Azure AI-050: Generative AI Solutions with Azure OpenAI

Generative AI enables capabilities that create human-like responses and original content across text, images, and code, powered by large language models via Azure OpenAI.

Learn how to request access to Azure OpenAI service, why access is limited to companies, and explore the OpenAI Studio with GPT and DALL-E demos.

Explore azure openai by creating and deploying a chat model in the openai studio, using models like GPT-4 and GPT-3.5-turbo-16k, with system messages and dalle image generation.

Learn how to configure Azure OpenAI parameters in OpenAI Studio, including maximum response, temperature and topic settings, plus frequency and presence penalties, with deployment and few-shot examples.

It demonstrates integrating Azure OpenAI into a Python app using OpenAI package, loading endpoint and deployment from dot env file, and using system and user messages to generate responses.

Show how to fetch the keys, endpoint, and deployment name from the Azure OpenAI service, and copy them into your application and OpenAI Studio deployment.

Explore prompt engineering, a technique that improves accuracy and relevancy by crafting descriptive prompts for AI tools like ChatGPT, with examples using simple language and markdown tables.

Generate code with Azure OpenAI service by using natural language prompts and prompt engineering, choosing Python or C sharp, and using the OpenAI Studio chart playground.

Demonstrates generating code with Azure OpenAI from a Python app. Load endpoint and deployment from a .env file, use system and user prompts to add comments, create unit tests, and fix a function, with results saved to results/app.txt.

Explore image generation with Azure OpenAI service using the DALL-E model, create deployments, and refine outputs with prompts and styling to produce themed images like an elephant on a skateboard.

Demonstrates generating images in a Python app using an OpenAI API endpoint, passing prompts, image count, and size, and displaying the image URL without using the SDK.

Learn to use your own data with Azure OpenAI service, compare fine tuning with retrieval augmented generation, and ground LLM responses via your data sources.

This demo shows using your own data with Azure OpenAI Service by uploading PDFs, indexing them with a data source and search resource, and querying the AI from those PDFs.

Apply responsible ai practices for azure openai using a four-stage framework—identify, measure, mitigate, operate—and content filters to block hate, sexual content, violence, and self-harm, with incident and rollback planning.

Activate and customize content filters in Azure OpenAI to block harmful prompts, support categories like hate and self-harm, and set severity levels from safe to high.

Discover Azure AI Foundry, the unified platform that merges the Azure OpenAI models with other Azure AI services. Learn to create an AI hub and projects and manage access.

Explore Azure AI Foundry, with models from OpenAI, Microsoft, Meta, Snowflake, and Databricks to generate text, perform image classification, object detection, and text-to-image tasks, hub setup in the Azure portal.

Deploy a GPT model in the foundry portal, create a hub and a project, and test the model in the playground using a system message to set the AI persona.