MultiModal RAG in LangChain

Repo for code of a DS@M medium blog. Blog: Building a multimodal Retrieval Augmented Generation (RAG) system: A technical overview | by Yash Khasbage | Data Science at Microsoft | Feb, 2025 | Medium

Running the code

1. Keep your testing PDFs in a folder papers in the root.
2. Create a deployment of GPT-4o from the azure portal.
3. Fill up the azure_oai.env environment file with the credentials of the 4o model deployed in Azure.
4. Create a python venv

python -m venv .venv
# activate the venv
# Linux/MacOS
source .venv/bin/activate
# Windows
.venv/Scripts/activate

5. Install the dependencies

pip install -r requirements.txt

6. Run the RAG and observe the output on terminal

python mm_rag.py