Github user jkbradley commented on the issue: https://github.com/apache/spark/pull/20825 I actually would prefer not to merge this change since it could blow up the size of the strings printed for some classification tasks with large numbers of labels. If people want to debug, they could trace through the tree manually. Alternatively, I'd be OK with adding an optional argument which tells toDebugString to include the stats.
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org