GroqWarp

GroqWarp is a Streamlit-based app designed to compare the performance of RAG (Retrieval-Augmented Generation) using two powerful models: Groq and Ollama. The app enables users to input a question, process it with both models using document retrieval from a custom dataset, and compare the response times and accuracy.

Tech Stack

Features

• Model Comparison: Compare Groq (Gemma 2-9B-IT) and Ollama (Gemma 2-9B) based on performance (speed and accuracy) for a given question.
• Document Retrieval: Uses FAISS to retrieve context from uploaded PDF documents.
• Side-by-Side Results: Display responses from both models along with the time taken to process.
• Performance Visualization: A bar chart for visual comparison of response times between the models.
• Textual Context: View the context from the documents used to generate responses.

Installation

To install and set up the GroqWarp project, follow these steps:

1. Clone the repository:

   git clone https://github.com/your-username/GroqWarp.git
   cd GroqWarp

2. Set up a virtual environment (using conda or venv):

   conda create --name groqwarp-env python=3.8
   conda activate groqwarp-env

3. Install dependencies:

   pip install -r requirements.txt

4. Ensure the following dependencies are included in requirements.txt:

   faiss-cpu==1.10.0
   groq==0.18.0
   langchain-groq==0.2.5
   langchain-ollama==0.2.3
   PyPDF2==3.0.1
   langchain-google-genai==2.0.11
   langchain==0.3.20
   streamlit==1.43.1
   langchain_community==0.3.19
   python-dotenv==1.0.1
   pypdf==5.3.1
   pandas
   altair

Prerequisites

Before running GroqWarp, make sure you have the following prerequisites:

• Python 3.12 or higher
• Conda (or virtual environment)
• Required libraries (faiss-cpu, groq, langchain, streamlit, pandas, altair, etc.)

To set up the environment using Conda and install dependencies from the requirements.txt file, use the following command:

conda create --name groqwarp-env python=3.12
conda activate groqwarp-env
pip install -r requirements.txt

Run the App

Make sure to start Ollama before running the app to ensure both models are properly initialized. You can start Ollama by running the following command:

ollama serve

Start the app by running the following command:

streamlit run app.py

The app will launch in your browser at http://localhost:8501.

Usage

Upload your dataset: Place your PDF files in the ./pdfs directory.

Ask a question: Input a question in the provided text field to compare how Groq and Ollama handle it.

View results: The app will display the answers from both models and the time taken to process the question.

Performance Comparison: A bar chart will show the time difference between the two models.

Contributing

Welcome contributions to GroqWarp! If you would like to contribute, please fork the repository, create a new branch, and submit a pull request with your changes.

License

GroqWarp is open-source software licensed under the MIT License. See the LICENSE file for more information.

Acknowledgements

Libraries Used:

• Streamlit for the app interface.
• FAISS for document similarity search.
• Groq and Ollama for language model inference.
• Langchain for document processing and retrieval.
• Altair for data visualizations.