spark-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From jkbradley <>
Subject [GitHub] spark pull request: [SPARK-7770] [ML] GBT validationTol change to ...
Date Wed, 23 Sep 2015 22:04:05 GMT
Github user jkbradley commented on a diff in the pull request:
    --- Diff: mllib/src/main/scala/org/apache/spark/mllib/tree/GradientBoostedTrees.scala
    @@ -262,7 +263,8 @@ object GradientBoostedTrees extends Logging {
               validationInput, validatePredError, baseLearnerWeights(m), baseLearners(m),
             val currentValidateError = validatePredError.values.mean()
    -        if (bestValidateError - currentValidateError < validationTol) {
    +        if (bestValidateError - currentValidateError < validationTol * Math.max(
    +          currentPredError, 0.01)) {
    --- End diff --
    I realized that it may make more sense to use currentValidateError instead of currentPredError
here.  That's what users would expect from "relative tolerance."  Could you please update
this + the doc?  Thanks.

If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at or file a JIRA ticket
with INFRA.

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message