[ https://issues.apache.org/jira/browse/SPARK-16404?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15365227#comment-15365227 ]
Seth Hendrickson commented on SPARK-16404: ------------------------------------------ cc [~dbtsai] I looked in to using the @transient tag, but this prevents the coefficients from being serialized and broadcast to the executors at all, resulting in a {{NullPointerException}}. I am not sure of a way around this. I can submit a patch utilizing the same strategy as in LoR later this week. > LeastSquaresAggregator in Linear Regression serializes unnecessary data > ----------------------------------------------------------------------- > > Key: SPARK-16404 > URL: https://issues.apache.org/jira/browse/SPARK-16404 > Project: Spark > Issue Type: Improvement > Components: ML > Reporter: Seth Hendrickson > > This is basically the same issue as > [SPARK-16008|https://issues.apache.org/jira/browse/SPARK-16008], but for > linear regression, where {{coefficients}} and {{featuresStd}} are > unnecessarily serialized between stages. -- This message was sent by Atlassian JIRA (v6.3.4#6332) --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org