Evaluations

Run and monitor model evaluations