Nested CV with vfolds outer and simple-split inner?

The problem

I would like to set up a nested CV where the outer loop is 5-fold CV and the inner loop is a simple train-validate split. Specifically, I'm trying to do a nested CV where the splits are 60-20-20 train-validate-test, and it loops five times over the mutually exclusive test sets.

It seems the way to do this is to use nested_cv(mtcars, outside = vfold_cv(v = 5), inside = validation_split(prop = .75)); however, validation_split() has been deprecated.

Is there a better way to do this type of split with the new three-way split functions?

Thanks!

Reproducible example

nested_cv(mtcars, outside = vfold_cv(v = 5), inside = validation_split(prop = .75))

Start with the nested_cv(), vfold_cv(), validation_split(), and newer three-way split functions named in the issue. Compare how the outer and inner resampling objects represent the requested 60-20-20 train-validation-test arrangement, then verify the resulting splits across all five outer folds. Done means the supported approach is clear and reproducible.