Different ways of logging model

Hello, Lightning newbie here. I am confused by the different ways of logging models in Pytorch.

I am using Lightning with mlflow so there is an option to pass log_model when you inherit from MLFlowLogger. You can also define log_models in mlflow.pytorch.autolog. Then finally, you can also call mlflow.pytorch.log_model separately.

Can someone please provide guidance on different ways of logging models.