Does the RAG feature in OpenSearch support streaming responses, similar to ChatGPT? Otherwise, the user has to wait too long for each query.
Each query takes several seconds.
It feels like the more relevant documents there are in the index, the slower the RAG queries become.
How to solve this problem