How can we improve Microsoft Azure Machine Learning?

Monitor hyperparameter tuning progress, abort and provide best values found so far

I used "Tuned Model Hyperparameters" and it's taking longer than I expected (hours), with no indication of remaining progress. I hesitate to stop it because I would like to have information about the best solution found so far. The log only shows when each training started and stopped, but doesn't show the parameters tested at each attempt and the results.

It would also be useful to have a graphical view of the progress: Plot with X = iteration, Y = Each of the possible metrics (AUC, F-Score,...). This graphical view should be available as parameter tuning progresses, to help decide if tuning should continue or not (if the metrics stay flat for several iterations, it's not worth continuing).

In any case, if tuning is aborted, we should have access to the best solution found.

1 vote
Vote
Sign in
Check!
(thinking…)
Reset
or sign in with
  • facebook
  • google
    Password icon
    I agree to the terms of service
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    Christos Karras shared this idea  ·   ·  Flag idea as inappropriate…  ·  Admin →

    0 comments

    Sign in
    Check!
    (thinking…)
    Reset
    or sign in with
    • facebook
    • google
      Password icon
      I agree to the terms of service
      Signed in as (Sign out)
      Submitting...

      Feedback and Knowledge Base