Applying last_fit function in tidy models

If I understood correctly, to apply last_fit function in tidy package, I need to have a split object that created using rsample::initial_split().

However in a situation that I have separate training data and test data at the very beginning, I don't want to use initial_split function to split the data into training and testing.

Since I cannot create a split object, Couldnt I use last_fit function?

If you want to create an rsplit object from existing testing and training sets, you can use make_splits():

library(rsample)
library(dplyr)
#> 
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#> 
#>     filter, lag
#> The following objects are masked from 'package:base':
#> 
#>     intersect, setdiff, setequal, union

data(cells, package = "modeldata")

make_splits(
  cells %>% filter(case == "Train"),
  cells %>% filter(case == "Test")
)
#> <Analysis/Assess/Total>
#> <1009/1010/2019>

^{Created on 2022-01-16 by the reprex package (v2.0.1)}

Alternatively you can not use last_fit() and manually fit() on the training and predict() on the testing set.