I am patching non-deterministic test behavior in these methods (als, at least) as part of M-1365. the scala side haven't been using RandomUtil sweepingly but probably should.
On Fri, Jun 6, 2014 at 2:35 PM, Andrew Palumbo <[email protected]> wrote: > looks like it was a random/threshold issue- ran it again and passed > > From: [email protected] > To: [email protected] > Subject: RE: Build > Date: Fri, 6 Jun 2014 17:33:09 -0400 > > > > > I didn't get any failures in mrlegacy but did get one in spark-shell: > > *** 1 TEST FAILED *** > [INFO] > ------------------------------------------------------------------------ > [INFO] Reactor Summary: > [INFO] > [INFO] Mahout Build Tools ................................ SUCCESS [4.056s] > [INFO] Apache Mahout ..................................... SUCCESS [0.787s] > [INFO] Mahout Math ....................................... SUCCESS > [2:36.258s] > [INFO] Mahout MapReduce Legacy ........................... SUCCESS > [17:52.964s] > [INFO] Mahout Integration ................................ SUCCESS > [1:36.630s] > [INFO] Mahout Examples ................................... SUCCESS > [21.628s] > [INFO] Mahout Release Package ............................ SUCCESS [0.153s] > [INFO] Mahout Math/Scala wrappers ........................ SUCCESS > [1:15.464s] > [INFO] Mahout Spark bindings ............................. FAILURE > [3:15.106s] > [INFO] Mahout Spark bindings shell ....................... SKIPPED > > > ALS factorized approximation block: > { > 0 => > {0:0.39474878632299804,1:-0.0869672266359451,2:-1.0574506867969315} > 1 => > {0:0.4076223671588186,1:0.013588403631779156,2:-0.6050329555295999} > 2 => > {0:0.15932451623364444,1:0.07468331598847106,2:-0.3796237843015263} > } > norm of residuals 0.012073 > train iteration rmses: List(2.0234162711717642E-7, 2.9153664240021547E-7) > - als *** FAILED *** > 0.012073455230384868 was not less than 0.01 (MathSuite.scala:208) > > > > > > >
