3 ms·Yes, google "llama-server split model multiple GPU" for some concrete examples.by walrus01 2mo agoYes, google "llama-server split model multiple GPU" for some concrete examples.