Find answers from the community

Updated 9 months ago

Memory

https://github.com/run-llama/llama_index/issues/10493

@Logan M Running into the same issue here. Happy to walk you through to re-produce the bug. Any idea when your guys plan to fix this?
L
s
5 comments
I know the issue πŸ˜… haven't had time to prioritize a fix yet
Basically openai requires tool messages to have a matching assistant message. But the memory buffer might disregard one of them if they don't fit in the window
Ok, I see.
Any suggestion for short term fix?
It's a weird bug for sure. Took me a while to find a way to reproduce the bug consistently.
Yea super weird bug. Also not sure on a short term fix. Tbh I would really recommend tackling a PR if you have the time πŸ˜…
Add a reply
Sign up and join the conversation on Discord