Kaveen Kumarasinghe

This is still relevant!

What langchain version are you on? Try updating langchain to at least langchain==0.0.105

The screenshot is the result of me doing a query on a GPTTreeIndex that contains content from the web about "Harald Baldr". The query was simply "Who is harald baldr"?

 Getting this also from the result of querying a simple vector index -- looks like the ChatGPT personality behaviour really breaks the internal dialogue when doing a query

Hmm, seems like the same issues if not worse now: These are queries from simple vector indexes with nodes=2

Eventually it starts just outright erroring out on the requests ``` 500 {'error': {'message': 'The server experienced an error while processing your request. Sorry about that! You can retry your...

 text-davinci-002 has the same issues unfortunately, knowledge graph index is entirely unusable for me

It seems like it's doing one request, for over the max token amount, instead of batching it into like two requests of smaller size?

Yup still also an issue for me

This is fine for me on 0.4.17, still quite slow but not entirely blocked