You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Calculating the partials of a model is currently the task that takes longest for me in dbarts (0.9-23 CRAN version). In 0.9-22 this would take a few hours in a model of about 7000 observations and 20 variables. Testing now for 0.9-23 if some of the changes under the hood changed that.
Is it possible to implement a parallelisation of the pdbart function? Maybe variables can be calculated in parallel? It seems to be using only a single CPU (out of the 18 available) right now (both in 0.9-22 and 0.9-23).
The text was updated successfully, but these errors were encountered:
Calculating the partials of a model is currently the task that takes longest for me in dbarts (0.9-23 CRAN version). In 0.9-22 this would take a few hours in a model of about 7000 observations and 20 variables. Testing now for 0.9-23 if some of the changes under the hood changed that.
Is it possible to implement a parallelisation of the pdbart function? Maybe variables can be calculated in parallel? It seems to be using only a single CPU (out of the 18 available) right now (both in 0.9-22 and 0.9-23).
The text was updated successfully, but these errors were encountered: