Closed Aariq closed 3 years ago
another possible speed enhancement would be adding support for models created by bam()
. Allow passing cluster
argument to predict()
.
since I've decided to use the gratia
package for most plots (plotting smooths directly rather than predicted model values), this is no longer very important.
This custom wrapper to predict() is quite slow because its running predict multiple times in a for loop. It would maybe be faster if I replaced the for loop with
map()
or possibly just created one big matrix to pass to predict once.