Has anyone had any experience using LlamaIndex for a production, somewhat scaled use case? I've been having issues with speed and reliability with the vector DB provider I'm using and wondering whether it's an issue with just that provider's integration, or in general how deep the rabbit hole might go.
06/28/2023, 12:44 AM
Is it an async (or lack of) issue? I've encountered Python libraries that focus on the synchronous use cases, but fail to adequately document the async use cases.
06/28/2023, 12:57 AM
It's less about sync vs async and moreso a performance issue
As in sometimes OpenAI API seems to persistently fail with no visible error (this might be purely an OpenAI issue but I haven't seen this problem in other OpenAI issues) and the time to index small text documents seems super high