for reference, I am using argoworkflows to schedule training of yolov5 (using yolov5's docker image) and would like to use yolov5's built-in clearml logger. Inside of the container I have access to the access tokens and secret through my own api, but need to properly initialize clearml before scheduling training.