What's the recommended way to test local models?

Hi! I'm wondering what's the recommended way to run Olmes against new models running locally. I can spin up a vLLM/SGLang instance to serve my model, but it seems like the only way to configure a new model is by adding it to [models.py](https://github.com/allenai/olmes/blob/main/oe_eval/configs/models.py). Is that correct or is there a more seamless approach?