Usage example causes session to abort

[Oliver-CISC]

Running the following example from the README.md causes the session to abort:

library(forester)

data("titanic", package = 'DALEX')
best_model <- forester(data = titanic, target = "survived", type = "classification", metric = "precision", tune = FALSE)

Forester proceeds with the first steps creating Ranger, XgBoost, and Catboost models then aborts the session.

My sessionInfo:

R version 4.1.1 (2021-08-10)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 19043)

Matrix products: default

locale:
[1] LC_COLLATE=English_United Kingdom.1252 LC_CTYPE=English_United Kingdom.1252
[3] LC_MONETARY=English_United Kingdom.1252 LC_NUMERIC=C
[5] LC_TIME=English_United Kingdom.1252

attached base packages:
[1] stats graphics grDevices utils datasets methods base

loaded via a namespace (and not attached):
[1] bit_4.0.4 odbc_1.3.2 compiler_4.1.1 ellipsis_0.3.2 generics_0.1.0 hms_1.1.0
[7] DBI_1.1.1 tools_4.1.1 Rcpp_1.0.7 bit64_4.0.5 lubridate_1.7.10 vctrs_0.3.8
[13] blob_1.2.2 lifecycle_1.0.0 pkgconfig_2.0.3 rlang_0.4.11

[Oliver-CISC]

Further investigation points to the lightgbm algorithm as the culprit. All other models can be constructed using the make_ function, but make_lightgbm() immediately crashes the session.

Stepping through the code reveals lightgbm::lightgbm() to be the culprit.

[Oliver-CISC]

I have found a workaround to this issue:

If you are experiencing session crashes when using forester it is likely due to the lightgbm algorithm. This can be checked by trying the individual make_ functions within forester.

The first time you try to create a forester() model it will always crash your session. In the re-started session, invoke lgb.unloader(wipe = TRUE) then proceed to run the forester() function. This should now work as expected. This problem is documented here.

[lhthien09]

Good morning,

Thank you for your information. We found out that the problem happened with LightGBM, as you pointed out. And in this case, LightGBM crashes with the default example sometimes; you may try this example from: https://lightgbm.readthedocs.io/en/latest/R/index.html

library(lightgbm)
data(agaricus.train, package='lightgbm')
train <- agaricus.train
dtrain <- lgb.Dataset(train$data, label = train$label)
model <- lgb.cv(
params = list(
objective = "regression"
, metric = "l2"
)
, data = dtrain
)

==> Solution is mentioned at this page: microsoft/LightGBM#3445 by restarting R session: .rs.restartR() and {forester} works.

[lhthien09]

Please feel free to report issues with {forester}, thank you once again for reporting this issue.

[Oliver-CISC]

Thanks for the solution - it works reproducibly on my system!