Many students were asking how to parallelize grid search

[cimentadaj]

• Figure out which interface uses tidymodels

• Can we just define the parallel interface outside the tidyflow, together with the cores? Or do we have to pass anything to the tidyflow --> tidymodels as arguments?

• Make sure you can test that the models are being ran in parallel. Don't just run the model without making sure.

• One approach is to allow each cross-validated set from tidymodels to be verbose and if many different CV sets are pushing text without the sequential order, they're being ran in parallel.

[cimentadaj]

We're overthinking this. Following the vignette at tidymodels (https://tune.tidymodels.org/articles/extras/optimizations.html), we just have to specify the parallel backend before running tidyflow. We can add a small vignette with parallelization instructions.

[cimentadaj]

In light of tidymodels/tune#275, it would be useful if we added a test to check that seeds replicate across parallelization. I believe this should work out of the box since the seed is run before any step in tidyflow but we should make sure this is the case in the tests.