srowen commented on issue #25576: [SPARK-28866][ML] Persist item factors RDD when checkpointing in ALS URL: https://github.com/apache/spark/pull/25576#issuecomment-525740768 This may be a dumb question, but I thought we only 'materialized' persisted RDDs, and they aren't persisted here. Is the .count() after .checkpoint() necessary in the implicit case? I didn't realize this interacted with checkpointing? well, it won't matter after this change, because it is needed to materialize the persisted RDD. Oh, nevermind my comment about persisting. I see that it does, just looked right past it at the end of the line. Would it make sense to persist/unpersist the user factors too then in the non-implicit case? or we're saying that its lineage is always short, as it always depends on the previous item factors?
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
