Simon P. Couch

Thanks for all of your work here! Really excited about these additions. I think these PRs are in a good spot to head to `main` and will look towards a...

Thanks for the issue! I'm on board. :) Related to tidymodels/parsnip#760, and tidymodels/parsnip#765. My response for the analogous parsnip issues reflects where my thinking is at with this in bonsai...

Thanks for the issue, @seb-mueller. This sounds like it may be a good fit for the bonsai parsnip extension package, so I will go ahead and move this issue there....

Could you clarify which engine/model supported by bonsai this question is in reference to?

Ah, gotcha! We ought to add support for that lightgbm parameter but have not yet. This will be included in a future release! Note to self: reading through [docs](https://lightgbm.readthedocs.io/en/latest/Parameters.html#dataset-parameters), this...

Awesome, thanks! :)

Closed by #72.

Thanks for the issue. Training for a multi-class task with early stopping indeed works fine: ``` r library(bonsai) #> Loading required package: parsnip library(modeldata) data(penguins) lgb_model % set_engine("lightgbm", validation =...

See: > bonsai handles setting the `obj` argument to `lgb.train`, and will warn if it's passed No need to pass that argument!👍

Just wanted to drop a note here and let you know this hasn't fallen off my radar! I'm hoping to spend some time with our `multi_predict` methods and put together...