So I went ahead and tried running the example script with "A CHRISTMAS CAROL" using the "meta-llama-3.1-8b-instruct" and "text-embedding-nomic-embed-text-v1.5" models locally. How long should it take to extract the subgraphs with this kind of setup?