Hello! A question about chunks. When querying the vector database, why pass the chunk_size? I get it when we created the index - chunk size determined the size of pieces of text to be stored, but here, why?
Another question is what happens when chunk_sizes are different: say, when we created the index, it was 1024 and when querying it's 256?
Thanks!
service_context = ServiceContext.from_defaults(llm_predictor=llm_predictor, chunk_size=chunk_size,
callback_manager=callback_manager)
index = VectorStoreIndex.from_vector_store(vector_store, service_context)
chat_engine = index.as_chat_engine()
response = chat_engine.chat(q)