2 ms·
Can the vectorization of chunks and finding context close to query be done with any LLMs and then only relevant chunks be sent to OpenAI?
by shekhar101 4y ago
Can the vectorization of chunks and finding context close to query be done with any LLMs and then only relevant chunks be sent to OpenAI?
- naveedjanmo 4y agoVectorisation is done via OpenAI's embedding API. And the chunking/querying is happens through the Langchain library. But there are a few different ways of doing it - another good library is LLamaIndex.
- shekhar101 4y agoThanks a lot! Do you _have_ to do vectorization and querying with the same LLM? Can someone do vectorization with 1 and do querying with reevant chunks with another?