Max kuhn
banner
topepo.bsky.social
Max kuhn
@topepo.bsky.social
Writing modeling packages at @posit.co (née RStudio). Opinions are my own. https://max-kuhn.org/
For some models that are known to tidymodels, you can trick it into just working:

gist.github.com/topepo/0d6f7...
November 5, 2025 at 7:33 PM
You could create a small function/class/package that encapsulates the object and allows for predictions. The pain is making it work with tidymodels, especially if that involves some novel model type.
Avoid estimation for a model in tidymodels
Avoid estimation for a model in tidymodels. GitHub Gist: instantly share code, notes, and snippets.
gist.github.com
November 5, 2025 at 7:32 PM
They also have simple vector interfaces so if you’re not interested in the tidy API’s that’s not a big deal
October 30, 2025 at 2:55 PM
Yardstick can use case ways for all of its functions, including sensitivity, specificity, and ROC curves.
October 30, 2025 at 10:33 AM
No. I generate the random seeds based on the wall clock (to better emulate true randomness).
October 23, 2025 at 4:54 PM
We had a discussion in my group about this a while back. Use actual random numbers as seeds or common values/sequences, such as 1, 123, 999, 321, etc.

I think to their dismay, I always had 3-4 digit random values, leading to this slide annotation:
October 22, 2025 at 9:34 PM
Emil!
October 18, 2025 at 7:36 PM
... and after
October 18, 2025 at 7:34 PM
I do the same thing over and over and over again.
October 16, 2025 at 11:28 PM
The kangaroos are nice, but those handles are 💯
October 7, 2025 at 1:22 PM
I hesitate to inflict bag envy on this whole site but here they are. Thanks to @visnut.bsky.social and @robjhyndman.com
October 7, 2025 at 1:02 PM
The new version of tune (out before posit::conf🤞) has support for both.

My testing so far shows that future(_lappy) and mirai(_map) are very similar in speed-ups most of the time. The times when they differ split in either direction, so I haven’t seen any clear signal.
August 27, 2025 at 11:29 PM