display training progress about tabnet HOT 5 CLOSED

if you are using tabnet_fit you could pass the verbose argument directly to the function.

from tabnet.

just noticed that it's there when you use the tidymodels API (before I was executing tabnet_fit) :-)

from tabnet.

as I'm at it - the whole issue is not that important :-) - may I suggest that if doing cross-validation, we display the fold we're working on?

from tabnet.

This can be done with tidymodels, see eg: https://tune.tidymodels.org/reference/control_grid.html
For displaying a progressbar when training you can pass verbose=TRUE when calling set_engine. Like this:

mod <- tabnet(epochs = 50, batch_size = 128) %>%
  set_engine("torch", verbose = TRUE) %>%
  set_mode("classification")

from tabnet.

thanks! (this issue originally referred to the code in the README, where there's no set_engine ... but in reality, I'd use the code from your answer anyway (((or even the workflow approach))))

from tabnet.