Implementation of Query-Specific Graph Neural Network

Prepare

1. Install environment acorrding to requirement.txt
2. Download NV-Embed_v2, we suggest download checkpoint from huggingface: https://huggingface.co/nvidia/NV-Embed-v2
3. Prepare your LLM, use API or download and delopy locally
4. All the downloaded models should be saved under directory named "model_download"

Run

Please refer to run_script.txt