I see. So it means it could perform well

It can, how much data do you have? I think it could make sense to use a 3rd party vector database if you have very large amounts of data

if you are using default vector store from us which is basically a simple in-memory knn search Vector DB. With larger amount documents, please try qdrant or pinecone

I see. Thanks. Will try that.

lets say my data is like this. which vector store is suitable for this?

@soap054 for your data it seems the type of vector store won't make too much difference. Since your data is bit of mixure of code and stack_trace and they are not Natrual language. I think you probably need to find better text/code encoder for your use purpose

I see. Do you have any suggestions?

So what I want to do is use a gpt model and ask it to analyze and give me information from the data. What would be a perfect combination to do that? thanks

can you try https://huggingface.co/docs/transformers/main/model_doc/code_llama