Summary
In this chapter, we saw how generative AI can involve using large pre-trained models to generate content from simple prompts.
We saw how chat completions allow you to generate text based on a prompt and conversation history, how image generation allows you to create serious and silly imagery, and how text embeddings help determine how similar or different the topics of two pieces of text are.
The models we interacted with in this chapter were all hosted on Azure in an Azure OpenAI Service, but generative AI can work with models directly on OpenAI, models hosted on another service (through a different set of APIs), or even models downloaded locally, using a library such as LLamaSharp. See the Further reading section for more information on these resources.
Working with generative AI does not require a notebook environment, but in this chapter, we saw how Notebooks can make the generative AI experience easier. Notebooks can specifically help you inspect complex result...
