This tool looks super interesting! I had a question- You mentioned the ability to run tests to ensure the fine-tuned model behaves as expected. What evaluation

This tool looks super interesting! I had a question- You mentioned the ability to run tests to ensure the fine-tuned model behaves as expected. What evaluation metrics are built in, and how customizable is the evaluation pipeline? Can I easily add my own metrics?