For a 0.8.6 release

[ablaom]

#193

Release notes:

• Add logger option to TunedModel wrapper, for logging internal model evaluations to an ML tracking platform. Default should be nothing for no logging (Adding loggers into TunedModels #193). The logger must support asynchronous messaging if TunedModel has acceleration != CPU1(). Such support is planned for MLflow integration via MLJFlow.jl; see Implementing channel control for experiment creation MLJFlow.jl#41 Adding loggers into TunedModels #193

cc @pebeto

[codecov]

Codecov Report

Attention: Patch coverage is 90.00000% with 1 lines in your changes are missing coverage. Please review.

Project coverage is 87.55%. Comparing base (1dd32b1) to head (2294da4).
Report is 33 commits behind head on master.

Files	Patch %	Lines
src/tuned_models.jl	90.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #218      +/-   ##
==========================================
+ Coverage   87.53%   87.55%   +0.01%     
==========================================
  Files          13       13              
  Lines         666      667       +1     
==========================================
+ Hits          583      584       +1     
  Misses         83       83              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.