RAG Architecture
A typical RAG application has two main components: Indexing: a pipeline for ingesting data from a source and indexing it. This usually happens offline. Retrieval and generation: the actual RAG chain, which takes the user query at run time and retrieves the relevant data from the index, then passes that to the model. The most […]