LLMs are slow.
Caching can help, but developers should have control over how caches operate at different scales
In next week's blog, we will show you how to write a @pydantic compatible cache mechanism for use with Instructor.
1. For a single session:... See more