When would you use learn.save() vs learn.export()?

I have created a learner object but I want to shut down my instance to start one with a GPU. How would I save my learner object so that I can reinstantiate it in a new instance to start training? learn.save or learn.export?

1 Like


I think you have to use learn.export()when you want to export your model and use it in another code and another data, where you only want to load your model and use it.
If I understand what you are doing, you want to run again the same code you created the model and train with the same data. and only want to save your model and start from where you stopped. so I think that learn.save() make more sense.