GPU Recommender & Research Chatbot

A full‑stack application to help users discover the best GPUs based on region, budget, and usage timeline, and to interact with a research‑paper–powered chatbot for detailed insights.

Video - https://drive.google.com/file/d/1YuwXrVpLt-1kdGy2C1AQuFUaU68KXBgv/view?usp=sharing

Documentation - https://drive.google.com/file/d/1yUdIYdmQZP2WEUQSlfVG41omk_jmYKGc/view?usp=sharing

---

🔗 Live Backend API

The recommendation engine backend is deployed on Render:

https://gpu-recommender.onrender.com/api/recommend-gpus

This API returns the top 5 GPUs for given parameters (region, operating system, timeline, budget).

---

🖥️ Frontend (Streamlit)

The Streamlit frontend combines three interactive steps:

1. Step 1: Fetch & Clean GPU Names

   • Collects user inputs (region, OS, timeline, budget)
   • Calls the backend API and displays cleaned GPU names

2. Step 2: GPU Recommendation (LLM-based)

   • Uses LangChain + Groq LLM to generate short 2–3 line recommendations for each GPU

3. Step 3: PDF ChatBot

   • Loads research papers from a local research_papers/ directory
   • Answers user questions with source snippets

Folder Structure

project-root/
├── backend/               # (Optional local copy of API code)
├── frontend/              # Streamlit client
│   ├── main.py            # Entry point for Streamlit
│   ├── research_papers/   # PDF files for chatbot context
│   └── .env               # GROQ_API_KEY, HF_TOKEN
└── README.md

Getting Started

1. Clone the repo

   git clone <repository-url>
   cd project-root

2. Install dependencies

   pip install -r frontend/requirements.txt

3. Set up environment

   Create a .env file inside frontend/:

   GROQ_API_KEY=<your_groq_api_key>
   HF_TOKEN=<your_huggingface_token>

4. Run the Streamlit App

   cd frontend
   streamlit run main.py

This will open the app in your default browser at http://localhost:8501.

---

🚀 Tech Stack

• Backend

  • Python, FastAPI (or Flask)
  • Hosted on Render.io

• Frontend

  • Streamlit
  • Requests for API calls
  • LangChain + ChatGroq for LLM integrations
  • FAISS for vector search over PDFs
  • SentenceTransformers for embeddings

• Research Papers

  • Local research_papers/ directory
  • RecursiveCharacterTextSplitter for chunking

---

⭐ Features

• Dynamic top‑5 GPU recommendation by region, OS, timeline, and budget
• Cleaned, human‑readable GPU names
• Short, LLM‑generated usage recommendations
• Interactive PDF‑based chatbot with snippet‑level context
• Easy deployment with Render (backend) and Streamlit (frontend)

---

🔧 Configuration & Deployment

• Backend: ensure RENDER environment and any API keys are set on Render.io
• Frontend: configure .env and run streamlit run main.py

---

🤝 Contributing

Feel free to open issues or submit pull requests for improvements!

---

File to run: frontend/main.py

Happy GPU‑hunting and research exploration! 🎉