Skip to content

Auto train pipeline Bug #5437

Closed
Closed

Description

System information

  • OS version/distro: Windows server 2012r2 - windows 10
  • .NET Version: .net core 3.1
  • Ml.net Version: Last Stable Version.

Issue

  • What did you do?: Set a auto Training pipeline using API, set a long MaxExperimentTimeInSeconds (For example 600) for small dataset.
  • What happened?: Program hanged without any response.
  • What did you expect?: the program should stop training after 600 seconds. and show me the result.
  • Am I Tested it with shorter time?: Yes! I Tested it with 120Seconds and 100 Seconds. Its work currectly!

Source code / logs

var experimentSettings = new MulticlassExperimentSettings()
            {
                MaxExperimentTimeInSeconds = 600,
                CacheDirectory = new DirectoryInfo("cache"),
                CacheBeforeTrainer = CacheBeforeTrainer.Auto,
                OptimizingMetric = MulticlassClassificationMetric.MicroAccuracy,
            };
        var experiment = mlContext.Auto().CreateMulticlassClassificationExperiment(experimentSettings);

        var crossValidationExperimentResult = experiment.Execute(
            trainData: mappedInputData,
            labelColumnName: "Label",
            progressHandler: new Progress<RunDetail<MulticlassClassificationMetrics>>()
        );

toplearn

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Assignees

Labels

AutoML.NETAutomating various steps of the machine learning processP2Priority of the issue for triage purpose: Needs to be fixed at some point.

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions