elasticsearch-labs icon indicating copy to clipboard operation
elasticsearch-labs copied to clipboard

Another example of RAG using LangChain and Llama3

Open ezimuel opened this issue 1 year ago • 1 comments

This PR contains another example for showing how to build a RAG system using Llama3 (executed in Google Colab), LangChain and Elasticsearch. The use case shows how to extend the knowledge of Llama3 answering to the question "How many moons has Neptune?". Recently, on 23 February 2024 three new moons around Uranus and Neptune. If we ask to Llama3 we get 14 moons as answer, using the RAG solution we get the correct answer that is 16.

You can try the Colab online here.

ezimuel avatar Jul 04 '24 07:07 ezimuel