Recent advancements in LLM APIs allow for rich, structured outputs beyond standard text completion APIs (e.g. the OpenAI Function Call API). With text completion APIs, developers needed to write unstructured natural language f-strings and finnicky output parsing to coerce the model to returning certain outputs. In this webinar we show how combining the OpenAI Function API with Pydantic allows users to easily extract a range of structured outputs enabling creative new application features.
Link to slides: https://tome.app/fivesixseven/structured-cljflbvaj0szao53bdmoki3ig
Add comment