Resources, including ML compute instances and ML storage volumes, that are configured for model training.