Remove ConcurrencyFactor from IHostEnvironment

[artidoro]

Fixes #2051.

In this PR I remove ConcurrencyFactor from IHostEnvironment and interfaces and classes deriving/implementing it.

I had to skip a RandomPredictor as I am not sure how I can require a single threaded behavior without the ConcurrencyFactor for this specific trainer.

[artidoro]

Note that the first and second commit contain the most important logical changes the rest are fixes to tests and samples to make sure they compile.

[codecov]

Codecov Report

Merging #2846 into master will decrease coverage by <.01%.
The diff coverage is 94.92%.

@@            Coverage Diff             @@
##           master    #2846      +/-   ##
==========================================
- Coverage   71.69%   71.69%   -0.01%     
==========================================
  Files         811      811              
  Lines      142546   142494      -52     
  Branches    16125    16112      -13     
==========================================
- Hits       102193   102155      -38     
+ Misses      35924    35918       -6     
+ Partials     4429     4421       -8

Flag	Coverage Δ
#Debug	71.69% <94.92%> (-0.01%)	⬇️
#production	67.91% <92%> (-0.01%)	⬇️
#test	85.91% <95.57%> (ø)	⬆️

Impacted Files	Coverage Δ
src/Microsoft.ML.Core/Data/IHostEnvironment.cs	95.12% <ø> (ø)	⬆️
...soft.ML.TestFramework/DataPipe/TestDataPipeBase.cs	73.67% <ø> (-0.09%)	⬇️
.../Standard/LogisticRegression/LbfgsPredictorBase.cs	71.26% <ø> (+0.96%)	⬆️
...t/Microsoft.ML.Benchmarks/PredictionEngineBench.cs	0% <0%> (ø)	⬆️
...c/Microsoft.ML.FastTree/SumupPerformanceCommand.cs	0% <0%> (ø)	⬆️
src/Microsoft.ML.FastTree/GamTrainer.cs	89.83% <0%> (-0.22%)	⬇️
...st/Microsoft.ML.Tests/Scenarios/TensorflowTests.cs	100% <100%> (ø)	⬆️
...icrosoft.ML.Functional.Tests/DataTransformation.cs	100% <100%> (ø)	⬆️
...st/Microsoft.ML.Tests/Scenarios/ClusteringTests.cs	100% <100%> (ø)	⬆️
...osoft.ML.TimeSeries.Tests/TimeSeriesStaticTests.cs	98.56% <100%> (ø)	⬆️
... and 48 more

[TomFinley]

I had to skip a RandomPredictor as I am not sure how I can require a single threaded behavior without the ConcurrencyFactor for this specific trainer.

Actually that's interesting. We need to hide this for now, just as a we had to hide GenerateNumberTransform. (Not necessarily part of your PR, but if you wanted to do it it wouldn't be that hard, we'd just need to hide RandomModelParameters and RandomTrainer.) RandomModelParameters claims it implements the internal interface IValueMapperDist, which means that it is claiming (via our infrastructure) that it is possible to express as an IRowToRowMapper, but it can't since the outputs are not independent.

Edit: To address this problem I've introduced issue #2848 and PR #2849. It had somehow escaped my notice that we had a random model just lying around claiming to be a value mapper. (And thereby introducing the bug you observed.) That's obviously not going to fly... #Resolved

[TomFinley]

[Ivanidzo4ka]

        //      The coordinates of centroid 0 are: (26, 6, 1)

I would double check output in samples.
Or just set numThreads in KMeans to 1. #Resolved

---

Refers to: docs/samples/Microsoft.ML.Samples/Dynamic/Trainers/Clustering/KMeans.cs:45 in a8c7541. [](commit_id = a8c7541, deletion_comment = False)

[Ivanidzo4ka]

        //      The coordinates of centroid 0 are: (26, 6, 1)

Please validate what output in samples are same across all files in this PR

---

In reply to: 469970156 [](ancestors = 469970156)

---

Refers to: docs/samples/Microsoft.ML.Samples/Dynamic/Trainers/Clustering/KMeans.cs:45 in a8c7541. [](commit_id = a8c7541, deletion_comment = False)

[wschin]

Please explain what seed and verbose are.