AutoML ExperimentResult plain TrainerName instead of "ReplaceMissingValues=>OneHotHashEncoding=>Concatenate=>Unknown=>SdcaMaximumEntropyMulti=>Unknown"

dotnet / machinelearning

ML.NET is an open source and cross-platform machine learning framework for .NET.

https://dot.net/ml

MIT License

9.03k stars 1.89k forks source link

AutoML ExperimentResult plain TrainerName instead of "ReplaceMissingValues=>OneHotHashEncoding=>Concatenate=>Unknown=>SdcaMaximumEntropyMulti=>Unknown" #6541

Open rzechu opened 1 year ago

rzechu commented 1 year ago

System Information (please complete the following information):

OS & Version: Windows 11
ML.NET Version: 0.20
.NET Version: .NET 6.0

Describe the bug Result of Microsoft.ML.AutoML.MulticlassClassificationExperiment public override ExperimentResult Execute(IDataView trainData, string labelColumnName = "Label", string samplingKeyColumn = null, IEstimator preFeaturizer = null, IProgress<RunDetail> progressHandler = null)

Expected behavior Trainer name should be used trainer name not this "concatated long value"

Screenshots, Code, Sample Projects Previous valid (0.17)

New version (0.20)

LittleLittleCloud commented 1 year ago

We also have similar feedback on having a seperate property that hold trainer name only. Would adding an extra TrainerName in TrialResult which held algo name be enough?

rzechu commented 1 year ago

No matter what just need easy accesible algo name to describe results in returned ExperimentResults. Another property with plain algo name should be enough

tugbargn commented 11 months ago

I am having the same issue. Microsoft.ML version 3.0.0 Microsoft.Auto.ML version 0.21.0

bettwedder commented 8 months ago

I also have the same issue.

.NET 8.0. AutoML 0.21.1 Microsoft.ML version 3.0.1

tugbargn commented 1 month ago

I still have this issue. TrainerName is like this: FeaturizeText=>Concatenate=>LbfgsLogisticRegressionBinary But I'm expacting LbfgsLogisticRegressionBinary.