Evaluating Retrieval-Augmented Generation (RAG) with any combination of LLMs, Vector DBs, and Ingestion Strategy

To help developers test their RAG systems, we added a RAG experiment class to our open-source library [PromptTools](https://github.com/hegelai/prompttools). It allows users to easily experiment with different combinations of LLMs and vector DBs, and evaluate the results of their whole pipeline.

In particular, you can experiment with:

1. Chunking up your documents into different sizes
2. Pre-processing those documents in various ways
3. Inserting those documents into your vector DBs with various vectorizer and embedding function, and accessing them with different …

aipromptprogramming anthropic chat chroma dbs documents embedding experiment function functions google google vertex llama llm openai palm pre-processing processing replicate them vector vertex weaviate