Arxiv-RAG: a retrieval-augmented generation (RAG) system that leverages Mistral-7B (& LLaMA-2) to search and summarize research papers from arXiv. It uses ChromaDB for efficient storage and retrieval.
app_file: streamlitapp.py
short_description: ArXiv research retrieval system
Features:
- Dynamic arXiv Paper Retrieval – Fetch research papers across multiple real-time categories.
- AI-Powered Summaries – Leverages Mistral-7B / LLaMA-2 for intelligent summarization.
- Vector-Based Search – Uses ChromaDB for fast and efficient research paper retrieval.
- Interactive Web UI – Built with Streamlit for an intuitive search experience.
- Deployment & Automation – Hosted on Railway.app with GitHub CI/CD integration.
git clone https://github.com/triiJU/arxiv_search_retrieval.git
cd arxiv_search_retrieval
conda create -n arxiv-rag python=3.10 -y
conda activate arxiv-rag
pip install -r requirements.txt
python fetch_arxiv.py
streamlit run streamlitapp.py
python query_chroma.py
| Query | Expected Output |
|---|---|
| "Transformer models" | List of papers + summaries |
| "Reinforcement learning" | Papers + AI-generated insights |
pip check
pip install --force-reinstall -r requirements.txt
rm -rf chroma_db/ && python fetch_arxiv.py
MIT License.