AI Chatbot with RAG and HuggingFace Models

A professional, modular chatbot app using Streamlit, supporting PDF/CSV/XLSX upload, Retrieval-Augmented Generation (RAG), and custom HuggingFace models (e.g., amiguel/GM_Qwen1.8B_Finetune).

Features

Modern Streamlit chat UI with Tw Cen MT font and company logo
Upload PDF, CSV, XLSX files for context
Uses RAG pipeline for document QA
Customizable HuggingFace text generation model
Modular, production-ready codebase

Quickstart

1. Clone and Install

git clone <your-repo-url>
cd chatbot_app
pip install -r requirements.txt

2. Run Locally

streamlit run app.py

3. Docker Deploy

docker build -t chatbot-app .
docker run -p 8501:8501 chatbot-app

Deployment

A. AWS (GPU)

Launch a GPU EC2 instance (e.g., g4dn.xlarge).
Install Docker and NVIDIA drivers.
Build and run the Docker image as above.
Open port 8501 for web access.

B. HuggingFace Spaces

Create a new Space (Streamlit SDK).
Upload all files (including app.py, requirements.txt, src/, and assets/).
The app will auto-deploy.

Customization

Change the logo in assets/logo.png.
Adjust font or colors in src/utils.py.
Swap out the default HuggingFace model in the sidebar.

License

MIT

Contact

[Your Company Name] | [[email protected]]