LocalPrompt

Description

This project is a simple and lightweight client-server program for interfacing with local LLMs using ollama, and LLMs in groq using groq api.

This project demonstrates a simple and lightweight client-server program for interfacing with Large Language Models (LLMs) on your local machine using ollama and also interfacing groq using groq api.

Prerequisites

• Python 3.7+
• aiohttp library
• ollama installed in the system

Installation

Clone the Repository

git clone https://github.com/im-pramesh10/LocalPrompt.git
cd LocalPrompt/backend

Create python virtual environment

python -m venv venv

else

python3 -m venv venv

or

virtualenv venv

Activate venv

• For Windows

.\venv\Scripts\activate

• For Linux and MacOs

source venv/bin/activate

Install from requirements.txt

pip install -r requirements.txt

Usage

• cd into backend folder
• Activate the virtual environment and run the following command:

python simple_async_server.py

• navigate to http://localhost:8000 to use the program
• you can change your default port from setting.py
• Ensure Ollama is running in the background and the Phi model is pulled. This example uses the Phi model.

Modifying for Other LLMs

For single prompt

To connect LocalPrompt setup with a different LLM using Ollama:

• Write your code inside custom_model_api function inside api_call.py file
• Return your response in the following format:

{'response':'You need to set up your custom model function inside api_call.py file inside backend folder'}

[!NOTE] Make sure to restart server after each changes.