SwayamInSync/S.E.A.R.C.H

S.E.A.R.C.H (Systematic Engine for Analyzed Retrieval and Contextual Handling): An advanced, multi-faceted information retrieval system.

S.E.A.R.C.H : Systematic Engine for Analyzed Retrieval and Contextual Handling

Table of Contents

• S.E.A.R.C.H : Systematic Engine for Analyzed Retrieval and Contextual Handling
  • Table of Contents
  • Introduction
  • Installation
  • Usage
    • For local machine
  • Demo

Introduction

S.E.A.R.C.H (Systematic Engine for Analyzed Retrieval and Contextual Handling) is an advanced, multi-faceted information retrieval system backed by Zephyr-7B. It integrates a Language Model with internet access for real-time, factual data acquisition, and employs a highly scalable Retrieval Augmented Generation (RAG) framework for efficient vector database processing. The application uniquely supports querying visual data, including images, and enhances reasoning through an embedded code interpreter. An innovative in-house sub-query engine significantly reduces hallucinations, ensuring precise, context-sensitive, and reliable responses

Installation

1. Clone repo and install dependencies

git clone https://github.com/SwayamInSync/S.E.A.R.C.H.git
cd S.E.A.R.C.H

pip install -r requirements.txt

2. Get API keys from SerpAPI and Cohere

Usage

• Below attached is a quickstart notebook to try it on Google Colab
  • 

For local machine

Run the following command inside terminal

export SERP_API=<SERP API KEY>
export COHERE_API=<COHERE API KEY>

Run the gradio interface

python app.py

If you want to adjust parameters in config.py

!python app.py --retriver_top_k 25 --reranker_top_k 10 --node_chunk_size 8000

Demo

demo.mp4