error with futuremice when n.core is not specified and exceeds m #533
Comments
|
Thanks,
Mice performs by default 5 imputations, which cannot be distributed in
`futuremice` over more than 5 cores. Hence the error. Either restrict the
number of cores or increase `m`.
I'd rather not make adjustments to the user specified call and override the
defaults as it may lull the user into a false sense of security. Now the
user is forced to make a choice.
…On Tue, 24 Jan 2023, 10:14 cdolladille, ***@***.***> wrote:
Hi all,
I'm pretty new to GitHub issues. I apologize if this is a duplicate or
already answered, but I could not find it fixed in another issue.
I have a 16 cores processor, when I try to run
imp <- futuremice(nhanes, maxit = 2, m = 2)
I have the following warning message
Number of cores not specified. Based on your machine a value of n.core =
15 is chosen; the imputations are distributed about equally over the cores.
Then this error
Error in (function (.x, .f, ..., .progress = FALSE) :
ℹ In index: 1.
Caused by error:
! Number of imputations (m) lower than 1.
When I explicitly set n.core to 15, I have the following message and no
error
In futuremice(nhanes, maxit = 2, m = 2, n.core = 15) :
The number of cores exceeds the number of imputations. The number of cores used is set equal to the number of imputations (m = 2 ).
My guess is that automatic setting of n.core should check m before.
Something like
if(machine.cores >m) {n.core <- m} else { n.core <- machine.cores - 1 }
Here is my sessionInfo()
R version 4.2.2 (2022-10-31 ucrt)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 22621)
Matrix products: default
locale:
[1] LC_COLLATE=French_France.utf8 LC_CTYPE=French_France.utf8 LC_MONETARY=French_France.utf8
[4] LC_NUMERIC=C LC_TIME=French_France.utf8
attached base packages:
[1] stats graphics grDevices utils datasets methods base
other attached packages:
[1] mice_3.15.0 forcats_0.5.2 stringr_1.5.0 dplyr_1.0.10 purrr_1.0.0 readr_2.1.3
[7] tidyr_1.2.1 tibble_3.1.8 ggplot2_3.4.0 tidyverse_1.3.2
loaded via a namespace (and not attached):
[1] TH.data_1.1-1 googledrive_2.0.0 colorspace_2.0-3 deldir_1.0-6
[5] ellipsis_0.3.2 class_7.3-20 rprojroot_2.0.3 htmlTable_2.4.1
[9] base64enc_0.1-3 fs_1.6.0 gld_2.6.6 rstudioapi_0.14
[13] proxy_0.4-27 listenv_0.9.0 furrr_0.3.1 MatrixModels_0.5-1
[17] fansi_1.0.3 mvtnorm_1.1-3 lubridate_1.9.0 xml2_1.3.3
[21] codetools_0.2-18 splines_4.2.2 rootSolve_1.8.2.3 knitr_1.41
[25] Formula_1.2-4 jsonlite_1.8.4 broom_1.0.2 cluster_2.1.4
[29] dbplyr_2.3.0 png_0.1-8 compiler_4.2.2 httr_1.4.4
[33] backports_1.4.1 assertthat_0.2.1 Matrix_1.5-3 fastmap_1.1.0
[37] gargle_1.2.1 cli_3.5.0 htmltools_0.5.4 quantreg_5.94
[41] tools_4.2.2 gtable_0.3.1 glue_1.6.2 lmom_2.9
[45] Rcpp_1.0.9 cellranger_1.1.0 vctrs_0.5.1 nlme_3.1-160
[49] xfun_0.36 globals_0.16.2 rvest_1.0.3 timechange_0.2.0
[53] lifecycle_1.0.3 googlesheets4_1.0.1 polspline_1.1.22 future_1.30.0
[57] zoo_1.8-11 MASS_7.3-58.1 scales_1.2.1 hms_1.1.2
[61] sandwich_3.0-2 parallel_4.2.2 expm_0.999-7 SparseM_1.81
[65] RColorBrewer_1.1-3 yaml_2.3.7 Exact_3.2 gridExtra_2.3
[69] rms_6.4-1 rpart_4.1.19 latticeExtra_0.6-30 stringi_1.7.8
[73] e1071_1.7-12 checkmate_2.1.0 boot_1.3-28 rlang_1.0.6
[77] pkgconfig_2.0.3 evaluate_0.20 lattice_0.20-45 htmlwidgets_1.6.1
[81] tidyselect_1.2.0 here_1.0.1 parallelly_1.34.0 magrittr_2.0.3
[85] R6_2.5.1 DescTools_0.99.47 generics_0.1.3 Hmisc_4.7-2
[89] multcomp_1.4-20 DBI_1.1.3 pillar_1.8.1 haven_2.5.1
[93] foreign_0.8-83 withr_2.5.0 survival_3.4-0 nnet_7.3-18
[97] modelr_0.1.10 crayon_1.5.2 interp_1.1-3 utf8_1.2.2
[101] tzdb_0.3.0 rmarkdown_2.20 jpeg_0.1-10 grid_4.2.2
[105] readxl_1.4.1 data.table_1.14.6 reprex_2.0.2 digest_0.6.31
[109] munsell_0.5.0
Thank you!
Charles
—
Reply to this email directly, view it on GitHub
<#533>, or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ABT2AKCIKA5BBNEZMZTQTQTWT6MRFANCNFSM6AAAAAAUE2QBS4>
.
You are receiving this because you are subscribed to this thread.Message
ID: ***@***.***>
|
|
Thank you for your reply, I acknowledge this might not be a frequent use case. |
|
I agree with @gerkovink, but I also think that we can easily tackle the issue by always falling back to the highest number of cores that makes sense given the number of imputed datasets, and informing the user that we did this by printing a message. |
Merged
|
@thomvolker Thanks. Closing because this is solved by #550. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Hi all,
I'm pretty new to GitHub issues. I apologize if this is a duplicate or already answered, but I could not find it fixed in another issue.
I have a 16 cores processor, when I try to run
imp <- futuremice(nhanes, maxit = 2, m = 2)I have the following warning message
Number of cores not specified. Based on your machine a value of n.core = 15 is chosen; the imputations are distributed about equally over the cores.Then this error
When I explicitly set n.core to 15, I have the following message and no error
My guess is that automatic setting of n.core should check m before. Something like
if(machine.cores >m) {n.core <- m} else { n.core <- machine.cores - 1 }Here is my sessionInfo()
Thank you!
Charles
The text was updated successfully, but these errors were encountered: