You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@spark.apache.org by "Sean R. Owen (Jira)" <ji...@apache.org> on 2019/11/10 19:21:00 UTC
[jira] [Resolved] (SPARK-29824) Missing persist on trainDataset in
ml.classification.GBTClassifier.train()
[ https://issues.apache.org/jira/browse/SPARK-29824?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Sean R. Owen resolved SPARK-29824.
----------------------------------
Resolution: Duplicate
> Missing persist on trainDataset in ml.classification.GBTClassifier.train()
> --------------------------------------------------------------------------
>
> Key: SPARK-29824
> URL: https://issues.apache.org/jira/browse/SPARK-29824
> Project: Spark
> Issue Type: Sub-task
> Components: ML
> Affects Versions: 2.4.3
> Reporter: Dong Wang
> Priority: Major
>
> The rdd trainDataset in ml.classification.GBTClassifier.train() is used by an action first and other actions in GradientBoostedTrees.run/runWithValidation, but it is not persisted, which will cause recomputation on trainDataset.
> {code:scala}
> override protected def train(
> dataset: Dataset[_]): GBTClassificationModel = instrumented { instr =>
> val categoricalFeatures: Map[Int, Int] =
> MetadataUtils.getCategoricalFeatures(dataset.schema($(featuresCol)))
> ...
> val numFeatures = trainDataset.first().features.size // first use trainDataset
> ...
> // trainDataset will be used by other actions in run methods.
> val (baseLearners, learnerWeights) = if (withValidation) {
> GradientBoostedTrees.runWithValidation(trainDataset, validationDataset, boostingStrategy,
> $(seed), $(featureSubsetStrategy))
> } else {
> GradientBoostedTrees.run(trainDataset, boostingStrategy, $(seed), $(featureSubsetStrategy))
> }
> {code}
> This issue is reported by our tool CacheCheck, which is used to dynamically detecting persist()/unpersist() api misuses.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org