Uploaded image for project: 'Spark'
  1. Spark
  2. SPARK-3382

GradientDescent convergence tolerance

    XMLWordPrintableJSON

Details

    • Improvement
    • Status: Resolved
    • Minor
    • Resolution: Fixed
    • 1.1.0
    • 1.5.0
    • MLlib
    • None

    Description

      GradientDescent should support a convergence tolerance setting. In general, for optimization, convergence tolerance should be preferred over a limit on the number of iterations since it is a somewhat data-adaptive or data-specific convergence criterion.

      Attachments

        Issue Links

          is related to

          New Feature - A new feature of the product, which has yet to be developed. SPARK-1503 Implement Nesterov's accelerated first-order method

          • Major - Major loss of function.
          • Resolved
          relates to

          New Feature - A new feature of the product, which has yet to be developed. SPARK-10194 SGD algorithms need convergenceTol parameter in Python

          • Major - Major loss of function.
          • Resolved
          links to

          Pull request #3636 [Github] Pull Request #3636 (Lewuathe)

          Activity

            People

              lewuathe Kai
              josephkb Joseph K. Bradley
              Votes:
              0 Vote for this issue
              Watchers:
              4 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: