How do I store and then load a model after training?

[bluelight773]

These are questions: After I've run training with something like the following per the code on the readme.md:

trainer.train(
    epochs=1,
    steps_per_epoch=500,
    batch_size=16,
    monitored_metrics={'val_PSNR_Y': 'max'}
)

How do I store the model? And how do I afterwards load it and use it for prediction? Also, do you have any guidelines for how long training should work, eg, if p3.2xlarge instance is used or some other recommended AWS instance?

[cfrancesco]

I think I use a p2.xlarge for training (about a day of training I believe). The model is automatically saved in the weights directory.