Problem SparkTrials is an extension of Hyperopt, which allows runs to be distributed to Spark workers. When you start an MLflow run with nested=True in the