search

JuliaAI / MLJTuning.jl

Hyperparameter optimization algorithms for use in the MLJ machine learning framework
MIT License
66 stars 12 forks source link

Create option to write `CompactPerformanceEvaluation` objects to history #215

Closed ablaom closed 2 months ago

ablaom commented 2 months ago

... and make that the default.

Requires:

and closes https://github.com/alan-turing-institute/MLJ.jl/issues/1105

Technically breaks a change made a few patch releases ago, but doesn't actually effect the published API.

After this PR, you can specify compact_history=false in TunedModel(...) to get full PerformanceEvaluation objects written to the history. Otherwise these are CompactPerformanceEvaluation objects.

codecov[bot] commented 2 months ago

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Project coverage is 87.53%. Comparing base (c597f27) to head (5143292).

Additional details and impacted files ```diff @@ Coverage Diff @@ ## dev #215 +/- ## ======================================= Coverage 87.53% 87.53% ======================================= Files 13 13 Lines 666 666 ======================================= Hits 583 583 Misses 83 83 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

ablaom commented 2 months ago

@dpaetzel Any chance you could review this?

dpaetzel commented 2 months ago

I looked over the change set and didn't find anything out of place. Since the build passes the tests and those cover the new behaviour I guess this looks good to merge? :slightly_smiling_face:

ablaom commented 2 months ago

Thanks @dpaetzel for the prompt review 🙏🏾