In this repo, i am able to chat with multiple pdfs. I have used streamlit at frontend. Used FAISS (Facebook AI Similarity Search), a library which is used for similarity search of vectors. This chatbot is trained in such a way that it will not provide random output, if answer is not available in context. It will just say i dont know.
- Python = 3.10
- streamlit = 1.32.2
- google-generativeai = 0.4.1
- langchain = 0.1.13
- faiss-cpu = 1.8.0
- First of all, needs to have conda (Anaconda) -package and env management tool.
- Once conda is installed create new virtual env using command conda create -p {env_name} python={python_version} e.g. conda create -p venv python=3.10
- Activate conda env using command conda activate {env}
- Create .env file as per sample env .For this you will need to create google gemini api key.
- Install packges using command pip install -r requirements.txt
- Run application using command streamlit run app.py