When running the estimate_shift_configuration function with multiple cores the profiles element in the output, where the different configurations and their scores should be, comes back empty. This doesn't happen when using a single core. Below is an example of how I'm getting this issue with example data:
library(l1ou)
data(lizard.tree, lizard.traits)
lizard <- adjust_data(lizard.tree, lizard.traits[,1])
# the new tree is normalized: each tip at distance 1 from the root.
# new Y: matrix of size 100 x 1
eModel <- estimate_shift_configuration(lizard$tree, lizard$Y)
# Starting first LASSO (alpha=0) to find a list of candidate configurations.
# Starting second LASSO (alpha= 0.61 ) for another list of candidates.
str(eModel$profile)
# List of 2
# $ scores : num [1:2399] 16.8 18.3 18.7 19.3 19.9 ...
# $ configurations:List of 2399
# ..$ : num [1:8] 14 32 55 74 77 98 118 164
# ..$ : num [1:7] 14 32 55 77 98 118 164
# ..$ : num [1:6] 14 32 77 98 118 164
# ..$ : num [1:9] 14 32 37 55 74 77 98 118 164
# [etc...]
Hello!
When running the
estimate_shift_configuration
function with multiple cores the profiles element in the output, where the different configurations and their scores should be, comes back empty. This doesn't happen when using a single core. Below is an example of how I'm getting this issue with example data:Now multithreaded, produces an empty profile
I'm running R v. 4.22 and l1ou v. 1.43.
Thanks!