportunity is yours to

46. 12 Lessons - Everything You Need to Know to Build Generative AI Applications Find the Repo Here: aka.ms / genai -beginners