RAG for documents: containing text paired with metadata

how to run

Step 1: environment setup

conda env create -f mmrag_env.yaml

conda activate mmrag

if encounting with cuda version problem, you can alse try this:

cd Multimodal-Rag

pip install torch torchvision torchaudio  # install your correct torch verion

pip install -r requirements.txt

Step 2: configurate

• fill in environment variables scripts\run.sh

Step 3:

bash script/run.sh

To change parameters, please change in configs.yml.