Deploying generative AI models on Azure

The two best ways of interacting with readily available large language models are currently either through OpenAI directly or the Azure OpenAI Service.

Both services require you to provide an API key to identify yourself and receive charges from requests to the LLM that you wish to interact with.

Both services also offer a variety of different models to interact with. These models are periodically updated with minor changes. Additionally, the model catalog regularly expands as new models are trained, evaluated, and made available to the public.

While both of these services are viable ways of interacting with LLMs, this book will focus on Azure OpenAI Service. This is because Azure OpenAI Service tends to be a better option for enterprise users, due to its privacy, security, and content safety features.

Additionally, many organizations use Azure for other services, and there can be strong advantages to having your LLMs hosted in the...