You can try out some of our schemas with Ollama if you want: https://github.com/vlm-run/vlmrun-hub (instructions in Readme)