Simon P. Couch

Do you think this is in-scope for tidyclust, @EmilHvitfeldt?

There are actually quite a bit of these: https://github.com/tidymodels/parsnip/blob/45624583a7c6c4cfc7839c8dad3752f9ac5ff1ce/tests/testthat/_snaps/mlp_keras.md?plain=1#L1-L7 https://github.com/tidymodels/parsnip/blob/45624583a7c6c4cfc7839c8dad3752f9ac5ff1ce/tests/testthat/_snaps/mlp_nnet.md?plain=1#L1-L7 Let's do all of these in one swoop if we do remove. :)

Another: https://github.com/tidymodels/parsnip/blob/45624583a7c6c4cfc7839c8dad3752f9ac5ff1ce/tests/testthat/_snaps/nearest_neighbor_kknn.md?plain=1#L11-L15

Ultimately, `recipes::juice()` and `hardhat::recompose()` are what handles running recipes preprocessors via `fit.action_recipe()` -> `mold()` -> `run_mold()` -> `mold_recipe_default_process()`, and they can avoid `model.matrix()` while they do so because recipes don't...

In order of most to least preferred, my thoughts on approaches we could take for `fit()`ting workflows with formula preprocessors and sparse tibbles: 1) **Maybe overopinionated**: Warn on `add_formula()` with...

Notes from chatting with Emil on this: parsnip experiences the same issue with its `fit()` interface: the sparse data is passed to `model.matrix()`, which is somewhat slow but, even more...

Ah, interesting! Thanks for the issue. The xgboost engine itself will return the same predictions given that objective function, but tidymodels then only knows how to post-process predictions from the...

I dig it! If folks would find this pedagogically useful, I think this is surely within scope and would have a low maintenance burden. :)

Long time no see😝 We've got some good news here, though—custom probability thresholds and other postprocessing functionality is now available via [tailor](https://github.com/tidymodels/tailor)s, which can be added to workflows in the...

Thanks for the issue! A few notes from my first time wrapping my head around this [here](https://github.com/tidymodels/bonsai/issues/12#issuecomment-1117651731). What makes this more straightforward for engines like lightgbm or xgboost is that...