Uploaded image for project: 'Data Management'
  1. Data Management
  2. DM-11169

Store more detailed history in GLMM

    Details

    • Type: Story
    • Status: Done
    • Resolution: Done
    • Fix Version/s: None
    • Component/s: meas_deblender
    • Labels:
      None

      Description

      When an NMF minimization using the GLMM algorithm fails to converge it is often useful to have more information from the history of the traceback to understand why convergence failed. This ticket will modify the Traceback object to properly store the X, Z, U matrices as well as the f and g steps used in each iteration.

        Attachments

          Activity

          fred3m Fred Moolekamp created issue -
          fred3m Fred Moolekamp made changes -
          Field Original Value New Value
          Epic Link DM-10353 [ 32070 ]
          Hide
          fred3m Fred Moolekamp added a comment -

          Hi Peter,
          I updated the proxmin repository to store more errors and make it easier to retrieve parameter and error histories. This should make it easier to chase down errors in convergence. This has been developed on the new new_traceback branch in proxmin. Have a look and let me know if you have any questions.

          -Fred

          Show
          fred3m Fred Moolekamp added a comment - Hi Peter, I updated the proxmin repository to store more errors and make it easier to retrieve parameter and error histories. This should make it easier to chase down errors in convergence. This has been developed on the new new_traceback branch in proxmin . Have a look and let me know if you have any questions. -Fred
          fred3m Fred Moolekamp made changes -
          Reviewers Peter Melchior [ pmelchior ]
          Status To Do [ 10001 ] In Review [ 10004 ]
          Hide
          pmelchior Peter Melchior added a comment -

          I put several design suggestions into a comment for the pull request. None of them are onerous but will make the structure more easily useable. Please make those changes, and then we can use the new structure consistently for all algorithms.

          Show
          pmelchior Peter Melchior added a comment - I put several design suggestions into a comment for the pull request. None of them are onerous but will make the structure more easily useable. Please make those changes, and then we can use the new structure consistently for all algorithms.
          fred3m Fred Moolekamp made changes -
          Story Points 2 4
          Hide
          pmelchior Peter Melchior added a comment -

          Modifications made and merged into master

          Show
          pmelchior Peter Melchior added a comment - Modifications made and merged into master
          pmelchior Peter Melchior made changes -
          Status In Review [ 10004 ] Reviewed [ 10101 ]
          fred3m Fred Moolekamp made changes -
          Resolution Done [ 10000 ]
          Status Reviewed [ 10101 ] Done [ 10002 ]

            People

            • Assignee:
              fred3m Fred Moolekamp
              Reporter:
              fred3m Fred Moolekamp
              Reviewers:
              Peter Melchior
              Watchers:
              Fred Moolekamp, Peter Melchior
            • Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved:

                Summary Panel