Hi, got more of a general question. So

At a glance

The community members are discussing the different ways to work with summary-based RAG (Retrieval Augmented Generation) in the llama_index library. The main options mentioned are:

SummaryExtractor, SummaryIndex, and DocumentSummaryIndex. The key differences are:

- SummaryExtractor uses a summary index to extract summaries.

- SummaryIndex gives all nodes to the language model in an iterative process to write a summary or answer a query.

- DocumentSummaryIndex summarizes each input document and then uses those summaries to select which documents should be used to answer a query.

The community members did not provide any explicit examples of use cases for each option.

aalexnn

Hi, got more of a general question. So there are at least 3 ways to work with summary based RAG in llama_index:

SummaryExtractor
SummaryIndex
DocumentSummaryIndex

What is the main difference in using those? Can someone provide some examples of use cases for each?

1 comment

LLogan M

A summary extractor just uses a summary index

A summary index gives all nodes to the LLM in an iterative process/response mode (usually I would use tree_summarize) to write a summary/answer a query

A document summary index summarizes each input document (using tree summarize), and then uses those summaries to select which documents should be used to answer a query

Add a reply

Find answers from the community

Hi, got more of a general question. So