mlr-org/mlr3

benchmark_grid should not allow learners with the same ID

Opened this issue · 0 comments

tt = tsk("iris")
lrn1 = lrn("classif.rpart", id = "foo")
lrn2 = lrn("classif.rpart", id = "foo")

xs = list(lrn1, lrn2)

bg = benchmark_grid(learners = xs, tasks = tt, resampling = rsmp("holdout"))
br = benchmark(bg)

this should probably be illegal. also then add a test.