Closed drozzy closed 3 years ago
Yeah, the test data isn't being used currently, so I should probably remove it as it's just confusing.
It would be nice if there there was a method called test
that has to be called manually by the user, as in the pytorch-lightning:
https://pytorch-lightning.readthedocs.io/en/stable/test_set.html
Just curious, I was reading the docs, and it says that I can pass the test dataset as part of the tuple: https://lorenzoh.github.io/FluxTraining.jl/dev/docstrings/FluxTraining.Learner.html
(traindata, valdata, [testdata])
But when I did that, I didn't notice any different command line outputs. Is the trainer supposed to run inference on test data automatically, or do I need to call it myself?
Thanks!