Chatbot created using Concept of RAG( Retrieval Augmented Generation)

Demo Video

Screencast.from.01-25-2024.11.53.14.PM.webm

Project Description

This is a chatbot implemented using concept of RAG with Tech-News Dataset on Zephyr Alpha Model.

• Dataset: TechNews dataset: BBC Dataset http://mlg.ucd.ie/datasets/bbc.html

• RAG Technique Used:

Hybrid Search and Re-ranking to retrieve document faster provided with the given context.

• Vector Database : FAISS

Vector databases are particularly optimized for similarity search. This means they are capable of quickly finding vectors that are closest (or most similar) to a given query vector.

FAISS (Facebook AI Similarity Search), developed by Facebook’s AI team is specifically engineered for the efficient similarity search and clustering of dense vectors. FAISS excels in handling large datasets, offering both speed and accuracy in retrieving similar items. This made it an ideal choice for this task that involves searching through large volumes of high-dimensional data i.e. identifying documents relevant to a particular query.

• LLM — Zephyr-7b-alpha

Read Medium article on:
Advanced RAG Implementation using Hybrid Search, Reranking with Zephyr Alpha LLM

Access colab notebook here:

Colab notebook