CPU info:
    CPU Model Name: Intel(R) Xeon(R) CPU W3565 @ 3.20GHz
    Hardware threads: 8
    Total Memory: 12580436 kB
-------------------------------------------------------------------
=== Running C:\Program Files\Microsoft MPI\Bin\/mpiexec.exe -n 2 C:\jenkins\workspace\CNTK-Test-Windows-W1\x64\release\cntk.exe configFile=C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\HTKDeserializers\DNN\ParallelBMWithAdjustLR/cntk.cntk currentDirectory=C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\Data RunDir=C:\Users\svcphil\AppData\Local\Temp\cntk-test-20161221052142.365503\Speech\HTKDeserializers\DNN_ParallelBMWithAdjustLR@release_cpu DataDir=C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\Data ConfigDir=C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\HTKDeserializers\DNN\ParallelBMWithAdjustLR OutputDir=C:\Users\svcphil\AppData\Local\Temp\cntk-test-20161221052142.365503\Speech\HTKDeserializers\DNN_ParallelBMWithAdjustLR@release_cpu DeviceId=-1 timestamping=true numCPUThreads=4 precision=double speechTrain=[SGD=[ParallelTrain=[parallelizationStartEpoch=2]]] stderr=C:\Users\svcphil\AppData\Local\Temp\cntk-test-20161221052142.365503\Speech\HTKDeserializers\DNN_ParallelBMWithAdjustLR@release_cpu/stderr
CNTK 2.0.beta6.0+ (HEAD 0a2e20, Dec 21 2016 04:21:26) on cntk-muc01 at 2016/12/21 05:27:44

C:\jenkins\workspace\CNTK-Test-Windows-W1\x64\release\cntk.exe  configFile=C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\HTKDeserializers\DNN\ParallelBMWithAdjustLR/cntk.cntk  currentDirectory=C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\Data  RunDir=C:\Users\svcphil\AppData\Local\Temp\cntk-test-20161221052142.365503\Speech\HTKDeserializers\DNN_ParallelBMWithAdjustLR@release_cpu  DataDir=C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\Data  ConfigDir=C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\HTKDeserializers\DNN\ParallelBMWithAdjustLR  OutputDir=C:\Users\svcphil\AppData\Local\Temp\cntk-test-20161221052142.365503\Speech\HTKDeserializers\DNN_ParallelBMWithAdjustLR@release_cpu  DeviceId=-1  timestamping=true  numCPUThreads=4  precision=double  speechTrain=[SGD=[ParallelTrain=[parallelizationStartEpoch=2]]]  stderr=C:\Users\svcphil\AppData\Local\Temp\cntk-test-20161221052142.365503\Speech\HTKDeserializers\DNN_ParallelBMWithAdjustLR@release_cpu/stderr
Changed current directory to C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\Data
requestnodes [MPIWrapper]: using 2 out of 2 MPI nodes on a single host (2 requested); we (0) are in (participating)
CNTK 2.0.beta6.0+ (HEAD 0a2e20, Dec 21 2016 04:21:26) on cntk-muc01 at 2016/12/21 05:27:44

C:\jenkins\workspace\CNTK-Test-Windows-W1\x64\release\cntk.exe  configFile=C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\HTKDeserializers\DNN\ParallelBMWithAdjustLR/cntk.cntk  currentDirectory=C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\Data  RunDir=C:\Users\svcphil\AppData\Local\Temp\cntk-test-20161221052142.365503\Speech\HTKDeserializers\DNN_ParallelBMWithAdjustLR@release_cpu  DataDir=C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\Data  ConfigDir=C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\HTKDeserializers\DNN\ParallelBMWithAdjustLR  OutputDir=C:\Users\svcphil\AppData\Local\Temp\cntk-test-20161221052142.365503\Speech\HTKDeserializers\DNN_ParallelBMWithAdjustLR@release_cpu  DeviceId=-1  timestamping=true  numCPUThreads=4  precision=double  speechTrain=[SGD=[ParallelTrain=[parallelizationStartEpoch=2]]]  stderr=C:\Users\svcphil\AppData\Local\Temp\cntk-test-20161221052142.365503\Speech\HTKDeserializers\DNN_ParallelBMWithAdjustLR@release_cpu/stderr
Changed current directory to C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\Data
requestnodes [MPIWrapper]: using 2 out of 2 MPI nodes on a single host (2 requested); we (1) are in (participating)
MPI Rank 0: 12/21/2016 05:27:44: Redirecting stderr to file C:\Users\svcphil\AppData\Local\Temp\cntk-test-20161221052142.365503\Speech\HTKDeserializers\DNN_ParallelBMWithAdjustLR@release_cpu/stderr_SpeechTrain.logrank0
MPI Rank 0: 12/21/2016 05:27:44: -------------------------------------------------------------------
MPI Rank 0: 12/21/2016 05:27:44: Build info: 
MPI Rank 0: 
MPI Rank 0: 12/21/2016 05:27:44: 		Built time: Dec 21 2016 04:21:26
MPI Rank 0: 12/21/2016 05:27:44: 		Last modified date: Tue Dec 20 18:55:12 2016
MPI Rank 0: 12/21/2016 05:27:44: 		Build type: Release
MPI Rank 0: 12/21/2016 05:27:44: 		Build target: GPU
MPI Rank 0: 12/21/2016 05:27:44: 		With ASGD: yes
MPI Rank 0: 12/21/2016 05:27:44: 		Math lib: mkl
MPI Rank 0: 12/21/2016 05:27:44: 		CUDA_PATH: C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v8.0
MPI Rank 0: 12/21/2016 05:27:44: 		CUB_PATH: c:\src\cub-1.4.1
MPI Rank 0: 12/21/2016 05:27:44: 		CUDNN_PATH: C:\local\cudnn-8.0-windows10-x64-v5.1
MPI Rank 0: 12/21/2016 05:27:44: 		Build Branch: HEAD
MPI Rank 0: 12/21/2016 05:27:44: 		Build SHA1: 0a2e20ddce32ca3cd458ef0358757e1489d9afe3 (modified)
MPI Rank 0: 12/21/2016 05:27:44: 		Built by svcphil on LIANA-09-w
MPI Rank 0: 12/21/2016 05:27:44: 		Build Path: C:\jenkins\workspace\CNTK-Build-Windows\Source\CNTK\
MPI Rank 0: 12/21/2016 05:27:44: -------------------------------------------------------------------
MPI Rank 0: 12/21/2016 05:27:45: -------------------------------------------------------------------
MPI Rank 0: 12/21/2016 05:27:45: GPU info:
MPI Rank 0: 
MPI Rank 0: 12/21/2016 05:27:45: 		Device[0]: cores = 2496; computeCapability = 5.2; type = "Quadro M4000"; memory = 8192 MB
MPI Rank 0: 12/21/2016 05:27:45: -------------------------------------------------------------------
MPI Rank 0: 
MPI Rank 0: Configuration After Processing and Variable Resolution:
MPI Rank 0: 
MPI Rank 0: configparameters: cntk.cntk:command=SpeechTrain
MPI Rank 0: configparameters: cntk.cntk:ConfigDir=C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\HTKDeserializers\DNN\ParallelBMWithAdjustLR
MPI Rank 0: configparameters: cntk.cntk:currentDirectory=C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\Data
MPI Rank 0: configparameters: cntk.cntk:DataDir=C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\Data
MPI Rank 0: configparameters: cntk.cntk:deviceId=-1
MPI Rank 0: configparameters: cntk.cntk:framemode=true
MPI Rank 0: configparameters: cntk.cntk:modelPath=C:\Users\svcphil\AppData\Local\Temp\cntk-test-20161221052142.365503\Speech\HTKDeserializers\DNN_ParallelBMWithAdjustLR@release_cpu/models/cntkSpeech.dnn
MPI Rank 0: configparameters: cntk.cntk:numCPUThreads=4
MPI Rank 0: configparameters: cntk.cntk:OutputDir=C:\Users\svcphil\AppData\Local\Temp\cntk-test-20161221052142.365503\Speech\HTKDeserializers\DNN_ParallelBMWithAdjustLR@release_cpu
MPI Rank 0: configparameters: cntk.cntk:parallelTrain=true
MPI Rank 0: configparameters: cntk.cntk:precision=double
MPI Rank 0: configparameters: cntk.cntk:RunDir=C:\Users\svcphil\AppData\Local\Temp\cntk-test-20161221052142.365503\Speech\HTKDeserializers\DNN_ParallelBMWithAdjustLR@release_cpu
MPI Rank 0: configparameters: cntk.cntk:SpeechTrain={
MPI Rank 0:     action = "train"
MPI Rank 0:     BrainScriptNetworkBuilder = {
MPI Rank 0:         layerSizes = 363:512:512:132
MPI Rank 0:         trainingCriterion = 'CE'
MPI Rank 0:         evalCriterion = 'Err'
MPI Rank 0:         applyMeanVarNorm = true
MPI Rank 0:         L = Length(layerSizes)-1    // number of model layers
MPI Rank 0:         features = Input { layerSizes[0] } // 1, tag='feature' 
MPI Rank 0:         labels = Input { layerSizes[L] } // 1, tag='label' 
MPI Rank 0:         featNorm = if applyMeanVarNorm
MPI Rank 0:                    then MeanVarNorm(features)
MPI Rank 0:                    else features
MPI Rank 0:         layers[layer:1..L-1] = if layer > 1
MPI Rank 0:                                then SBFF(layers[layer-1].Eh, layerSizes[layer], layerSizes[layer-1])
MPI Rank 0:                                else SBFF(featNorm, layerSizes[layer], layerSizes[layer-1])
MPI Rank 0:         outLayer = BFF(layers[L-1].Eh, layerSizes[L], layerSizes[L-1])
MPI Rank 0:         outZ = outLayer.z        // + PastValue(layerSizes[L], 1, outLayer.z)
MPI Rank 0:         CE = if trainingCriterion == 'CE'
MPI Rank 0:              then CrossEntropyWithSoftmax(labels, outZ, tag='criterion')
MPI Rank 0:              else Fail('unknown trainingCriterion ' + trainingCriterion)
MPI Rank 0:         Err = if evalCriterion == 'Err' then
MPI Rank 0:               ClassificationError(labels, outZ, tag='evaluation')
MPI Rank 0:               else Fail('unknown evalCriterion ' + evalCriterion)
MPI Rank 0:         logPrior = LogPrior(labels)
MPI Rank 0:         // TODO: how to add a tag to an infix operation?
MPI Rank 0:         ScaledLogLikelihood = Minus (outZ, logPrior, tag='output')
MPI Rank 0:     }
MPI Rank 0:     SGD = {
MPI Rank 0:         epochSize = 20480
MPI Rank 0:         minibatchSize = 64:256:1024
MPI Rank 0:         learningRatesPerMB = 1.0:0.5:0.1
MPI Rank 0:         numMBsToShowResult = 1
MPI Rank 0:         momentumPerMB = 0.9:0.656119
MPI Rank 0:         dropoutRate = 0.0
MPI Rank 0:         maxEpochs = 5
MPI Rank 0:         keepCheckPointFiles = true
MPI Rank 0:         clippingThresholdPerSample = 1#INF
MPI Rank 0:         ParallelTrain = {
MPI Rank 0:             parallelizationMethod = "BlockMomentumSGD"
MPI Rank 0:             distributedMBReading = true
MPI Rank 0:             syncPerfStats=1
MPI Rank 0:             BlockMomentumSGD = {
MPI Rank 0:                 blockSizePerWorker=2048
MPI Rank 0:                 resetSGDMomentum=true
MPI Rank 0:                 useNesterovMomentum=true
MPI Rank 0:             }
MPI Rank 0:         }
MPI Rank 0:         AutoAdjust = {
MPI Rank 0:             reduceLearnRateIfImproveLessThan = 0
MPI Rank 0:             loadBestModel = true
MPI Rank 0:             increaseLearnRateIfImproveMoreThan = 1000000000
MPI Rank 0:             learnRateDecreaseFactor = 0.5
MPI Rank 0:             learnRateIncreaseFactor = 1.382
MPI Rank 0:             autoAdjustLR = "searchBeforeEpoch"
MPI Rank 0:             numMiniBatch4LRSearch = 20
MPI Rank 0:             numPrevLearnRates = 3            
MPI Rank 0:         }
MPI Rank 0:     }
MPI Rank 0:     reader = {
MPI Rank 0:         verbosity = 0 ; randomize = true
MPI Rank 0:         deserializers = ({
MPI Rank 0:             type = "HTKFeatureDeserializer"
MPI Rank 0:             module = "HTKDeserializers"                
MPI Rank 0:             input = { features = { dim = 363 ; scpFile = "glob_0000.scp" } }
MPI Rank 0:         }:{
MPI Rank 0:             type = "HTKMLFDeserializer"
MPI Rank 0:             module = "HTKDeserializers"
MPI Rank 0:             input = { labels = { mlfFile = "C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\Data/glob_0000.mlf" ; labelMappingFile = "C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\Data/state.list" ; labelDim = 132 } }
MPI Rank 0:         })
MPI Rank 0:     }
MPI Rank 0: } [SGD=[ParallelTrain=[parallelizationStartEpoch=2]]]
MPI Rank 0: 
MPI Rank 0: configparameters: cntk.cntk:stderr=C:\Users\svcphil\AppData\Local\Temp\cntk-test-20161221052142.365503\Speech\HTKDeserializers\DNN_ParallelBMWithAdjustLR@release_cpu/stderr
MPI Rank 0: configparameters: cntk.cntk:timestamping=true
MPI Rank 0: configparameters: cntk.cntk:traceLevel=1
MPI Rank 0: 12/21/2016 05:27:45: Commands: SpeechTrain
MPI Rank 0: 12/21/2016 05:27:45: precision = "double"
MPI Rank 0: 12/21/2016 05:27:45: Using 4 CPU threads.
MPI Rank 0: 
MPI Rank 0: 12/21/2016 05:27:45: ##############################################################################
MPI Rank 0: 12/21/2016 05:27:45: #                                                                            #
MPI Rank 0: 12/21/2016 05:27:45: # SpeechTrain command (train action)                                         #
MPI Rank 0: 12/21/2016 05:27:45: #                                                                            #
MPI Rank 0: 12/21/2016 05:27:45: ##############################################################################
MPI Rank 0: 
MPI Rank 0: 12/21/2016 05:27:45: WARNING: 'numMiniBatch4LRSearch' is deprecated, please remove it and use 'numSamples4Search' instead.
MPI Rank 0: 12/21/2016 05:27:45: 
MPI Rank 0: Creating virgin network.
MPI Rank 0: 
MPI Rank 0: Post-processing network...
MPI Rank 0: 
MPI Rank 0: 6 roots:
MPI Rank 0: 	CE = CrossEntropyWithSoftmax()
MPI Rank 0: 	Err = ClassificationError()
MPI Rank 0: 	ScaledLogLikelihood = Minus()
MPI Rank 0: 	featNorm.invStdDev = InvStdDev()
MPI Rank 0: 	featNorm.mean = Mean()
MPI Rank 0: 	logPrior._ = Mean()
MPI Rank 0: 
MPI Rank 0: Validating network. 25 nodes to process in pass 1.
MPI Rank 0: 
MPI Rank 0: Validating --> labels = InputValue() :  -> [132 x *]
MPI Rank 0: Validating --> outLayer.W = LearnableParameter() :  -> [132 x 512]
MPI Rank 0: Validating --> layers[2].Eh._.W = LearnableParameter() :  -> [512 x 512]
MPI Rank 0: Validating --> layers[1].Eh._.W = LearnableParameter() :  -> [512 x 363]
MPI Rank 0: Validating --> features = InputValue() :  -> [363 x *]
MPI Rank 0: Validating --> featNorm.mean = Mean (features) : [363 x *] -> [363]
MPI Rank 0: Validating --> featNorm.ElementTimesArgs[0] = Minus (features, featNorm.mean) : [363 x *], [363] -> [363 x *]
MPI Rank 0: Validating --> featNorm.invStdDev = InvStdDev (features) : [363 x *] -> [363]
MPI Rank 0: Validating --> featNorm = ElementTimes (featNorm.ElementTimesArgs[0], featNorm.invStdDev) : [363 x *], [363] -> [363 x *]
MPI Rank 0: Validating --> layers[1].Eh._.z.PlusArgs[0] = Times (layers[1].Eh._.W, featNorm) : [512 x 363], [363 x *] -> [512 x *]
MPI Rank 0: Validating --> layers[1].Eh._.B = LearnableParameter() :  -> [512 x 1]
MPI Rank 0: Validating --> layers[1].Eh._.z = Plus (layers[1].Eh._.z.PlusArgs[0], layers[1].Eh._.B) : [512 x *], [512 x 1] -> [512 x 1 x *]
MPI Rank 0: Validating --> layers[1].Eh = Sigmoid (layers[1].Eh._.z) : [512 x 1 x *] -> [512 x 1 x *]
MPI Rank 0: Validating --> layers[2].Eh._.z.PlusArgs[0] = Times (layers[2].Eh._.W, layers[1].Eh) : [512 x 512], [512 x 1 x *] -> [512 x 1 x *]
MPI Rank 0: Validating --> layers[2].Eh._.B = LearnableParameter() :  -> [512 x 1]
MPI Rank 0: Validating --> layers[2].Eh._.z = Plus (layers[2].Eh._.z.PlusArgs[0], layers[2].Eh._.B) : [512 x 1 x *], [512 x 1] -> [512 x 1 x *]
MPI Rank 0: Validating --> layers[2].Eh = Sigmoid (layers[2].Eh._.z) : [512 x 1 x *] -> [512 x 1 x *]
MPI Rank 0: Validating --> outLayer.z.PlusArgs[0] = Times (outLayer.W, layers[2].Eh) : [132 x 512], [512 x 1 x *] -> [132 x 1 x *]
MPI Rank 0: Validating --> outLayer.B = LearnableParameter() :  -> [132 x 1]
MPI Rank 0: Validating --> outZ = Plus (outLayer.z.PlusArgs[0], outLayer.B) : [132 x 1 x *], [132 x 1] -> [132 x 1 x *]
MPI Rank 0: Validating --> CE = CrossEntropyWithSoftmax (labels, outZ) : [132 x *], [132 x 1 x *] -> [1]
MPI Rank 0: Validating --> Err = ClassificationError (labels, outZ) : [132 x *], [132 x 1 x *] -> [1]
MPI Rank 0: Validating --> logPrior._ = Mean (labels) : [132 x *] -> [132]
MPI Rank 0: Validating --> logPrior = Log (logPrior._) : [132] -> [132]
MPI Rank 0: Validating --> ScaledLogLikelihood = Minus (outZ, logPrior) : [132 x 1 x *], [132] -> [132 x 1 x *]
MPI Rank 0: 
MPI Rank 0: Validating network. 17 nodes to process in pass 2.
MPI Rank 0: 
MPI Rank 0: 
MPI Rank 0: Validating network, final pass.
MPI Rank 0: 
MPI Rank 0: 
MPI Rank 0: 
MPI Rank 0: 
MPI Rank 0: Post-processing network complete.
MPI Rank 0: 
MPI Rank 0: Reading script file glob_0000.scp ... 948 entries
MPI Rank 0: HTKDataDeserializer::HTKDataDeserializer: selected 948 utterances grouped into 3 chunks, average chunk size: 316.0 utterances, 84244.7 frames (for I/O: 316.0 utterances, 84244.7 frames)
MPI Rank 0: HTKDataDeserializer::HTKDataDeserializer: determined feature kind as 33-dimensional 'USER' with frame shift 10.0 ms
MPI Rank 0: total 132 state names in state list C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\Data/state.list
MPI Rank 0: htkmlfreader: reading MLF file C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\Data/glob_0000.mlf ... total 948 entries
MPI Rank 0: MLFDataDeserializer::MLFDataDeserializer: 948 utterances with 252734 frames in 129 classes
MPI Rank 0: 12/21/2016 05:27:45: 
MPI Rank 0: Model has 25 nodes. Using CPU.
MPI Rank 0: 
MPI Rank 0: 12/21/2016 05:27:45: Training criterion:   CE = CrossEntropyWithSoftmax
MPI Rank 0: 12/21/2016 05:27:45: Evaluation criterion: Err = ClassificationError
MPI Rank 0: 
MPI Rank 0: 
MPI Rank 0: Allocating matrices for forward and/or backward propagation.
MPI Rank 0: 
MPI Rank 0: Memory Sharing: Out of 40 matrices, 19 are shared as 8, and 21 are not shared.
MPI Rank 0: 
MPI Rank 0: 	{ layers[2].Eh : [512 x 1 x *] (gradient)
MPI Rank 0: 	  layers[2].Eh._.B : [512 x 1] (gradient)
MPI Rank 0: 	  outZ : [132 x 1 x *] (gradient) }
MPI Rank 0: 	{ layers[1].Eh : [512 x 1 x *]
MPI Rank 0: 	  layers[1].Eh._.z.PlusArgs[0] : [512 x *] (gradient) }
MPI Rank 0: 	{ outLayer.W : [132 x 512] (gradient)
MPI Rank 0: 	  outZ : [132 x 1 x *] }
MPI Rank 0: 	{ layers[1].Eh : [512 x 1 x *] (gradient)
MPI Rank 0: 	  layers[1].Eh._.B : [512 x 1] (gradient)
MPI Rank 0: 	  layers[2].Eh._.z : [512 x 1 x *] (gradient)
MPI Rank 0: 	  outLayer.z.PlusArgs[0] : [132 x 1 x *] }
MPI Rank 0: 	{ layers[1].Eh._.W : [512 x 363] (gradient)
MPI Rank 0: 	  layers[1].Eh._.z : [512 x 1 x *] }
MPI Rank 0: 	{ layers[2].Eh : [512 x 1 x *]
MPI Rank 0: 	  layers[2].Eh._.z.PlusArgs[0] : [512 x 1 x *] (gradient) }
MPI Rank 0: 	{ layers[1].Eh._.z : [512 x 1 x *] (gradient)
MPI Rank 0: 	  layers[2].Eh._.z.PlusArgs[0] : [512 x 1 x *] }
MPI Rank 0: 	{ layers[2].Eh._.W : [512 x 512] (gradient)
MPI Rank 0: 	  layers[2].Eh._.z : [512 x 1 x *] }
MPI Rank 0: 
MPI Rank 0: 
MPI Rank 0: 12/21/2016 05:27:45: Training 516740 parameters in 6 out of 6 parameter tensors and 15 nodes with gradient:
MPI Rank 0: 
MPI Rank 0: 12/21/2016 05:27:45: 	Node 'layers[1].Eh._.B' (LearnableParameter operation) : [512 x 1]
MPI Rank 0: 12/21/2016 05:27:45: 	Node 'layers[1].Eh._.W' (LearnableParameter operation) : [512 x 363]
MPI Rank 0: 12/21/2016 05:27:45: 	Node 'layers[2].Eh._.B' (LearnableParameter operation) : [512 x 1]
MPI Rank 0: 12/21/2016 05:27:45: 	Node 'layers[2].Eh._.W' (LearnableParameter operation) : [512 x 512]
MPI Rank 0: 12/21/2016 05:27:45: 	Node 'outLayer.B' (LearnableParameter operation) : [132 x 1]
MPI Rank 0: 12/21/2016 05:27:45: 	Node 'outLayer.W' (LearnableParameter operation) : [132 x 512]
MPI Rank 0: 
MPI Rank 0: 
MPI Rank 0: 12/21/2016 05:27:45: Precomputing --> 3 PreCompute nodes found.
MPI Rank 0: 
MPI Rank 0: 12/21/2016 05:27:45: 	featNorm.mean = Mean()
MPI Rank 0: 12/21/2016 05:27:45: 	featNorm.invStdDev = InvStdDev()
MPI Rank 0: 12/21/2016 05:27:45: 	logPrior._ = Mean()
MPI Rank 0: 
MPI Rank 0: 12/21/2016 05:27:49: Precomputing --> Completed.
MPI Rank 0: 
MPI Rank 0: 
MPI Rank 0: 12/21/2016 05:27:49: Starting Epoch 1: learning rate per sample = 0.015625  effective momentum = 0.900000  momentum as time constant = 607.4 samples
MPI Rank 0: 
MPI Rank 0: 12/21/2016 05:27:49: Starting minibatch loop.
MPI Rank 0: 12/21/2016 05:27:49:  Epoch[ 1 of 5]-Minibatch[   1-   1, 0.31%]: CE = 4.91295596 * 64; Err = 0.96875000 * 64; time = 0.0346s; samplesPerSecond = 1848.4
MPI Rank 0: 12/21/2016 05:27:49:  Epoch[ 1 of 5]-Minibatch[   2-   2, 0.63%]: CE = 4.78498529 * 64; Err = 1.00000000 * 64; time = 0.0281s; samplesPerSecond = 2275.2
MPI Rank 0: 12/21/2016 05:27:49:  Epoch[ 1 of 5]-Minibatch[   3-   3, 0.94%]: CE = 4.19018696 * 64; Err = 0.81250000 * 64; time = 0.0301s; samplesPerSecond = 2124.6
MPI Rank 0: 12/21/2016 05:27:49:  Epoch[ 1 of 5]-Minibatch[   4-   4, 1.25%]: CE = 4.46135476 * 64; Err = 0.82812500 * 64; time = 0.0342s; samplesPerSecond = 1871.3
MPI Rank 0: 12/21/2016 05:27:49:  Epoch[ 1 of 5]-Minibatch[   5-   5, 1.56%]: CE = 4.72788003 * 64; Err = 0.92187500 * 64; time = 0.0280s; samplesPerSecond = 2287.2
MPI Rank 0: 12/21/2016 05:27:49:  Epoch[ 1 of 5]-Minibatch[   6-   6, 1.88%]: CE = 4.07654096 * 64; Err = 0.89062500 * 64; time = 0.0291s; samplesPerSecond = 2197.8
MPI Rank 0: 12/21/2016 05:27:49:  Epoch[ 1 of 5]-Minibatch[   7-   7, 2.19%]: CE = 4.50165607 * 64; Err = 0.96875000 * 64; time = 0.0259s; samplesPerSecond = 2468.2
MPI Rank 0: 12/21/2016 05:27:49:  Epoch[ 1 of 5]-Minibatch[   8-   8, 2.50%]: CE = 4.93153999 * 64; Err = 0.89062500 * 64; time = 0.0357s; samplesPerSecond = 1794.9
MPI Rank 0: 12/21/2016 05:27:49:  Epoch[ 1 of 5]-Minibatch[   9-   9, 2.81%]: CE = 4.79817443 * 64; Err = 0.93750000 * 64; time = 0.0298s; samplesPerSecond = 2147.7
MPI Rank 0: 12/21/2016 05:27:49:  Epoch[ 1 of 5]-Minibatch[  10-  10, 3.13%]: CE = 4.46089875 * 64; Err = 0.96875000 * 64; time = 0.0271s; samplesPerSecond = 2358.3
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  11-  11, 3.44%]: CE = 4.34462020 * 64; Err = 0.90625000 * 64; time = 0.1605s; samplesPerSecond = 398.7
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  12-  12, 3.75%]: CE = 3.91243070 * 64; Err = 0.87500000 * 64; time = 0.0175s; samplesPerSecond = 3658.2
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  13-  13, 4.06%]: CE = 4.73715179 * 64; Err = 0.92187500 * 64; time = 0.0153s; samplesPerSecond = 4172.1
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  14-  14, 4.38%]: CE = 4.42160986 * 64; Err = 0.93750000 * 64; time = 0.0262s; samplesPerSecond = 2443.0
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  15-  15, 4.69%]: CE = 4.14675744 * 64; Err = 0.85937500 * 64; time = 0.0161s; samplesPerSecond = 3978.6
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  16-  16, 5.00%]: CE = 4.50951186 * 64; Err = 0.95312500 * 64; time = 0.0242s; samplesPerSecond = 2641.6
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  17-  17, 5.31%]: CE = 4.30758210 * 64; Err = 0.85937500 * 64; time = 0.0235s; samplesPerSecond = 2721.9
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  18-  18, 5.63%]: CE = 4.34534841 * 64; Err = 1.00000000 * 64; time = 0.0199s; samplesPerSecond = 3221.7
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  19-  19, 5.94%]: CE = 4.19517128 * 64; Err = 0.96875000 * 64; time = 0.0243s; samplesPerSecond = 2639.0
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  20-  20, 6.25%]: CE = 4.41248710 * 64; Err = 0.98437500 * 64; time = 0.0198s; samplesPerSecond = 3234.8
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  21-  21, 6.56%]: CE = 4.10891079 * 64; Err = 0.92187500 * 64; time = 0.0252s; samplesPerSecond = 2538.1
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  22-  22, 6.88%]: CE = 4.16379766 * 64; Err = 0.85937500 * 64; time = 0.0209s; samplesPerSecond = 3069.2
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  23-  23, 7.19%]: CE = 4.09455579 * 64; Err = 0.92187500 * 64; time = 0.0253s; samplesPerSecond = 2532.6
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  24-  24, 7.50%]: CE = 3.95980469 * 64; Err = 0.89062500 * 64; time = 0.0278s; samplesPerSecond = 2299.6
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  25-  25, 7.81%]: CE = 4.05428109 * 64; Err = 0.87500000 * 64; time = 0.0569s; samplesPerSecond = 1125.2
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  26-  26, 8.13%]: CE = 4.16245451 * 64; Err = 0.84375000 * 64; time = 0.0590s; samplesPerSecond = 1085.5
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  27-  27, 8.44%]: CE = 3.71756327 * 64; Err = 0.84375000 * 64; time = 0.0197s; samplesPerSecond = 3246.1
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  28-  28, 8.75%]: CE = 3.80779138 * 64; Err = 0.87500000 * 64; time = 0.0302s; samplesPerSecond = 2119.3
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  29-  29, 9.06%]: CE = 3.72564857 * 64; Err = 0.81250000 * 64; time = 0.1538s; samplesPerSecond = 416.2
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  30-  30, 9.38%]: CE = 4.01963243 * 64; Err = 0.87500000 * 64; time = 0.0280s; samplesPerSecond = 2287.2
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  31-  31, 9.69%]: CE = 3.68590709 * 64; Err = 0.89062500 * 64; time = 0.0249s; samplesPerSecond = 2568.9
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  32-  32, 10.00%]: CE = 3.81516754 * 64; Err = 0.78125000 * 64; time = 0.0240s; samplesPerSecond = 2668.9
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  33-  33, 10.31%]: CE = 3.93685037 * 64; Err = 0.87500000 * 64; time = 0.0255s; samplesPerSecond = 2506.8
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  34-  34, 10.63%]: CE = 3.96481462 * 64; Err = 0.90625000 * 64; time = 0.0239s; samplesPerSecond = 2679.4
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  35-  35, 10.94%]: CE = 3.57865409 * 64; Err = 0.84375000 * 64; time = 0.0251s; samplesPerSecond = 2554.2
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  36-  36, 11.25%]: CE = 3.72265528 * 64; Err = 0.85937500 * 64; time = 0.0239s; samplesPerSecond = 2680.6
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  37-  37, 11.56%]: CE = 3.71485627 * 64; Err = 0.84375000 * 64; time = 0.0246s; samplesPerSecond = 2605.7
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  38-  38, 11.88%]: CE = 4.04042687 * 64; Err = 0.87500000 * 64; time = 0.0243s; samplesPerSecond = 2638.4
MPI Rank 0: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  39-  39, 12.19%]: CE = 3.48663283 * 64; Err = 0.76562500 * 64; time = 0.0250s; samplesPerSecond = 2554.9
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  40-  40, 12.50%]: CE = 3.48828968 * 64; Err = 0.81250000 * 64; time = 0.0242s; samplesPerSecond = 2640.2
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  41-  41, 12.81%]: CE = 3.46883616 * 64; Err = 0.78125000 * 64; time = 0.0248s; samplesPerSecond = 2575.6
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  42-  42, 13.13%]: CE = 4.12832965 * 64; Err = 0.90625000 * 64; time = 0.0239s; samplesPerSecond = 2681.6
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  43-  43, 13.44%]: CE = 3.82286476 * 64; Err = 0.90625000 * 64; time = 0.0247s; samplesPerSecond = 2589.4
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  44-  44, 13.75%]: CE = 3.99396471 * 64; Err = 0.90625000 * 64; time = 0.0241s; samplesPerSecond = 2650.2
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  45-  45, 14.06%]: CE = 3.84953256 * 64; Err = 0.89062500 * 64; time = 0.0255s; samplesPerSecond = 2512.7
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  46-  46, 14.37%]: CE = 3.57917953 * 64; Err = 0.79687500 * 64; time = 0.0238s; samplesPerSecond = 2686.1
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  47-  47, 14.69%]: CE = 3.86079148 * 64; Err = 0.84375000 * 64; time = 0.0248s; samplesPerSecond = 2575.9
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  48-  48, 15.00%]: CE = 3.88891763 * 64; Err = 0.85937500 * 64; time = 0.0243s; samplesPerSecond = 2638.8
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  49-  49, 15.31%]: CE = 3.94662742 * 64; Err = 0.89062500 * 64; time = 0.1398s; samplesPerSecond = 457.9
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  50-  50, 15.63%]: CE = 3.83644301 * 64; Err = 0.87500000 * 64; time = 0.0240s; samplesPerSecond = 2661.3
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  51-  51, 15.94%]: CE = 3.66716866 * 64; Err = 0.89062500 * 64; time = 0.0254s; samplesPerSecond = 2523.5
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  52-  52, 16.25%]: CE = 4.00651571 * 64; Err = 0.90625000 * 64; time = 0.0239s; samplesPerSecond = 2675.5
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  53-  53, 16.56%]: CE = 3.80511656 * 64; Err = 0.81250000 * 64; time = 0.0247s; samplesPerSecond = 2594.8
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  54-  54, 16.88%]: CE = 3.93380989 * 64; Err = 0.85937500 * 64; time = 0.0240s; samplesPerSecond = 2663.1
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  55-  55, 17.19%]: CE = 3.49394937 * 64; Err = 0.84375000 * 64; time = 0.0249s; samplesPerSecond = 2573.3
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  56-  56, 17.50%]: CE = 3.42224075 * 64; Err = 0.84375000 * 64; time = 0.0239s; samplesPerSecond = 2676.1
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  57-  57, 17.81%]: CE = 3.76078536 * 64; Err = 0.85937500 * 64; time = 0.0251s; samplesPerSecond = 2554.1
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  58-  58, 18.13%]: CE = 3.80639497 * 64; Err = 0.87500000 * 64; time = 0.0280s; samplesPerSecond = 2285.4
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  59-  59, 18.44%]: CE = 3.55543971 * 64; Err = 0.89062500 * 64; time = 0.0248s; samplesPerSecond = 2577.5
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  60-  60, 18.75%]: CE = 3.55947249 * 64; Err = 0.82812500 * 64; time = 0.0240s; samplesPerSecond = 2669.2
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  61-  61, 19.06%]: CE = 3.21133907 * 64; Err = 0.79687500 * 64; time = 0.0251s; samplesPerSecond = 2550.4
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  62-  62, 19.38%]: CE = 3.30807309 * 64; Err = 0.68750000 * 64; time = 0.0240s; samplesPerSecond = 2662.8
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  63-  63, 19.69%]: CE = 3.54643060 * 64; Err = 0.78125000 * 64; time = 0.0248s; samplesPerSecond = 2583.4
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  64-  64, 20.00%]: CE = 3.48819921 * 64; Err = 0.85937500 * 64; time = 0.0239s; samplesPerSecond = 2682.7
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  65-  65, 20.31%]: CE = 3.53098379 * 64; Err = 0.81250000 * 64; time = 0.0252s; samplesPerSecond = 2536.9
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  66-  66, 20.63%]: CE = 3.18218574 * 64; Err = 0.70312500 * 64; time = 0.0241s; samplesPerSecond = 2654.4
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  67-  67, 20.94%]: CE = 3.62919777 * 64; Err = 0.79687500 * 64; time = 0.0246s; samplesPerSecond = 2601.5
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  68-  68, 21.25%]: CE = 3.30344749 * 64; Err = 0.76562500 * 64; time = 0.0249s; samplesPerSecond = 2571.6
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  69-  69, 21.56%]: CE = 3.11192070 * 64; Err = 0.75000000 * 64; time = 0.1352s; samplesPerSecond = 473.3
MPI Rank 0: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  70-  70, 21.88%]: CE = 3.70063691 * 64; Err = 0.79687500 * 64; time = 0.0241s; samplesPerSecond = 2653.6
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  71-  71, 22.19%]: CE = 3.76244503 * 64; Err = 0.84375000 * 64; time = 0.0247s; samplesPerSecond = 2590.4
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  72-  72, 22.50%]: CE = 3.52103388 * 64; Err = 0.81250000 * 64; time = 0.0241s; samplesPerSecond = 2654.8
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  73-  73, 22.81%]: CE = 3.73227550 * 64; Err = 0.87500000 * 64; time = 0.0249s; samplesPerSecond = 2574.7
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  74-  74, 23.13%]: CE = 3.28056294 * 64; Err = 0.75000000 * 64; time = 0.0239s; samplesPerSecond = 2677.8
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  75-  75, 23.44%]: CE = 3.88497398 * 64; Err = 0.89062500 * 64; time = 0.0248s; samplesPerSecond = 2580.2
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  76-  76, 23.75%]: CE = 3.62146548 * 64; Err = 0.85937500 * 64; time = 0.0239s; samplesPerSecond = 2675.9
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  77-  77, 24.06%]: CE = 3.11930348 * 64; Err = 0.73437500 * 64; time = 0.0251s; samplesPerSecond = 2545.7
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  78-  78, 24.38%]: CE = 3.34530218 * 64; Err = 0.87500000 * 64; time = 0.0239s; samplesPerSecond = 2678.2
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  79-  79, 24.69%]: CE = 3.51426589 * 64; Err = 0.84375000 * 64; time = 0.0248s; samplesPerSecond = 2575.6
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  80-  80, 25.00%]: CE = 3.40713594 * 64; Err = 0.81250000 * 64; time = 0.0239s; samplesPerSecond = 2679.3
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  81-  81, 25.31%]: CE = 3.59134827 * 64; Err = 0.82812500 * 64; time = 0.0252s; samplesPerSecond = 2542.6
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  82-  82, 25.62%]: CE = 3.52703040 * 64; Err = 0.82812500 * 64; time = 0.0238s; samplesPerSecond = 2688.7
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  83-  83, 25.94%]: CE = 3.22259624 * 64; Err = 0.79687500 * 64; time = 0.0250s; samplesPerSecond = 2557.6
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  84-  84, 26.25%]: CE = 3.64961943 * 64; Err = 0.82812500 * 64; time = 0.0240s; samplesPerSecond = 2666.2
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  85-  85, 26.56%]: CE = 3.70782192 * 64; Err = 0.76562500 * 64; time = 0.0250s; samplesPerSecond = 2554.9
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  86-  86, 26.88%]: CE = 3.53921564 * 64; Err = 0.89062500 * 64; time = 0.0239s; samplesPerSecond = 2682.5
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  87-  87, 27.19%]: CE = 3.38712792 * 64; Err = 0.81250000 * 64; time = 0.0252s; samplesPerSecond = 2542.0
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  88-  88, 27.50%]: CE = 3.66470493 * 64; Err = 0.78125000 * 64; time = 0.0240s; samplesPerSecond = 2666.1
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  89-  89, 27.81%]: CE = 3.12758734 * 64; Err = 0.84375000 * 64; time = 0.1442s; samplesPerSecond = 443.8
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  90-  90, 28.13%]: CE = 3.52072988 * 64; Err = 0.82812500 * 64; time = 0.0241s; samplesPerSecond = 2655.5
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  91-  91, 28.44%]: CE = 3.45630741 * 64; Err = 0.76562500 * 64; time = 0.0252s; samplesPerSecond = 2541.5
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  92-  92, 28.75%]: CE = 3.19535282 * 64; Err = 0.78125000 * 64; time = 0.0241s; samplesPerSecond = 2657.7
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  93-  93, 29.06%]: CE = 3.40545723 * 64; Err = 0.81250000 * 64; time = 0.0248s; samplesPerSecond = 2578.7
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  94-  94, 29.38%]: CE = 3.47518793 * 64; Err = 0.70312500 * 64; time = 0.0239s; samplesPerSecond = 2673.9
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  95-  95, 29.69%]: CE = 3.32919398 * 64; Err = 0.78125000 * 64; time = 0.0248s; samplesPerSecond = 2581.7
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  96-  96, 30.00%]: CE = 3.86499937 * 64; Err = 0.93750000 * 64; time = 0.0240s; samplesPerSecond = 2668.1
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  97-  97, 30.31%]: CE = 3.42288014 * 64; Err = 0.84375000 * 64; time = 0.0249s; samplesPerSecond = 2569.8
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  98-  98, 30.63%]: CE = 3.31506114 * 64; Err = 0.82812500 * 64; time = 0.0241s; samplesPerSecond = 2655.8
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  99-  99, 30.94%]: CE = 3.28863365 * 64; Err = 0.76562500 * 64; time = 0.0249s; samplesPerSecond = 2573.1
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[ 100- 100, 31.25%]: CE = 3.20182099 * 64; Err = 0.76562500 * 64; time = 0.0240s; samplesPerSecond = 2661.3
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[ 101- 101, 31.56%]: CE = 3.75128437 * 64; Err = 0.89062500 * 64; time = 0.0249s; samplesPerSecond = 2575.2
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[ 102- 102, 31.87%]: CE = 3.57333316 * 64; Err = 0.84375000 * 64; time = 0.0244s; samplesPerSecond = 2625.4
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[ 103- 103, 32.19%]: CE = 3.65041879 * 64; Err = 0.81250000 * 64; time = 0.0249s; samplesPerSecond = 2570.6
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[ 104- 104, 32.50%]: CE = 3.45052191 * 64; Err = 0.82812500 * 64; time = 0.0241s; samplesPerSecond = 2654.3
MPI Rank 0: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[ 105- 105, 32.81%]: CE = 3.57278549 * 64; Err = 0.85937500 * 64; time = 0.0246s; samplesPerSecond = 2604.9
MPI Rank 0: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 106- 106, 33.13%]: CE = 3.35244169 * 64; Err = 0.87500000 * 64; time = 0.0241s; samplesPerSecond = 2658.2
MPI Rank 0: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 107- 107, 33.44%]: CE = 3.29949185 * 64; Err = 0.76562500 * 64; time = 0.0249s; samplesPerSecond = 2569.4
MPI Rank 0: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 108- 108, 33.75%]: CE = 3.78609758 * 64; Err = 0.82812500 * 64; time = 0.0245s; samplesPerSecond = 2613.5
MPI Rank 0: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 109- 109, 34.06%]: CE = 3.22622650 * 64; Err = 0.78125000 * 64; time = 0.1662s; samplesPerSecond = 385.1
MPI Rank 0: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 110- 110, 34.38%]: CE = 3.29821989 * 64; Err = 0.79687500 * 64; time = 0.0241s; samplesPerSecond = 2657.9
MPI Rank 0: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 111- 111, 34.69%]: CE = 3.44143907 * 64; Err = 0.82812500 * 64; time = 0.0253s; samplesPerSecond = 2527.1
MPI Rank 0: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 112- 112, 35.00%]: CE = 3.44276929 * 64; Err = 0.85937500 * 64; time = 0.0235s; samplesPerSecond = 2725.4
MPI Rank 0: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 113- 113, 35.31%]: CE = 3.18216790 * 64; Err = 0.76562500 * 64; time = 0.0607s; samplesPerSecond = 1053.8
MPI Rank 0: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 114- 114, 35.63%]: CE = 3.18609709 * 64; Err = 0.78125000 * 64; time = 0.0244s; samplesPerSecond = 2618.2
MPI Rank 0: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 115- 115, 35.94%]: CE = 3.06550821 * 64; Err = 0.73437500 * 64; time = 0.0261s; samplesPerSecond = 2455.2
MPI Rank 0: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 116- 116, 36.25%]: CE = 3.43583629 * 64; Err = 0.79687500 * 64; time = 0.0247s; samplesPerSecond = 2593.4
MPI Rank 0: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 117- 117, 36.56%]: CE = 3.10193105 * 64; Err = 0.78125000 * 64; time = 0.0268s; samplesPerSecond = 2391.2
MPI Rank 0: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 118- 118, 36.88%]: CE = 3.42968418 * 64; Err = 0.81250000 * 64; time = 0.0253s; samplesPerSecond = 2525.9
MPI Rank 0: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 119- 119, 37.19%]: CE = 2.85043824 * 64; Err = 0.60937500 * 64; time = 0.0270s; samplesPerSecond = 2371.1
MPI Rank 0: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 120- 120, 37.50%]: CE = 3.50428373 * 64; Err = 0.85937500 * 64; time = 0.0225s; samplesPerSecond = 2847.2
MPI Rank 0: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 121- 121, 37.81%]: CE = 3.28751701 * 64; Err = 0.82812500 * 64; time = 0.0285s; samplesPerSecond = 2246.9
MPI Rank 0: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 122- 122, 38.13%]: CE = 3.79916343 * 64; Err = 0.89062500 * 64; time = 0.0211s; samplesPerSecond = 3029.9
MPI Rank 0: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 123- 123, 38.44%]: CE = 3.55702537 * 64; Err = 0.82812500 * 64; time = 0.1733s; samplesPerSecond = 369.4
MPI Rank 0: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 124- 124, 38.75%]: CE = 3.00217445 * 64; Err = 0.71875000 * 64; time = 0.0183s; samplesPerSecond = 3506.5
MPI Rank 0: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 125- 125, 39.06%]: CE = 3.07327108 * 64; Err = 0.75000000 * 64; time = 0.0272s; samplesPerSecond = 2350.4
MPI Rank 0: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 126- 126, 39.38%]: CE = 2.88353063 * 64; Err = 0.59375000 * 64; time = 0.0274s; samplesPerSecond = 2333.2
MPI Rank 0: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 127- 127, 39.69%]: CE = 3.13059468 * 64; Err = 0.79687500 * 64; time = 0.0267s; samplesPerSecond = 2400.9
MPI Rank 0: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 128- 128, 40.00%]: CE = 3.21732650 * 64; Err = 0.85937500 * 64; time = 0.0239s; samplesPerSecond = 2675.5
MPI Rank 0: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 129- 129, 40.31%]: CE = 2.97299345 * 64; Err = 0.71875000 * 64; time = 0.0251s; samplesPerSecond = 2552.8
MPI Rank 0: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 130- 130, 40.63%]: CE = 2.93691495 * 64; Err = 0.79687500 * 64; time = 0.0185s; samplesPerSecond = 3458.9
MPI Rank 0: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 131- 131, 40.94%]: CE = 3.31837783 * 64; Err = 0.68750000 * 64; time = 0.0253s; samplesPerSecond = 2526.0
MPI Rank 0: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 132- 132, 41.25%]: CE = 2.91929775 * 64; Err = 0.78125000 * 64; time = 0.0307s; samplesPerSecond = 2085.7
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 133- 133, 41.56%]: CE = 3.07940161 * 64; Err = 0.68750000 * 64; time = 0.0234s; samplesPerSecond = 2740.3
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 134- 134, 41.88%]: CE = 3.28344492 * 64; Err = 0.75000000 * 64; time = 0.0262s; samplesPerSecond = 2440.1
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 135- 135, 42.19%]: CE = 3.18447176 * 64; Err = 0.78125000 * 64; time = 0.0215s; samplesPerSecond = 2982.4
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 136- 136, 42.50%]: CE = 2.79093256 * 64; Err = 0.71875000 * 64; time = 0.0278s; samplesPerSecond = 2298.8
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 137- 137, 42.81%]: CE = 2.87937588 * 64; Err = 0.70312500 * 64; time = 0.0219s; samplesPerSecond = 2927.7
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 138- 138, 43.13%]: CE = 2.64594163 * 64; Err = 0.68750000 * 64; time = 0.0215s; samplesPerSecond = 2975.4
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 139- 139, 43.44%]: CE = 2.94206439 * 64; Err = 0.84375000 * 64; time = 0.0266s; samplesPerSecond = 2408.6
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 140- 140, 43.75%]: CE = 3.51285987 * 64; Err = 0.82812500 * 64; time = 0.0295s; samplesPerSecond = 2172.1
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 141- 141, 44.06%]: CE = 3.04888687 * 64; Err = 0.81250000 * 64; time = 0.0236s; samplesPerSecond = 2714.2
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 142- 142, 44.38%]: CE = 3.13123367 * 64; Err = 0.76562500 * 64; time = 0.0271s; samplesPerSecond = 2364.6
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 143- 143, 44.69%]: CE = 2.92926400 * 64; Err = 0.71875000 * 64; time = 0.1586s; samplesPerSecond = 403.4
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 144- 144, 45.00%]: CE = 3.00144780 * 64; Err = 0.71875000 * 64; time = 0.0173s; samplesPerSecond = 3703.5
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 145- 145, 45.31%]: CE = 2.90962694 * 64; Err = 0.67187500 * 64; time = 0.0298s; samplesPerSecond = 2148.2
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 146- 146, 45.63%]: CE = 3.03283171 * 64; Err = 0.79687500 * 64; time = 0.0270s; samplesPerSecond = 2370.9
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 147- 147, 45.94%]: CE = 3.06942741 * 64; Err = 0.73437500 * 64; time = 0.0243s; samplesPerSecond = 2633.3
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 148- 148, 46.25%]: CE = 2.86661978 * 64; Err = 0.65625000 * 64; time = 0.0244s; samplesPerSecond = 2623.9
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 149- 149, 46.56%]: CE = 2.76894440 * 64; Err = 0.68750000 * 64; time = 0.0252s; samplesPerSecond = 2542.8
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 150- 150, 46.88%]: CE = 2.71313692 * 64; Err = 0.59375000 * 64; time = 0.0252s; samplesPerSecond = 2543.6
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 151- 151, 47.19%]: CE = 2.74131048 * 64; Err = 0.65625000 * 64; time = 0.0249s; samplesPerSecond = 2569.0
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 152- 152, 47.50%]: CE = 3.28257238 * 64; Err = 0.71875000 * 64; time = 0.0176s; samplesPerSecond = 3626.3
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 153- 153, 47.81%]: CE = 3.08491448 * 64; Err = 0.76562500 * 64; time = 0.0253s; samplesPerSecond = 2525.8
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 154- 154, 48.13%]: CE = 2.98917665 * 64; Err = 0.71875000 * 64; time = 0.0244s; samplesPerSecond = 2619.6
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 155- 155, 48.44%]: CE = 2.90881148 * 64; Err = 0.81250000 * 64; time = 0.0282s; samplesPerSecond = 2267.4
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 156- 156, 48.75%]: CE = 3.58531995 * 64; Err = 0.78125000 * 64; time = 0.0214s; samplesPerSecond = 2988.3
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 157- 157, 49.06%]: CE = 3.28706069 * 64; Err = 0.75000000 * 64; time = 0.0315s; samplesPerSecond = 2031.2
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 158- 158, 49.38%]: CE = 3.06029676 * 64; Err = 0.81250000 * 64; time = 0.0237s; samplesPerSecond = 2705.6
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 159- 159, 49.69%]: CE = 2.95483403 * 64; Err = 0.68750000 * 64; time = 0.0236s; samplesPerSecond = 2709.8
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 160- 160, 50.00%]: CE = 3.07409648 * 64; Err = 0.76562500 * 64; time = 0.0271s; samplesPerSecond = 2363.6
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 161- 161, 50.31%]: CE = 2.69786051 * 64; Err = 0.67187500 * 64; time = 0.0244s; samplesPerSecond = 2623.5
MPI Rank 0: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 162- 162, 50.63%]: CE = 2.80402381 * 64; Err = 0.70312500 * 64; time = 0.0180s; samplesPerSecond = 3547.5
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 163- 163, 50.94%]: CE = 2.62768914 * 64; Err = 0.62500000 * 64; time = 0.1611s; samplesPerSecond = 397.2
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 164- 164, 51.25%]: CE = 2.64449167 * 64; Err = 0.67187500 * 64; time = 0.0285s; samplesPerSecond = 2244.2
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 165- 165, 51.56%]: CE = 3.08919011 * 64; Err = 0.79687500 * 64; time = 0.0209s; samplesPerSecond = 3062.8
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 166- 166, 51.88%]: CE = 3.07122141 * 64; Err = 0.70312500 * 64; time = 0.0255s; samplesPerSecond = 2510.0
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 167- 167, 52.19%]: CE = 3.05111668 * 64; Err = 0.73437500 * 64; time = 0.0161s; samplesPerSecond = 3963.1
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 168- 168, 52.50%]: CE = 2.90345804 * 64; Err = 0.73437500 * 64; time = 0.0261s; samplesPerSecond = 2456.2
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 169- 169, 52.81%]: CE = 2.58801822 * 64; Err = 0.62500000 * 64; time = 0.0157s; samplesPerSecond = 4076.4
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 170- 170, 53.13%]: CE = 2.68278033 * 64; Err = 0.68750000 * 64; time = 0.0243s; samplesPerSecond = 2630.5
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 171- 171, 53.44%]: CE = 2.89664835 * 64; Err = 0.70312500 * 64; time = 0.0237s; samplesPerSecond = 2702.2
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 172- 172, 53.75%]: CE = 2.61913736 * 64; Err = 0.64062500 * 64; time = 0.0209s; samplesPerSecond = 3055.9
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 173- 173, 54.06%]: CE = 2.68386883 * 64; Err = 0.65625000 * 64; time = 0.0233s; samplesPerSecond = 2741.3
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 174- 174, 54.37%]: CE = 2.63044619 * 64; Err = 0.65625000 * 64; time = 0.0202s; samplesPerSecond = 3164.7
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 175- 175, 54.69%]: CE = 2.39899721 * 64; Err = 0.60937500 * 64; time = 0.0249s; samplesPerSecond = 2567.2
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 176- 176, 55.00%]: CE = 2.88430255 * 64; Err = 0.67187500 * 64; time = 0.0216s; samplesPerSecond = 2959.4
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 177- 177, 55.31%]: CE = 2.83595866 * 64; Err = 0.70312500 * 64; time = 0.0231s; samplesPerSecond = 2766.1
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 178- 178, 55.63%]: CE = 2.79519571 * 64; Err = 0.64062500 * 64; time = 0.0200s; samplesPerSecond = 3195.8
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 179- 179, 55.94%]: CE = 2.76600024 * 64; Err = 0.67187500 * 64; time = 0.0249s; samplesPerSecond = 2574.6
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 180- 180, 56.25%]: CE = 2.59895511 * 64; Err = 0.54687500 * 64; time = 0.0205s; samplesPerSecond = 3116.0
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 181- 181, 56.56%]: CE = 2.93763654 * 64; Err = 0.75000000 * 64; time = 0.1794s; samplesPerSecond = 356.7
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 182- 182, 56.88%]: CE = 2.93634742 * 64; Err = 0.73437500 * 64; time = 0.0206s; samplesPerSecond = 3111.2
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 183- 183, 57.19%]: CE = 2.59901571 * 64; Err = 0.68750000 * 64; time = 0.0218s; samplesPerSecond = 2929.9
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 184- 184, 57.50%]: CE = 2.81753002 * 64; Err = 0.73437500 * 64; time = 0.0200s; samplesPerSecond = 3204.8
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 185- 185, 57.81%]: CE = 3.04424260 * 64; Err = 0.73437500 * 64; time = 0.0218s; samplesPerSecond = 2937.8
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 186- 186, 58.13%]: CE = 2.49622625 * 64; Err = 0.64062500 * 64; time = 0.0242s; samplesPerSecond = 2646.4
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 187- 187, 58.44%]: CE = 2.94745408 * 64; Err = 0.71875000 * 64; time = 0.0248s; samplesPerSecond = 2584.7
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 188- 188, 58.75%]: CE = 2.80802583 * 64; Err = 0.71875000 * 64; time = 0.0239s; samplesPerSecond = 2677.4
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 189- 189, 59.06%]: CE = 2.54977638 * 64; Err = 0.67187500 * 64; time = 0.0252s; samplesPerSecond = 2543.2
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 190- 190, 59.38%]: CE = 2.90849909 * 64; Err = 0.68750000 * 64; time = 0.0241s; samplesPerSecond = 2660.9
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 191- 191, 59.69%]: CE = 2.89470021 * 64; Err = 0.71875000 * 64; time = 0.0252s; samplesPerSecond = 2542.9
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 192- 192, 60.00%]: CE = 2.55056761 * 64; Err = 0.64062500 * 64; time = 0.0240s; samplesPerSecond = 2665.9
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 193- 193, 60.31%]: CE = 2.39014720 * 64; Err = 0.59375000 * 64; time = 0.0252s; samplesPerSecond = 2542.4
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 194- 194, 60.62%]: CE = 2.61720826 * 64; Err = 0.65625000 * 64; time = 0.0239s; samplesPerSecond = 2676.4
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 195- 195, 60.94%]: CE = 2.59802571 * 64; Err = 0.65625000 * 64; time = 0.0251s; samplesPerSecond = 2551.1
MPI Rank 0: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 196- 196, 61.25%]: CE = 2.94597696 * 64; Err = 0.75000000 * 64; time = 0.0243s; samplesPerSecond = 2631.8
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 197- 197, 61.56%]: CE = 2.79771307 * 64; Err = 0.75000000 * 64; time = 0.1759s; samplesPerSecond = 363.8
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 198- 198, 61.88%]: CE = 3.20417932 * 64; Err = 0.71875000 * 64; time = 0.0195s; samplesPerSecond = 3277.5
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 199- 199, 62.19%]: CE = 2.27155558 * 64; Err = 0.53125000 * 64; time = 0.0194s; samplesPerSecond = 3305.3
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 200- 200, 62.50%]: CE = 2.87449908 * 64; Err = 0.68750000 * 64; time = 0.0242s; samplesPerSecond = 2649.6
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 201- 201, 62.81%]: CE = 2.71210245 * 64; Err = 0.65625000 * 64; time = 0.0191s; samplesPerSecond = 3352.5
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 202- 202, 63.13%]: CE = 2.44766371 * 64; Err = 0.57812500 * 64; time = 0.0231s; samplesPerSecond = 2776.2
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 203- 203, 63.44%]: CE = 2.68243088 * 64; Err = 0.70312500 * 64; time = 0.0220s; samplesPerSecond = 2905.3
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 204- 204, 63.75%]: CE = 2.40962202 * 64; Err = 0.54687500 * 64; time = 0.0264s; samplesPerSecond = 2421.5
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 205- 205, 64.06%]: CE = 2.48400547 * 64; Err = 0.59375000 * 64; time = 0.0239s; samplesPerSecond = 2673.9
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 206- 206, 64.38%]: CE = 2.49121254 * 64; Err = 0.60937500 * 64; time = 0.0257s; samplesPerSecond = 2485.7
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 207- 207, 64.69%]: CE = 2.84691899 * 64; Err = 0.75000000 * 64; time = 0.0230s; samplesPerSecond = 2786.7
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 208- 208, 65.00%]: CE = 2.45273493 * 64; Err = 0.59375000 * 64; time = 0.0262s; samplesPerSecond = 2445.8
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 209- 209, 65.31%]: CE = 2.75036440 * 64; Err = 0.68750000 * 64; time = 0.0224s; samplesPerSecond = 2858.9
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 210- 210, 65.63%]: CE = 2.49555051 * 64; Err = 0.71875000 * 64; time = 0.0256s; samplesPerSecond = 2504.0
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 211- 211, 65.94%]: CE = 2.71109113 * 64; Err = 0.68750000 * 64; time = 0.0231s; samplesPerSecond = 2774.3
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 212- 212, 66.25%]: CE = 2.38218216 * 64; Err = 0.59375000 * 64; time = 0.0260s; samplesPerSecond = 2457.1
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 213- 213, 66.56%]: CE = 2.60308722 * 64; Err = 0.57812500 * 64; time = 0.0233s; samplesPerSecond = 2741.7
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 214- 214, 66.88%]: CE = 2.65611547 * 64; Err = 0.68750000 * 64; time = 0.1694s; samplesPerSecond = 377.8
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 215- 215, 67.19%]: CE = 2.49633370 * 64; Err = 0.57812500 * 64; time = 0.0369s; samplesPerSecond = 1733.3
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 216- 216, 67.50%]: CE = 2.23315412 * 64; Err = 0.60937500 * 64; time = 0.0172s; samplesPerSecond = 3714.0
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 217- 217, 67.81%]: CE = 2.94093183 * 64; Err = 0.73437500 * 64; time = 0.0281s; samplesPerSecond = 2276.7
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 218- 218, 68.13%]: CE = 2.69840742 * 64; Err = 0.65625000 * 64; time = 0.0215s; samplesPerSecond = 2978.1
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 219- 219, 68.44%]: CE = 2.57215231 * 64; Err = 0.60937500 * 64; time = 0.0278s; samplesPerSecond = 2298.4
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 220- 220, 68.75%]: CE = 2.57160696 * 64; Err = 0.68750000 * 64; time = 0.0229s; samplesPerSecond = 2790.7
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 221- 221, 69.06%]: CE = 2.57776681 * 64; Err = 0.65625000 * 64; time = 0.0281s; samplesPerSecond = 2280.9
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 222- 222, 69.38%]: CE = 2.32289644 * 64; Err = 0.57812500 * 64; time = 0.0279s; samplesPerSecond = 2296.4
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 223- 223, 69.69%]: CE = 2.66432343 * 64; Err = 0.70312500 * 64; time = 0.0250s; samplesPerSecond = 2559.3
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 224- 224, 70.00%]: CE = 2.20387606 * 64; Err = 0.65625000 * 64; time = 0.0239s; samplesPerSecond = 2681.6
MPI Rank 0: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 225- 225, 70.31%]: CE = 2.39888933 * 64; Err = 0.59375000 * 64; time = 0.0249s; samplesPerSecond = 2572.6
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 226- 226, 70.63%]: CE = 2.80393339 * 64; Err = 0.70312500 * 64; time = 0.0239s; samplesPerSecond = 2673.7
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 227- 227, 70.94%]: CE = 2.71082242 * 64; Err = 0.68750000 * 64; time = 0.0249s; samplesPerSecond = 2566.5
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 228- 228, 71.25%]: CE = 2.62244612 * 64; Err = 0.70312500 * 64; time = 0.0239s; samplesPerSecond = 2677.3
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 229- 229, 71.56%]: CE = 2.29777087 * 64; Err = 0.62500000 * 64; time = 0.0246s; samplesPerSecond = 2600.4
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 230- 230, 71.88%]: CE = 2.51121239 * 64; Err = 0.65625000 * 64; time = 0.0241s; samplesPerSecond = 2651.3
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 231- 231, 72.19%]: CE = 2.76103008 * 64; Err = 0.64062500 * 64; time = 0.1737s; samplesPerSecond = 368.5
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 232- 232, 72.50%]: CE = 3.01432561 * 64; Err = 0.76562500 * 64; time = 0.0277s; samplesPerSecond = 2314.5
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 233- 233, 72.81%]: CE = 2.99024474 * 64; Err = 0.76562500 * 64; time = 0.0228s; samplesPerSecond = 2808.5
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 234- 234, 73.13%]: CE = 2.86664042 * 64; Err = 0.81250000 * 64; time = 0.0263s; samplesPerSecond = 2437.4
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 235- 235, 73.44%]: CE = 2.60998588 * 64; Err = 0.67187500 * 64; time = 0.0223s; samplesPerSecond = 2865.7
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 236- 236, 73.75%]: CE = 2.18201917 * 64; Err = 0.53125000 * 64; time = 0.0290s; samplesPerSecond = 2206.0
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 237- 237, 74.06%]: CE = 2.17418609 * 64; Err = 0.57812500 * 64; time = 0.0214s; samplesPerSecond = 2991.2
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 238- 238, 74.38%]: CE = 2.25759717 * 64; Err = 0.64062500 * 64; time = 0.0314s; samplesPerSecond = 2039.8
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 239- 239, 74.69%]: CE = 2.17788677 * 64; Err = 0.60937500 * 64; time = 0.0244s; samplesPerSecond = 2624.5
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 240- 240, 75.00%]: CE = 2.20328249 * 64; Err = 0.54687500 * 64; time = 0.0237s; samplesPerSecond = 2701.2
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 241- 241, 75.31%]: CE = 2.60590014 * 64; Err = 0.60937500 * 64; time = 0.0246s; samplesPerSecond = 2602.5
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 242- 242, 75.63%]: CE = 2.09884739 * 64; Err = 0.56250000 * 64; time = 0.0255s; samplesPerSecond = 2513.8
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 243- 243, 75.94%]: CE = 2.10587746 * 64; Err = 0.54687500 * 64; time = 0.0249s; samplesPerSecond = 2572.3
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 244- 244, 76.25%]: CE = 2.64457627 * 64; Err = 0.73437500 * 64; time = 0.0271s; samplesPerSecond = 2357.4
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 245- 245, 76.56%]: CE = 2.47600990 * 64; Err = 0.64062500 * 64; time = 0.0255s; samplesPerSecond = 2510.3
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 246- 246, 76.88%]: CE = 2.88789135 * 64; Err = 0.68750000 * 64; time = 0.0323s; samplesPerSecond = 1983.6
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 247- 247, 77.19%]: CE = 2.51823068 * 64; Err = 0.56250000 * 64; time = 0.1575s; samplesPerSecond = 406.4
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 248- 248, 77.50%]: CE = 2.24877264 * 64; Err = 0.62500000 * 64; time = 0.0244s; samplesPerSecond = 2619.0
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 249- 249, 77.81%]: CE = 2.51043156 * 64; Err = 0.71875000 * 64; time = 0.0187s; samplesPerSecond = 3421.7
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 250- 250, 78.13%]: CE = 2.54234511 * 64; Err = 0.70312500 * 64; time = 0.0320s; samplesPerSecond = 1996.9
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 251- 251, 78.44%]: CE = 2.68548933 * 64; Err = 0.70312500 * 64; time = 0.0182s; samplesPerSecond = 3519.4
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 252- 252, 78.75%]: CE = 2.23175466 * 64; Err = 0.57812500 * 64; time = 0.0312s; samplesPerSecond = 2048.6
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 253- 253, 79.06%]: CE = 2.24553589 * 64; Err = 0.60937500 * 64; time = 0.0190s; samplesPerSecond = 3373.4
MPI Rank 0: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 254- 254, 79.38%]: CE = 2.28765068 * 64; Err = 0.62500000 * 64; time = 0.0305s; samplesPerSecond = 2097.3
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 255- 255, 79.69%]: CE = 2.54161451 * 64; Err = 0.62500000 * 64; time = 0.0192s; samplesPerSecond = 3339.1
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 256- 256, 80.00%]: CE = 2.35401834 * 64; Err = 0.59375000 * 64; time = 0.0327s; samplesPerSecond = 1959.1
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 257- 257, 80.31%]: CE = 2.18137731 * 64; Err = 0.56250000 * 64; time = 0.0248s; samplesPerSecond = 2575.7
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 258- 258, 80.63%]: CE = 2.51499174 * 64; Err = 0.59375000 * 64; time = 0.0257s; samplesPerSecond = 2494.3
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 259- 259, 80.94%]: CE = 2.12242410 * 64; Err = 0.65625000 * 64; time = 0.0247s; samplesPerSecond = 2592.9
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 260- 260, 81.25%]: CE = 2.57230724 * 64; Err = 0.68750000 * 64; time = 0.0275s; samplesPerSecond = 2328.7
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 261- 261, 81.56%]: CE = 2.24717210 * 64; Err = 0.57812500 * 64; time = 0.0229s; samplesPerSecond = 2789.2
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 262- 262, 81.88%]: CE = 2.46805084 * 64; Err = 0.60937500 * 64; time = 0.0311s; samplesPerSecond = 2056.7
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 263- 263, 82.19%]: CE = 1.94672270 * 64; Err = 0.48437500 * 64; time = 0.0243s; samplesPerSecond = 2630.1
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 264- 264, 82.50%]: CE = 2.34898229 * 64; Err = 0.67187500 * 64; time = 0.0259s; samplesPerSecond = 2472.9
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 265- 265, 82.81%]: CE = 2.19361248 * 64; Err = 0.57812500 * 64; time = 0.0246s; samplesPerSecond = 2599.5
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 266- 266, 83.13%]: CE = 1.96058399 * 64; Err = 0.46875000 * 64; time = 0.1748s; samplesPerSecond = 366.0
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 267- 267, 83.44%]: CE = 2.02827934 * 64; Err = 0.53125000 * 64; time = 0.0250s; samplesPerSecond = 2563.9
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 268- 268, 83.75%]: CE = 2.16395773 * 64; Err = 0.56250000 * 64; time = 0.0242s; samplesPerSecond = 2640.0
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 269- 269, 84.06%]: CE = 2.42837196 * 64; Err = 0.64062500 * 64; time = 0.0230s; samplesPerSecond = 2786.2
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 270- 270, 84.38%]: CE = 2.56277231 * 64; Err = 0.75000000 * 64; time = 0.0256s; samplesPerSecond = 2499.8
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 271- 271, 84.69%]: CE = 2.35831855 * 64; Err = 0.59375000 * 64; time = 0.0240s; samplesPerSecond = 2669.8
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 272- 272, 85.00%]: CE = 2.48323539 * 64; Err = 0.70312500 * 64; time = 0.0244s; samplesPerSecond = 2627.1
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 273- 273, 85.31%]: CE = 2.66412354 * 64; Err = 0.67187500 * 64; time = 0.0229s; samplesPerSecond = 2791.7
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 274- 274, 85.63%]: CE = 2.35827343 * 64; Err = 0.65625000 * 64; time = 0.0254s; samplesPerSecond = 2516.9
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 275- 275, 85.94%]: CE = 2.35993611 * 64; Err = 0.59375000 * 64; time = 0.0241s; samplesPerSecond = 2651.6
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 276- 276, 86.25%]: CE = 2.27682017 * 64; Err = 0.59375000 * 64; time = 0.0244s; samplesPerSecond = 2624.3
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 277- 277, 86.56%]: CE = 2.58742110 * 64; Err = 0.70312500 * 64; time = 0.0262s; samplesPerSecond = 2439.6
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 278- 278, 86.88%]: CE = 2.59364573 * 64; Err = 0.70312500 * 64; time = 0.0249s; samplesPerSecond = 2571.6
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 279- 279, 87.19%]: CE = 2.58154982 * 64; Err = 0.67187500 * 64; time = 0.0267s; samplesPerSecond = 2394.3
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 280- 280, 87.50%]: CE = 2.65251947 * 64; Err = 0.71875000 * 64; time = 0.0253s; samplesPerSecond = 2525.8
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 281- 281, 87.81%]: CE = 2.42794113 * 64; Err = 0.56250000 * 64; time = 0.0284s; samplesPerSecond = 2253.0
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 282- 282, 88.13%]: CE = 2.31306675 * 64; Err = 0.56250000 * 64; time = 0.0213s; samplesPerSecond = 3005.1
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 283- 283, 88.44%]: CE = 2.30780317 * 64; Err = 0.57812500 * 64; time = 0.0282s; samplesPerSecond = 2267.6
MPI Rank 0: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 284- 284, 88.75%]: CE = 2.20092907 * 64; Err = 0.71875000 * 64; time = 0.0214s; samplesPerSecond = 2990.4
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 285- 285, 89.06%]: CE = 2.37127008 * 64; Err = 0.60937500 * 64; time = 0.1698s; samplesPerSecond = 377.0
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 286- 286, 89.38%]: CE = 1.96581596 * 64; Err = 0.51562500 * 64; time = 0.0237s; samplesPerSecond = 2702.5
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 287- 287, 89.69%]: CE = 2.38139796 * 64; Err = 0.68750000 * 64; time = 0.0195s; samplesPerSecond = 3284.9
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 288- 288, 90.00%]: CE = 2.17378766 * 64; Err = 0.56250000 * 64; time = 0.0303s; samplesPerSecond = 2109.3
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 289- 289, 90.31%]: CE = 2.43769870 * 64; Err = 0.62500000 * 64; time = 0.0192s; samplesPerSecond = 3326.2
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 290- 290, 90.63%]: CE = 1.92877315 * 64; Err = 0.48437500 * 64; time = 0.0306s; samplesPerSecond = 2089.9
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 291- 291, 90.94%]: CE = 2.40592700 * 64; Err = 0.62500000 * 64; time = 0.0197s; samplesPerSecond = 3250.9
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 292- 292, 91.25%]: CE = 2.08578061 * 64; Err = 0.59375000 * 64; time = 0.0311s; samplesPerSecond = 2054.8
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 293- 293, 91.56%]: CE = 2.00803832 * 64; Err = 0.51562500 * 64; time = 0.0201s; samplesPerSecond = 3181.5
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 294- 294, 91.88%]: CE = 2.17692353 * 64; Err = 0.57812500 * 64; time = 0.0298s; samplesPerSecond = 2145.3
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 295- 295, 92.19%]: CE = 2.50142509 * 64; Err = 0.70312500 * 64; time = 0.0203s; samplesPerSecond = 3159.2
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 296- 296, 92.50%]: CE = 2.23106504 * 64; Err = 0.60937500 * 64; time = 0.0296s; samplesPerSecond = 2161.0
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 297- 297, 92.81%]: CE = 2.15600594 * 64; Err = 0.59375000 * 64; time = 0.0194s; samplesPerSecond = 3293.2
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 298- 298, 93.13%]: CE = 2.57861376 * 64; Err = 0.68750000 * 64; time = 0.0318s; samplesPerSecond = 2009.7
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 299- 299, 93.44%]: CE = 2.07193617 * 64; Err = 0.56250000 * 64; time = 0.0197s; samplesPerSecond = 3248.9
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 300- 300, 93.75%]: CE = 2.16370481 * 64; Err = 0.60937500 * 64; time = 0.0305s; samplesPerSecond = 2099.0
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 301- 301, 94.06%]: CE = 2.24899831 * 64; Err = 0.56250000 * 64; time = 0.0197s; samplesPerSecond = 3252.2
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 302- 302, 94.38%]: CE = 1.87617314 * 64; Err = 0.54687500 * 64; time = 0.0304s; samplesPerSecond = 2102.8
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 303- 303, 94.69%]: CE = 2.22035878 * 64; Err = 0.56250000 * 64; time = 0.0183s; samplesPerSecond = 3502.8
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 304- 304, 95.00%]: CE = 2.23859583 * 64; Err = 0.65625000 * 64; time = 0.1704s; samplesPerSecond = 375.5
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 305- 305, 95.31%]: CE = 2.36221656 * 64; Err = 0.59375000 * 64; time = 0.0189s; samplesPerSecond = 3383.6
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 306- 306, 95.63%]: CE = 2.11637634 * 64; Err = 0.54687500 * 64; time = 0.0220s; samplesPerSecond = 2904.9
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 307- 307, 95.94%]: CE = 2.32528810 * 64; Err = 0.57812500 * 64; time = 0.0215s; samplesPerSecond = 2974.5
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 308- 308, 96.25%]: CE = 2.06869602 * 64; Err = 0.50000000 * 64; time = 0.0211s; samplesPerSecond = 3032.3
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 309- 309, 96.56%]: CE = 2.10471025 * 64; Err = 0.56250000 * 64; time = 0.0282s; samplesPerSecond = 2271.0
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 310- 310, 96.88%]: CE = 2.69881704 * 64; Err = 0.71875000 * 64; time = 0.0173s; samplesPerSecond = 3709.5
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 311- 311, 97.19%]: CE = 2.21301732 * 64; Err = 0.65625000 * 64; time = 0.0128s; samplesPerSecond = 4988.3
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 312- 312, 97.50%]: CE = 2.34597297 * 64; Err = 0.60937500 * 64; time = 0.0120s; samplesPerSecond = 5348.0
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 313- 313, 97.81%]: CE = 2.08858265 * 64; Err = 0.57812500 * 64; time = 0.0125s; samplesPerSecond = 5115.5
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 314- 314, 98.13%]: CE = 2.10805385 * 64; Err = 0.54687500 * 64; time = 0.0120s; samplesPerSecond = 5348.9
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 315- 315, 98.44%]: CE = 2.29975623 * 64; Err = 0.60937500 * 64; time = 0.0121s; samplesPerSecond = 5291.9
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 316- 316, 98.75%]: CE = 2.29188916 * 64; Err = 0.60937500 * 64; time = 0.0122s; samplesPerSecond = 5261.9
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 317- 317, 99.06%]: CE = 2.03062764 * 64; Err = 0.50000000 * 64; time = 0.0122s; samplesPerSecond = 5246.8
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 318- 318, 99.38%]: CE = 2.29874982 * 64; Err = 0.59375000 * 64; time = 0.0119s; samplesPerSecond = 5394.0
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 319- 319, 99.69%]: CE = 2.22342700 * 64; Err = 0.68750000 * 64; time = 0.0121s; samplesPerSecond = 5271.8
MPI Rank 0: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 320- 320, 100.00%]: CE = 2.32233814 * 64; Err = 0.59375000 * 64; time = 0.0118s; samplesPerSecond = 5416.4
MPI Rank 0: 12/21/2016 05:27:59: Finished Epoch[ 1 of 5]: [Training] CE = 3.02444900 * 20480; Err = 0.72885742 * 20480; totalSamplesSeen = 20480; learningRatePerSample = 0.015625; epochTime=10.3028s
MPI Rank 0: 12/21/2016 05:28:00: SGD: Saving checkpoint model 'C:\Users\svcphil\AppData\Local\Temp\cntk-test-20161221052142.365503\Speech\HTKDeserializers\DNN_ParallelBMWithAdjustLR@release_cpu/models/cntkSpeech.dnn.1'
MPI Rank 0: 
MPI Rank 0: 12/21/2016 05:28:00: Starting Epoch 2: learning rate per sample = 0.001953  effective momentum = 0.656119  momentum as time constant = 607.5 samples
MPI Rank 0: Parallel training (2 workers) using BlockMomentumSGD with block momentum = 0.5000, block momentum time constant (per worker) = 2954.6394, block learning rate = 1.0000, block size per worker = 2048 samples, using Nesterov-style block momentum, resetting SGD momentum after sync.
MPI Rank 0: 
MPI Rank 0: 12/21/2016 05:28:00: Starting minibatch loop, distributed reading is ENABLED.
MPI Rank 0: 12/21/2016 05:28:00:  Epoch[ 2 of 5]-Minibatch[   1-   1, 1.25%]: CE = 2.04621069 * 128; Err = 0.56250000 * 128; time = 0.0347s; samplesPerSecond = 3687.6
MPI Rank 0: 12/21/2016 05:28:00:  Epoch[ 2 of 5]-Minibatch[   2-   2, 2.50%]: CE = 2.17708778 * 128; Err = 0.58593750 * 128; time = 0.1858s; samplesPerSecond = 689.0
MPI Rank 0: 12/21/2016 05:28:00:  Epoch[ 2 of 5]-Minibatch[   3-   3, 3.75%]: CE = 2.34813693 * 128; Err = 0.65625000 * 128; time = 0.0482s; samplesPerSecond = 2654.4
MPI Rank 0: 12/21/2016 05:28:00:  Epoch[ 2 of 5]-Minibatch[   4-   4, 5.00%]: CE = 2.32449083 * 128; Err = 0.64062500 * 128; time = 0.0577s; samplesPerSecond = 2216.6
MPI Rank 0: 12/21/2016 05:28:00:  Epoch[ 2 of 5]-Minibatch[   5-   5, 6.25%]: CE = 1.88510886 * 128; Err = 0.50781250 * 128; time = 0.0358s; samplesPerSecond = 3578.8
MPI Rank 0: 12/21/2016 05:28:00:  Epoch[ 2 of 5]-Minibatch[   6-   6, 7.50%]: CE = 2.37002367 * 128; Err = 0.60156250 * 128; time = 0.0592s; samplesPerSecond = 2164.0
MPI Rank 0: 12/21/2016 05:28:00:  Epoch[ 2 of 5]-Minibatch[   7-   7, 8.75%]: CE = 2.26305040 * 128; Err = 0.61718750 * 128; time = 0.0343s; samplesPerSecond = 3728.2
MPI Rank 0: 12/21/2016 05:28:00:  Epoch[ 2 of 5]-Minibatch[   8-   8, 10.00%]: CE = 2.29855550 * 128; Err = 0.64062500 * 128; time = 0.0551s; samplesPerSecond = 2323.0
MPI Rank 0: 12/21/2016 05:28:00:  Epoch[ 2 of 5]-Minibatch[   9-   9, 11.25%]: CE = 2.46680236 * 128; Err = 0.68750000 * 128; time = 0.0407s; samplesPerSecond = 3141.6
MPI Rank 0: 12/21/2016 05:28:00:  Epoch[ 2 of 5]-Minibatch[  10-  10, 12.50%]: CE = 1.99182081 * 128; Err = 0.52343750 * 128; time = 0.0426s; samplesPerSecond = 3006.7
MPI Rank 0: 12/21/2016 05:28:00:  Epoch[ 2 of 5]-Minibatch[  11-  11, 13.75%]: CE = 1.94202161 * 128; Err = 0.57031250 * 128; time = 0.0482s; samplesPerSecond = 2658.3
MPI Rank 0: 12/21/2016 05:28:00:  Epoch[ 2 of 5]-Minibatch[  12-  12, 15.00%]: CE = 2.20437533 * 128; Err = 0.58593750 * 128; time = 0.1527s; samplesPerSecond = 838.1
MPI Rank 0: 12/21/2016 05:28:00:  Epoch[ 2 of 5]-Minibatch[  13-  13, 16.25%]: CE = 2.31128223 * 128; Err = 0.60937500 * 128; time = 0.0462s; samplesPerSecond = 2772.7
MPI Rank 0: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  14-  14, 17.50%]: CE = 2.29976865 * 128; Err = 0.62500000 * 128; time = 0.0456s; samplesPerSecond = 2804.3
MPI Rank 0: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  15-  15, 18.75%]: CE = 2.01466608 * 128; Err = 0.52343750 * 128; time = 0.0460s; samplesPerSecond = 2781.3
MPI Rank 0: 		(model aggregation stats): 1-th sync point was hit, introducing a 0.00-seconds latency this time; accumulated time on sync point = 0.00 seconds , average latency = 0.00 seconds
MPI Rank 0: 		(model aggregation stats) 1-th sync:     1.03 seconds since last report (0.02 seconds on comm.); 4096 samples processed by 2 workers (2048 by me);
MPI Rank 0: 		(model aggregation stats) 1-th sync: totalThroughput = 3.99k samplesPerSecond , throughputPerWorker = 2.00k samplesPerSecond
MPI Rank 0: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  16-  16, 20.00%]: CE = 2.24454636 * 128; Err = 0.59375000 * 128; time = 0.0785s; samplesPerSecond = 1631.4
MPI Rank 0: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  17-  17, 21.25%]: CE = 2.25498615 * 128; Err = 0.58593750 * 128; time = 0.0388s; samplesPerSecond = 3296.3
MPI Rank 0: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  18-  18, 22.50%]: CE = 2.07039205 * 128; Err = 0.57812500 * 128; time = 0.0533s; samplesPerSecond = 2402.8
MPI Rank 0: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  19-  19, 23.75%]: CE = 1.54332140 * 128; Err = 0.39843750 * 128; time = 0.0454s; samplesPerSecond = 2820.2
MPI Rank 0: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  20-  20, 25.00%]: CE = 2.28595867 * 128; Err = 0.61718750 * 128; time = 0.0453s; samplesPerSecond = 2823.6
MPI Rank 0: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  21-  21, 26.25%]: CE = 2.00097462 * 128; Err = 0.55468750 * 128; time = 0.0455s; samplesPerSecond = 2814.4
MPI Rank 0: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  22-  22, 27.50%]: CE = 1.89541624 * 128; Err = 0.49218750 * 128; time = 0.1309s; samplesPerSecond = 978.0
MPI Rank 0: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  23-  23, 28.75%]: CE = 2.17076816 * 128; Err = 0.60156250 * 128; time = 0.0404s; samplesPerSecond = 3170.1
MPI Rank 0: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  24-  24, 30.00%]: CE = 2.14484872 * 128; Err = 0.52343750 * 128; time = 0.0461s; samplesPerSecond = 2775.1
MPI Rank 0: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  25-  25, 31.25%]: CE = 1.98272942 * 128; Err = 0.51562500 * 128; time = 0.0493s; samplesPerSecond = 2597.2
MPI Rank 0: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  26-  26, 32.50%]: CE = 2.15380788 * 128; Err = 0.53906250 * 128; time = 0.0452s; samplesPerSecond = 2833.7
MPI Rank 0: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  27-  27, 33.75%]: CE = 2.00961418 * 128; Err = 0.57031250 * 128; time = 0.0469s; samplesPerSecond = 2726.9
MPI Rank 0: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  28-  28, 35.00%]: CE = 1.99086509 * 128; Err = 0.48437500 * 128; time = 0.0569s; samplesPerSecond = 2249.2
MPI Rank 0: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  29-  29, 36.25%]: CE = 2.30500580 * 128; Err = 0.60156250 * 128; time = 0.0461s; samplesPerSecond = 2776.2
MPI Rank 0: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  30-  30, 37.50%]: CE = 2.22538862 * 128; Err = 0.58593750 * 128; time = 0.0448s; samplesPerSecond = 2860.3
MPI Rank 0: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  31-  31, 38.75%]: CE = 2.02304991 * 128; Err = 0.51562500 * 128; time = 0.0463s; samplesPerSecond = 2762.4
MPI Rank 0: 		(model aggregation stats): 2-th sync point was hit, introducing a 0.00-seconds latency this time; accumulated time on sync point = 0.00 seconds , average latency = 0.00 seconds
MPI Rank 0: 		(model aggregation stats) 2-th sync:     1.01 seconds since last report (0.15 seconds on comm.); 4096 samples processed by 2 workers (2048 by me);
MPI Rank 0: 		(model aggregation stats) 2-th sync: totalThroughput = 4.06k samplesPerSecond , throughputPerWorker = 2.03k samplesPerSecond
MPI Rank 0: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  32-  32, 40.00%]: CE = 2.08122418 * 128; Err = 0.54687500 * 128; time = 0.2202s; samplesPerSecond = 581.4
MPI Rank 0: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  33-  33, 41.25%]: CE = 2.11152507 * 128; Err = 0.53906250 * 128; time = 0.0359s; samplesPerSecond = 3567.2
MPI Rank 0: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  34-  34, 42.50%]: CE = 1.95276514 * 128; Err = 0.57031250 * 128; time = 0.0464s; samplesPerSecond = 2761.6
MPI Rank 0: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  35-  35, 43.75%]: CE = 1.97726051 * 128; Err = 0.56250000 * 128; time = 0.0497s; samplesPerSecond = 2573.9
MPI Rank 0: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  36-  36, 45.00%]: CE = 2.17844485 * 128; Err = 0.61718750 * 128; time = 0.0517s; samplesPerSecond = 2476.1
MPI Rank 0: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  37-  37, 46.25%]: CE = 1.98699662 * 128; Err = 0.53906250 * 128; time = 0.0484s; samplesPerSecond = 2646.5
MPI Rank 0: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  38-  38, 47.50%]: CE = 1.99237188 * 128; Err = 0.53906250 * 128; time = 0.0458s; samplesPerSecond = 2795.7
MPI Rank 0: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  39-  39, 48.75%]: CE = 2.01392744 * 128; Err = 0.52343750 * 128; time = 0.0280s; samplesPerSecond = 4565.2
MPI Rank 0: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  40-  40, 50.00%]: CE = 1.95863719 * 128; Err = 0.53906250 * 128; time = 0.0461s; samplesPerSecond = 2775.5
MPI Rank 0: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  41-  41, 51.25%]: CE = 2.04755340 * 128; Err = 0.53906250 * 128; time = 0.0277s; samplesPerSecond = 4613.3
MPI Rank 0: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  42-  42, 52.50%]: CE = 2.17596474 * 128; Err = 0.64843750 * 128; time = 0.0448s; samplesPerSecond = 2859.5
MPI Rank 0: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  43-  43, 53.75%]: CE = 2.22751313 * 128; Err = 0.63281250 * 128; time = 0.0277s; samplesPerSecond = 4617.3
MPI Rank 0: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  44-  44, 55.00%]: CE = 2.01731189 * 128; Err = 0.52343750 * 128; time = 0.1687s; samplesPerSecond = 758.6
MPI Rank 0: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  45-  45, 56.25%]: CE = 1.96079757 * 128; Err = 0.58593750 * 128; time = 0.0277s; samplesPerSecond = 4622.8
MPI Rank 0: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  46-  46, 57.50%]: CE = 2.18325683 * 128; Err = 0.56250000 * 128; time = 0.0461s; samplesPerSecond = 2777.2
MPI Rank 0: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  47-  47, 58.75%]: CE = 2.08112152 * 128; Err = 0.57812500 * 128; time = 0.0279s; samplesPerSecond = 4594.9
MPI Rank 0: 		(model aggregation stats): 3-th sync point was hit, introducing a 0.00-seconds latency this time; accumulated time on sync point = 0.00 seconds , average latency = 0.00 seconds
MPI Rank 0: 		(model aggregation stats) 3-th sync:     0.85 seconds since last report (0.06 seconds on comm.); 4096 samples processed by 2 workers (2048 by me);
MPI Rank 0: 		(model aggregation stats) 3-th sync: totalThroughput = 4.80k samplesPerSecond , throughputPerWorker = 2.40k samplesPerSecond
MPI Rank 0: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  48-  48, 60.00%]: CE = 1.89594937 * 128; Err = 0.52343750 * 128; time = 0.1241s; samplesPerSecond = 1031.2
MPI Rank 0: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  49-  49, 61.25%]: CE = 1.93867044 * 128; Err = 0.51562500 * 128; time = 0.0360s; samplesPerSecond = 3558.5
MPI Rank 0: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  50-  50, 62.50%]: CE = 1.94696868 * 128; Err = 0.54687500 * 128; time = 0.0554s; samplesPerSecond = 2308.5
MPI Rank 0: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  51-  51, 63.75%]: CE = 1.89376884 * 128; Err = 0.55468750 * 128; time = 0.0437s; samplesPerSecond = 2928.3
MPI Rank 0: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  52-  52, 65.00%]: CE = 1.97393916 * 128; Err = 0.57031250 * 128; time = 0.0433s; samplesPerSecond = 2954.4
MPI Rank 0: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  53-  53, 66.25%]: CE = 2.16872511 * 128; Err = 0.53906250 * 128; time = 0.1541s; samplesPerSecond = 830.7
MPI Rank 0: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  54-  54, 67.50%]: CE = 2.15993970 * 128; Err = 0.60937500 * 128; time = 0.0579s; samplesPerSecond = 2210.6
MPI Rank 0: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  55-  55, 68.75%]: CE = 2.00906572 * 128; Err = 0.53125000 * 128; time = 0.0801s; samplesPerSecond = 1597.9
MPI Rank 0: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  56-  56, 70.00%]: CE = 1.89000808 * 128; Err = 0.54687500 * 128; time = 0.0468s; samplesPerSecond = 2733.9
MPI Rank 0: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  57-  57, 71.25%]: CE = 1.90423684 * 128; Err = 0.55468750 * 128; time = 0.0238s; samplesPerSecond = 5378.6
MPI Rank 0: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  58-  58, 72.50%]: CE = 1.91510403 * 128; Err = 0.54687500 * 128; time = 0.0215s; samplesPerSecond = 5952.4
MPI Rank 0: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  59-  59, 73.75%]: CE = 2.00206690 * 128; Err = 0.53125000 * 128; time = 0.0229s; samplesPerSecond = 5592.0
MPI Rank 0: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  60-  60, 75.00%]: CE = 2.04667281 * 128; Err = 0.56250000 * 128; time = 0.0215s; samplesPerSecond = 5941.1
MPI Rank 0: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  61-  61, 76.25%]: CE = 1.97803721 * 128; Err = 0.51562500 * 128; time = 0.0218s; samplesPerSecond = 5865.9
MPI Rank 0: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  62-  62, 77.50%]: CE = 1.83661828 * 128; Err = 0.52343750 * 128; time = 0.0212s; samplesPerSecond = 6039.4
MPI Rank 0: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  63-  63, 78.75%]: CE = 1.88327931 * 128; Err = 0.54687500 * 128; time = 0.0218s; samplesPerSecond = 5876.7
MPI Rank 0: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  64-  64, 80.00%]: CE = 2.11534140 * 128; Err = 0.60937500 * 128; time = 0.0212s; samplesPerSecond = 6030.3
MPI Rank 0: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  65-  65, 81.25%]: CE = 2.16230390 * 128; Err = 0.58593750 * 128; time = 0.0238s; samplesPerSecond = 5373.0
MPI Rank 0: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  66-  66, 82.50%]: CE = 2.13342620 * 128; Err = 0.61718750 * 128; time = 0.0212s; samplesPerSecond = 6028.6
MPI Rank 0: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  67-  67, 83.75%]: CE = 1.80170369 * 128; Err = 0.49218750 * 128; time = 0.0218s; samplesPerSecond = 5884.2
MPI Rank 0: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  68-  68, 85.00%]: CE = 2.13547951 * 128; Err = 0.59375000 * 128; time = 0.0214s; samplesPerSecond = 5978.8
MPI Rank 0: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  69-  69, 86.25%]: CE = 1.87771364 * 128; Err = 0.53906250 * 128; time = 0.0218s; samplesPerSecond = 5878.0
MPI Rank 0: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  70-  70, 87.50%]: CE = 1.80406584 * 128; Err = 0.50000000 * 128; time = 0.0949s; samplesPerSecond = 1348.3
MPI Rank 0: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  71-  71, 88.75%]: CE = 2.02216089 * 128; Err = 0.51562500 * 128; time = 0.0416s; samplesPerSecond = 3080.0
MPI Rank 0: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  72-  72, 90.00%]: CE = 2.05838722 * 128; Err = 0.55468750 * 128; time = 0.0277s; samplesPerSecond = 4614.4
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[  73-  73, 91.25%]: CE = 1.83577660 * 128; Err = 0.50000000 * 128; time = 0.0243s; samplesPerSecond = 5269.4
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[  74-  74, 92.50%]: CE = 1.95166728 * 128; Err = 0.53906250 * 128; time = 0.0239s; samplesPerSecond = 5358.3
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[  75-  75, 93.75%]: CE = 1.92254694 * 128; Err = 0.55468750 * 128; time = 0.0243s; samplesPerSecond = 5267.7
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[  76-  76, 95.00%]: CE = 1.91965482 * 128; Err = 0.52343750 * 128; time = 0.0238s; samplesPerSecond = 5374.1
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[  77-  77, 96.25%]: CE = 1.95992305 * 128; Err = 0.50781250 * 128; time = 0.0243s; samplesPerSecond = 5264.7
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[  78-  78, 97.50%]: CE = 2.11772801 * 128; Err = 0.58593750 * 128; time = 0.0238s; samplesPerSecond = 5375.2
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[  79-  79, 98.75%]: CE = 1.87594063 * 128; Err = 0.48437500 * 128; time = 0.0243s; samplesPerSecond = 5269.7
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[  80-  80, 100.00%]: CE = 1.93475436 * 128; Err = 0.51562500 * 128; time = 0.0238s; samplesPerSecond = 5367.3
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[  81-  81, 101.25%]: CE = 1.98965998 * 128; Err = 0.52343750 * 128; time = 0.0236s; samplesPerSecond = 5431.3
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[  82-  82, 102.50%]: CE = 1.96250616 * 128; Err = 0.58593750 * 128; time = 0.0241s; samplesPerSecond = 5317.6
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[  83-  83, 103.75%]: CE = 1.89156264 * 128; Err = 0.55468750 * 128; time = 0.0234s; samplesPerSecond = 5461.2
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[  84-  84, 105.00%]: CE = 1.98132112 * 128; Err = 0.55468750 * 128; time = 0.0230s; samplesPerSecond = 5573.9
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[  85-  85, 106.25%]: CE = 2.04288157 * 128; Err = 0.53125000 * 128; time = 0.0241s; samplesPerSecond = 5305.3
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[  86-  86, 107.50%]: CE = 2.01657890 * 128; Err = 0.58593750 * 128; time = 0.0232s; samplesPerSecond = 5520.3
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[  87-  87, 108.75%]: CE = 2.05797887 * 128; Err = 0.52343750 * 128; time = 0.0238s; samplesPerSecond = 5378.6
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[  88-  88, 110.00%]: CE = 1.87100099 * 128; Err = 0.52343750 * 128; time = 0.0231s; samplesPerSecond = 5538.2
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[  89-  89, 111.25%]: CE = 2.03788831 * 128; Err = 0.56250000 * 128; time = 0.0223s; samplesPerSecond = 5740.2
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[  90-  90, 112.50%]: CE = 1.98743413 * 128; Err = 0.57031250 * 128; time = 0.0217s; samplesPerSecond = 5906.2
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[  91-  91, 113.75%]: CE = 2.02666813 * 128; Err = 0.49218750 * 128; time = 0.0226s; samplesPerSecond = 5651.5
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[  92-  92, 115.00%]: CE = 1.96605680 * 128; Err = 0.53125000 * 128; time = 0.1157s; samplesPerSecond = 1106.2
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[  93-  93, 116.25%]: CE = 2.07576441 * 128; Err = 0.59375000 * 128; time = 0.0392s; samplesPerSecond = 3262.1
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[  94-  94, 117.50%]: CE = 1.90720948 * 128; Err = 0.58593750 * 128; time = 0.0282s; samplesPerSecond = 4541.1
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[  95-  95, 118.75%]: CE = 2.00432920 * 128; Err = 0.57031250 * 128; time = 0.0214s; samplesPerSecond = 5978.2
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[  96-  96, 120.00%]: CE = 1.88575775 * 128; Err = 0.52343750 * 128; time = 0.0238s; samplesPerSecond = 5369.4
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[  97-  97, 121.25%]: CE = 1.88803903 * 128; Err = 0.58593750 * 128; time = 0.0239s; samplesPerSecond = 5360.8
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[  98-  98, 122.50%]: CE = 1.72565000 * 128; Err = 0.49218750 * 128; time = 0.0240s; samplesPerSecond = 5327.1
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[  99-  99, 123.75%]: CE = 2.05912371 * 128; Err = 0.53125000 * 128; time = 0.0222s; samplesPerSecond = 5772.0
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[ 100- 100, 125.00%]: CE = 2.00262941 * 128; Err = 0.50000000 * 128; time = 0.0224s; samplesPerSecond = 5724.5
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[ 101- 101, 126.25%]: CE = 2.01561997 * 128; Err = 0.53906250 * 128; time = 0.0225s; samplesPerSecond = 5682.1
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[ 102- 102, 127.50%]: CE = 1.99238381 * 128; Err = 0.54687500 * 128; time = 0.0225s; samplesPerSecond = 5677.5
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[ 103- 103, 128.75%]: CE = 1.88524397 * 128; Err = 0.49218750 * 128; time = 0.0220s; samplesPerSecond = 5828.5
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[ 104- 104, 130.00%]: CE = 2.03128662 * 91; Err = 0.51648352 * 91; time = 0.0154s; samplesPerSecond = 5917.5
MPI Rank 0: 12/21/2016 05:28:04:  Epoch[ 2 of 5]-Minibatch[ 105- 105, 131.25%]: CE = 0.00000000 * 0; Err = 0.00000000 * 0; time = 3.7000e-005s; samplesPerSecond = 0.0
MPI Rank 0: 		(model aggregation stats): 4-th sync point was hit, introducing a 0.00-seconds latency this time; accumulated time on sync point = 0.00 seconds , average latency = 0.00 seconds
MPI Rank 0: 		(model aggregation stats) 4-th sync:     1.92 seconds since last report (0.06 seconds on comm.); 8192 samples processed by 2 workers (7131 by me);
MPI Rank 0: 		(model aggregation stats) 4-th sync: totalThroughput = 4.27k samplesPerSecond , throughputPerWorker = 2.13k samplesPerSecond
MPI Rank 0: 12/21/2016 05:28:04: Finished Epoch[ 2 of 5]: [Training] CE = 2.03791679 * 20480; Err = 0.55712891 * 20480; totalSamplesSeen = 40960; learningRatePerSample = 0.001953125; epochTime=4.81786s
MPI Rank 0: 12/21/2016 05:28:05: SGD: Saving checkpoint model 'C:\Users\svcphil\AppData\Local\Temp\cntk-test-20161221052142.365503\Speech\HTKDeserializers\DNN_ParallelBMWithAdjustLR@release_cpu/models/cntkSpeech.dnn.2'
MPI Rank 0: 
MPI Rank 0: 12/21/2016 05:28:05: Starting Epoch 3: learning rate per sample = 0.000098  effective momentum = 0.656119  momentum as time constant = 2429.9 samples
MPI Rank 0: Parallel training (2 workers) using BlockMomentumSGD with block momentum = 0.5000, block momentum time constant (per worker) = 2954.6394, block learning rate = 1.0000, block size per worker = 2048 samples, using Nesterov-style block momentum, resetting SGD momentum after sync.
MPI Rank 0: 
MPI Rank 0: 12/21/2016 05:28:05: Starting minibatch loop, distributed reading is ENABLED.
MPI Rank 0: 12/21/2016 05:28:05:  Epoch[ 3 of 5]-Minibatch[   1-   1, 5.00%]: CE = 1.92970354 * 512; Err = 0.50585938 * 512; time = 0.1760s; samplesPerSecond = 2908.6
MPI Rank 0: 12/21/2016 05:28:05:  Epoch[ 3 of 5]-Minibatch[   2-   2, 10.00%]: CE = 1.92764867 * 512; Err = 0.52343750 * 512; time = 0.1552s; samplesPerSecond = 3299.5
MPI Rank 0: 12/21/2016 05:28:05:  Epoch[ 3 of 5]-Minibatch[   3-   3, 15.00%]: CE = 2.00687305 * 512; Err = 0.55078125 * 512; time = 0.1570s; samplesPerSecond = 3261.7
MPI Rank 0: 		(model aggregation stats): 1-th sync point was hit, introducing a 0.00-seconds latency this time; accumulated time on sync point = 0.00 seconds , average latency = 0.00 seconds
MPI Rank 0: 		(model aggregation stats) 1-th sync:     0.83 seconds since last report (0.06 seconds on comm.); 4096 samples processed by 2 workers (2048 by me);
MPI Rank 0: 		(model aggregation stats) 1-th sync: totalThroughput = 4.94k samplesPerSecond , throughputPerWorker = 2.47k samplesPerSecond
MPI Rank 0: 12/21/2016 05:28:05:  Epoch[ 3 of 5]-Minibatch[   4-   4, 20.00%]: CE = 1.87892688 * 512; Err = 0.50585938 * 512; time = 0.3316s; samplesPerSecond = 1543.9
MPI Rank 0: 12/21/2016 05:28:06:  Epoch[ 3 of 5]-Minibatch[   5-   5, 25.00%]: CE = 1.97706875 * 512; Err = 0.54687500 * 512; time = 0.1549s; samplesPerSecond = 3305.4
MPI Rank 0: 12/21/2016 05:28:06:  Epoch[ 3 of 5]-Minibatch[   6-   6, 30.00%]: CE = 1.98393011 * 512; Err = 0.56640625 * 512; time = 0.1459s; samplesPerSecond = 3508.4
MPI Rank 0: 12/21/2016 05:28:06:  Epoch[ 3 of 5]-Minibatch[   7-   7, 35.00%]: CE = 1.92396216 * 512; Err = 0.54101563 * 512; time = 0.2098s; samplesPerSecond = 2440.7
MPI Rank 0: 		(model aggregation stats): 2-th sync point was hit, introducing a 0.05-seconds latency this time; accumulated time on sync point = 0.05 seconds , average latency = 0.02 seconds
MPI Rank 0: 		(model aggregation stats) 2-th sync:     0.80 seconds since last report (0.07 seconds on comm.); 4096 samples processed by 2 workers (2048 by me);
MPI Rank 0: 		(model aggregation stats) 2-th sync: totalThroughput = 5.15k samplesPerSecond , throughputPerWorker = 2.57k samplesPerSecond
MPI Rank 0: 12/21/2016 05:28:06:  Epoch[ 3 of 5]-Minibatch[   8-   8, 40.00%]: CE = 1.90143255 * 512; Err = 0.54296875 * 512; time = 0.2838s; samplesPerSecond = 1803.8
MPI Rank 0: 12/21/2016 05:28:06:  Epoch[ 3 of 5]-Minibatch[   9-   9, 45.00%]: CE = 1.90191854 * 512; Err = 0.53320313 * 512; time = 0.1278s; samplesPerSecond = 4007.5
MPI Rank 0: 12/21/2016 05:28:07:  Epoch[ 3 of 5]-Minibatch[  10-  10, 50.00%]: CE = 2.02959458 * 512; Err = 0.55664063 * 512; time = 0.2683s; samplesPerSecond = 1908.4
MPI Rank 0: 12/21/2016 05:28:07:  Epoch[ 3 of 5]-Minibatch[  11-  11, 55.00%]: CE = 2.02800892 * 512; Err = 0.56445313 * 512; time = 0.1384s; samplesPerSecond = 3699.9
MPI Rank 0: 		(model aggregation stats): 3-th sync point was hit, introducing a 0.00-seconds latency this time; accumulated time on sync point = 0.05 seconds , average latency = 0.02 seconds
MPI Rank 0: 		(model aggregation stats) 3-th sync:     0.84 seconds since last report (0.13 seconds on comm.); 4096 samples processed by 2 workers (2048 by me);
MPI Rank 0: 		(model aggregation stats) 3-th sync: totalThroughput = 4.90k samplesPerSecond , throughputPerWorker = 2.45k samplesPerSecond
MPI Rank 0: 12/21/2016 05:28:07:  Epoch[ 3 of 5]-Minibatch[  12-  12, 60.00%]: CE = 1.90570964 * 512; Err = 0.54296875 * 512; time = 0.2995s; samplesPerSecond = 1709.5
MPI Rank 0: 12/21/2016 05:28:07:  Epoch[ 3 of 5]-Minibatch[  13-  13, 65.00%]: CE = 2.09045613 * 512; Err = 0.57031250 * 512; time = 0.0988s; samplesPerSecond = 5183.3
MPI Rank 0: 12/21/2016 05:28:07:  Epoch[ 3 of 5]-Minibatch[  14-  14, 70.00%]: CE = 1.99217449 * 512; Err = 0.55078125 * 512; time = 0.1446s; samplesPerSecond = 3540.2
MPI Rank 0: 12/21/2016 05:28:07:  Epoch[ 3 of 5]-Minibatch[  15-  15, 75.00%]: CE = 1.78680596 * 512; Err = 0.52734375 * 512; time = 0.1573s; samplesPerSecond = 3255.8
MPI Rank 0: 12/21/2016 05:28:08:  Epoch[ 3 of 5]-Minibatch[  16-  16, 80.00%]: CE = 1.93041321 * 512; Err = 0.55664063 * 512; time = 0.2401s; samplesPerSecond = 2132.6
MPI Rank 0: 12/21/2016 05:28:08:  Epoch[ 3 of 5]-Minibatch[  17-  17, 85.00%]: CE = 1.94051124 * 512; Err = 0.56445313 * 512; time = 0.0775s; samplesPerSecond = 6606.4
MPI Rank 0: 12/21/2016 05:28:08:  Epoch[ 3 of 5]-Minibatch[  18-  18, 90.00%]: CE = 1.99870822 * 512; Err = 0.58789063 * 512; time = 0.0719s; samplesPerSecond = 7125.7
MPI Rank 0: 12/21/2016 05:28:08:  Epoch[ 3 of 5]-Minibatch[  19-  19, 95.00%]: CE = 2.00679952 * 512; Err = 0.56835938 * 512; time = 0.0912s; samplesPerSecond = 5614.2
MPI Rank 0: 12/21/2016 05:28:08:  Epoch[ 3 of 5]-Minibatch[  20-  20, 100.00%]: CE = 1.92222693 * 512; Err = 0.54296875 * 512; time = 0.0794s; samplesPerSecond = 6447.5
MPI Rank 0: 12/21/2016 05:28:08:  Epoch[ 3 of 5]-Minibatch[  21-  21, 105.00%]: CE = 1.93006837 * 512; Err = 0.54687500 * 512; time = 0.1458s; samplesPerSecond = 3511.7
MPI Rank 0: 12/21/2016 05:28:08:  Epoch[ 3 of 5]-Minibatch[  22-  22, 110.00%]: CE = 1.95262711 * 512; Err = 0.53125000 * 512; time = 0.0828s; samplesPerSecond = 6183.0
MPI Rank 0: 12/21/2016 05:28:08:  Epoch[ 3 of 5]-Minibatch[  23-  23, 115.00%]: CE = 1.99713868 * 512; Err = 0.54687500 * 512; time = 0.0844s; samplesPerSecond = 6062.8
MPI Rank 0: 12/21/2016 05:28:08:  Epoch[ 3 of 5]-Minibatch[  24-  24, 120.00%]: CE = 2.00598038 * 512; Err = 0.55664063 * 512; time = 0.0805s; samplesPerSecond = 6362.1
MPI Rank 0: 12/21/2016 05:28:09:  Epoch[ 3 of 5]-Minibatch[  25-  25, 125.00%]: CE = 1.91932725 * 501; Err = 0.56686627 * 501; time = 0.0779s; samplesPerSecond = 6434.9
MPI Rank 0: 12/21/2016 05:28:09:  Epoch[ 3 of 5]-Minibatch[  26-  26, 130.00%]: CE = 0.00000000 * 0; Err = 0.00000000 * 0; time = 3.8000e-005s; samplesPerSecond = 0.0
MPI Rank 0: 		(model aggregation stats): 4-th sync point was hit, introducing a 0.00-seconds latency this time; accumulated time on sync point = 0.05 seconds , average latency = 0.01 seconds
MPI Rank 0: 		(model aggregation stats) 4-th sync:     1.46 seconds since last report (0.01 seconds on comm.); 8192 samples processed by 2 workers (6645 by me);
MPI Rank 0: 		(model aggregation stats) 4-th sync: totalThroughput = 5.60k samplesPerSecond , throughputPerWorker = 2.80k samplesPerSecond
MPI Rank 0: 12/21/2016 05:28:09: Finished Epoch[ 3 of 5]: [Training] CE = 1.94906734 * 20480; Err = 0.54541016 * 20480; totalSamplesSeen = 61440; learningRatePerSample = 9.7656251e-005; epochTime=3.93516s
MPI Rank 0: 12/21/2016 05:28:09: SGD: Saving checkpoint model 'C:\Users\svcphil\AppData\Local\Temp\cntk-test-20161221052142.365503\Speech\HTKDeserializers\DNN_ParallelBMWithAdjustLR@release_cpu/models/cntkSpeech.dnn.3'
MPI Rank 0: ValidateSubNetwork: featNorm.invStdDev InvStdDev operation changed, from [363 x 1] to [363].ValidateSubNetwork: featNorm.mean Mean operation changed, from [363 x 1] to [363].ValidateSubNetwork: logPrior._ Mean operation changed, from [132 x 1] to [132].Parallel training (2 workers) using BlockMomentumSGD with block momentum = 0.5000, block momentum time constant (per worker) = 2954.6394, block learning rate = 1.0000, block size per worker = 2048 samples, using Nesterov-style block momentum, resetting SGD momentum after sync.
MPI Rank 0: 
MPI Rank 0: 12/21/2016 05:28:09: Starting minibatch loop, distributed reading is ENABLED.
MPI Rank 0: 12/21/2016 05:28:09:   BaseAdaptiveLearnRateSearch: Epoch[ 4 of 5]-Minibatch[   1-   1, 5.00%]: CE = 1.83662591 * 512; Err = 0.50390625 * 512; time = 0.0679s; samplesPerSecond = 7538.4
MPI Rank 0: 12/21/2016 05:28:09:   BaseAdaptiveLearnRateSearch: Epoch[ 4 of 5]-Minibatch[   2-   2, 10.00%]: CE = 1.95013576 * 512; Err = 0.53125000 * 512; time = 0.0840s; samplesPerSecond = 6093.1
MPI Rank 0: 12/21/2016 05:28:09:   BaseAdaptiveLearnRateSearch: Epoch[ 4 of 5]-Minibatch[   3-   3, 15.00%]: CE = 0.00000000 * 0; Err = 0.00000000 * 0; time = 8.8000e-005s; samplesPerSecond = 0.0
MPI Rank 0: 		(model aggregation stats): 1-th sync point was hit, introducing a 0.00-seconds latency this time; accumulated time on sync point = 0.00 seconds , average latency = 0.00 seconds
MPI Rank 0: 		(model aggregation stats) 1-th sync:     0.25 seconds since last report (0.06 seconds on comm.); 1536 samples processed by 2 workers (1024 by me);
MPI Rank 0: 		(model aggregation stats) 1-th sync: totalThroughput = 6.16k samplesPerSecond , throughputPerWorker = 3.08k samplesPerSecond
MPI Rank 0: 12/21/2016 05:28:09:  Finished Mini-Epoch[4]: CE = 1.91381689 * 1536; Err = 0.52148438 * 1536; learningRatePerSample = 0; minibatchSize = 1024
MPI Rank 0: ValidateSubNetwork: featNorm.invStdDev InvStdDev operation changed, from [363 x 1] to [363].ValidateSubNetwork: featNorm.mean Mean operation changed, from [363 x 1] to [363].ValidateSubNetwork: logPrior._ Mean operation changed, from [132 x 1] to [132].Parallel training (2 workers) using BlockMomentumSGD with block momentum = 0.5000, block momentum time constant (per worker) = 2954.6394, block learning rate = 1.0000, block size per worker = 2048 samples, using Nesterov-style block momentum, resetting SGD momentum after sync.
MPI Rank 0: 
MPI Rank 0: 12/21/2016 05:28:09: Starting minibatch loop, distributed reading is ENABLED.
MPI Rank 0: 12/21/2016 05:28:10:   AdaptiveLearnRateSearch: Epoch[ 4 of 5]-Minibatch[   1-   1, 5.00%]: CE = 1.83662591 * 512; Err = 0.50390625 * 512; time = 0.1124s; samplesPerSecond = 4555.2
MPI Rank 0: 12/21/2016 05:28:10:   AdaptiveLearnRateSearch: Epoch[ 4 of 5]-Minibatch[   2-   2, 10.00%]: CE = 1.94792524 * 512; Err = 0.52929688 * 512; time = 0.1412s; samplesPerSecond = 3627.2
MPI Rank 0: 12/21/2016 05:28:10:   AdaptiveLearnRateSearch: Epoch[ 4 of 5]-Minibatch[   3-   3, 15.00%]: CE = 0.00000000 * 0; Err = 0.00000000 * 0; time = 9.4000e-005s; samplesPerSecond = 0.0
MPI Rank 0: 		(model aggregation stats): 1-th sync point was hit, introducing a 0.00-seconds latency this time; accumulated time on sync point = 0.00 seconds , average latency = 0.00 seconds
MPI Rank 0: 		(model aggregation stats) 1-th sync:     0.36 seconds since last report (0.08 seconds on comm.); 1536 samples processed by 2 workers (1024 by me);
MPI Rank 0: 		(model aggregation stats) 1-th sync: totalThroughput = 4.24k samplesPerSecond , throughputPerWorker = 2.12k samplesPerSecond
MPI Rank 0: 12/21/2016 05:28:10:  Finished Mini-Epoch[4]: CE = 1.91308005 * 1536; Err = 0.52083333 * 1536; learningRatePerSample = 0.0039062502; minibatchSize = 1024
MPI Rank 0: ValidateSubNetwork: featNorm.invStdDev InvStdDev operation changed, from [363 x 1] to [363].ValidateSubNetwork: featNorm.mean Mean operation changed, from [363 x 1] to [363].ValidateSubNetwork: logPrior._ Mean operation changed, from [132 x 1] to [132].12/21/2016 05:28:10:  SearchForBestLearnRate Epoch[4]: Best learningRatePerSample = 0.003906250186, baseCriterion=1.922629502
MPI Rank 0: 
MPI Rank 0: 12/21/2016 05:28:10: Starting Epoch 4: learning rate per sample = 0.003906  effective momentum = 0.656119  momentum as time constant = 2429.9 samples
MPI Rank 0: Parallel training (2 workers) using BlockMomentumSGD with block momentum = 0.5000, block momentum time constant (per worker) = 2954.6394, block learning rate = 1.0000, block size per worker = 2048 samples, using Nesterov-style block momentum, resetting SGD momentum after sync.
MPI Rank 0: 
MPI Rank 0: 12/21/2016 05:28:10: Starting minibatch loop, distributed reading is ENABLED.
MPI Rank 0: 12/21/2016 05:28:10:  Epoch[ 4 of 5]-Minibatch[   1-   1, 5.00%]: CE = 1.83662591 * 512; Err = 0.50390625 * 512; time = 0.1559s; samplesPerSecond = 3283.1
MPI Rank 0: 12/21/2016 05:28:11:  Epoch[ 4 of 5]-Minibatch[   2-   2, 10.00%]: CE = 1.94792524 * 512; Err = 0.52929688 * 512; time = 0.2720s; samplesPerSecond = 1882.2
MPI Rank 0: 12/21/2016 05:28:11:  Epoch[ 4 of 5]-Minibatch[   3-   3, 15.00%]: CE = 1.84996779 * 512; Err = 0.50195313 * 512; time = 0.1560s; samplesPerSecond = 3282.6
MPI Rank 0: 		(model aggregation stats): 1-th sync point was hit, introducing a 0.00-seconds latency this time; accumulated time on sync point = 0.00 seconds , average latency = 0.00 seconds
MPI Rank 0: 		(model aggregation stats) 1-th sync:     0.83 seconds since last report (0.07 seconds on comm.); 4096 samples processed by 2 workers (2048 by me);
MPI Rank 0: 		(model aggregation stats) 1-th sync: totalThroughput = 4.96k samplesPerSecond , throughputPerWorker = 2.48k samplesPerSecond
MPI Rank 0: 12/21/2016 05:28:11:  Epoch[ 4 of 5]-Minibatch[   4-   4, 20.00%]: CE = 1.91533561 * 512; Err = 0.52929688 * 512; time = 0.2340s; samplesPerSecond = 2187.7
MPI Rank 0: 12/21/2016 05:28:11:  Epoch[ 4 of 5]-Minibatch[   5-   5, 25.00%]: CE = 1.91106814 * 512; Err = 0.55273438 * 512; time = 0.1504s; samplesPerSecond = 3403.2
MPI Rank 0: 12/21/2016 05:28:11:  Epoch[ 4 of 5]-Minibatch[   6-   6, 30.00%]: CE = 1.89016030 * 512; Err = 0.51171875 * 512; time = 0.1249s; samplesPerSecond = 4099.9
MPI Rank 0: 12/21/2016 05:28:11:  Epoch[ 4 of 5]-Minibatch[   7-   7, 35.00%]: CE = 1.85983068 * 512; Err = 0.51953125 * 512; time = 0.1390s; samplesPerSecond = 3684.0
MPI Rank 0: 		(model aggregation stats): 2-th sync point was hit, introducing a 0.25-seconds latency this time; accumulated time on sync point = 0.25 seconds , average latency = 0.13 seconds
MPI Rank 0: 		(model aggregation stats) 2-th sync:     0.91 seconds since last report (0.07 seconds on comm.); 4096 samples processed by 2 workers (2048 by me);
MPI Rank 0: 		(model aggregation stats) 2-th sync: totalThroughput = 4.48k samplesPerSecond , throughputPerWorker = 2.24k samplesPerSecond
MPI Rank 0: 12/21/2016 05:28:12:  Epoch[ 4 of 5]-Minibatch[   8-   8, 40.00%]: CE = 1.82250013 * 512; Err = 0.51171875 * 512; time = 0.4982s; samplesPerSecond = 1027.7
MPI Rank 0: 12/21/2016 05:28:12:  Epoch[ 4 of 5]-Minibatch[   9-   9, 45.00%]: CE = 1.92651054 * 512; Err = 0.52929688 * 512; time = 0.1589s; samplesPerSecond = 3222.3
MPI Rank 0: 12/21/2016 05:28:12:  Epoch[ 4 of 5]-Minibatch[  10-  10, 50.00%]: CE = 1.89119121 * 512; Err = 0.55664063 * 512; time = 0.2776s; samplesPerSecond = 1844.1
MPI Rank 0: 12/21/2016 05:28:13:  Epoch[ 4 of 5]-Minibatch[  11-  11, 55.00%]: CE = 1.85766465 * 512; Err = 0.50585938 * 512; time = 0.1462s; samplesPerSecond = 3502.6
MPI Rank 0: 		(model aggregation stats): 3-th sync point was hit, introducing a 0.00-seconds latency this time; accumulated time on sync point = 0.25 seconds , average latency = 0.08 seconds
MPI Rank 0: 		(model aggregation stats) 3-th sync:     0.82 seconds since last report (0.06 seconds on comm.); 4096 samples processed by 2 workers (2048 by me);
MPI Rank 0: 		(model aggregation stats) 3-th sync: totalThroughput = 4.97k samplesPerSecond , throughputPerWorker = 2.49k samplesPerSecond
MPI Rank 0: 12/21/2016 05:28:13:  Epoch[ 4 of 5]-Minibatch[  12-  12, 60.00%]: CE = 1.91245983 * 512; Err = 0.53710938 * 512; time = 0.2391s; samplesPerSecond = 2141.1
MPI Rank 0: 12/21/2016 05:28:13:  Epoch[ 4 of 5]-Minibatch[  13-  13, 65.00%]: CE = 1.83122854 * 512; Err = 0.50976563 * 512; time = 0.2934s; samplesPerSecond = 1744.9
MPI Rank 0: 12/21/2016 05:28:13:  Epoch[ 4 of 5]-Minibatch[  14-  14, 70.00%]: CE = 1.82747049 * 512; Err = 0.52539063 * 512; time = 0.1579s; samplesPerSecond = 3243.4
MPI Rank 0: 12/21/2016 05:28:13:  Epoch[ 4 of 5]-Minibatch[  15-  15, 75.00%]: CE = 1.84116676 * 512; Err = 0.50390625 * 512; time = 0.1681s; samplesPerSecond = 3044.9
MPI Rank 0: 12/21/2016 05:28:14:  Epoch[ 4 of 5]-Minibatch[  16-  16, 80.00%]: CE = 1.86439812 * 512; Err = 0.52734375 * 512; time = 0.2040s; samplesPerSecond = 2510.1
MPI Rank 0: 12/21/2016 05:28:14:  Epoch[ 4 of 5]-Minibatch[  17-  17, 85.00%]: CE = 1.87971759 * 512; Err = 0.51757813 * 512; time = 0.0766s; samplesPerSecond = 6686.9
MPI Rank 0: 12/21/2016 05:28:14:  Epoch[ 4 of 5]-Minibatch[  18-  18, 90.00%]: CE = 1.77581930 * 512; Err = 0.53515625 * 512; time = 0.0794s; samplesPerSecond = 6446.4
MPI Rank 0: 12/21/2016 05:28:14:  Epoch[ 4 of 5]-Minibatch[  19-  19, 95.00%]: CE = 1.78889741 * 512; Err = 0.51757813 * 512; time = 0.0746s; samplesPerSecond = 6865.1
MPI Rank 0: 12/21/2016 05:28:14:  Epoch[ 4 of 5]-Minibatch[  20-  20, 100.00%]: CE = 1.83607952 * 512; Err = 0.54101563 * 512; time = 0.0723s; samplesPerSecond = 7080.0
MPI Rank 0: 12/21/2016 05:28:14:  Epoch[ 4 of 5]-Minibatch[  21-  21, 105.00%]: CE = 1.96701587 * 512; Err = 0.53515625 * 512; time = 0.0771s; samplesPerSecond = 6644.6
MPI Rank 0: 12/21/2016 05:28:14:  Epoch[ 4 of 5]-Minibatch[  22-  22, 110.00%]: CE = 1.74090893 * 512; Err = 0.49023438 * 512; time = 0.1585s; samplesPerSecond = 3230.4
MPI Rank 0: 12/21/2016 05:28:14:  Epoch[ 4 of 5]-Minibatch[  23-  23, 115.00%]: CE = 1.79959982 * 512; Err = 0.50390625 * 512; time = 0.0815s; samplesPerSecond = 6283.2
MPI Rank 0: 12/21/2016 05:28:14:  Epoch[ 4 of 5]-Minibatch[  24-  24, 120.00%]: CE = 1.83947195 * 512; Err = 0.49609375 * 512; time = 0.0792s; samplesPerSecond = 6464.3
MPI Rank 0: 12/21/2016 05:28:14:  Epoch[ 4 of 5]-Minibatch[  25-  25, 125.00%]: CE = 1.93199441 * 306; Err = 0.54575163 * 306; time = 0.0463s; samplesPerSecond = 6604.1
MPI Rank 0: 12/21/2016 05:28:14:  Epoch[ 4 of 5]-Minibatch[  26-  26, 130.00%]: CE = 0.00000000 * 0; Err = 0.00000000 * 0; time = 3.8000e-005s; samplesPerSecond = 0.0
MPI Rank 0: 		(model aggregation stats): 4-th sync point was hit, introducing a 0.00-seconds latency this time; accumulated time on sync point = 0.25 seconds , average latency = 0.06 seconds
MPI Rank 0: 		(model aggregation stats) 4-th sync:     1.59 seconds since last report (0.00 seconds on comm.); 8192 samples processed by 2 workers (6450 by me);
MPI Rank 0: 		(model aggregation stats) 4-th sync: totalThroughput = 5.14k samplesPerSecond , throughputPerWorker = 2.57k samplesPerSecond
MPI Rank 0: 12/21/2016 05:28:14: Finished Epoch[ 4 of 5]: [Training] CE = 1.86802921 * 20480; Err = 0.52246094 * 20480; totalSamplesSeen = 81920; learningRatePerSample = 0.0039062502; epochTime=5.68031s
MPI Rank 0: 12/21/2016 05:28:14: SGD: Saving checkpoint model 'C:\Users\svcphil\AppData\Local\Temp\cntk-test-20161221052142.365503\Speech\HTKDeserializers\DNN_ParallelBMWithAdjustLR@release_cpu/models/cntkSpeech.dnn.4'
MPI Rank 0: ValidateSubNetwork: featNorm.invStdDev InvStdDev operation changed, from [363 x 1] to [363].ValidateSubNetwork: featNorm.mean Mean operation changed, from [363 x 1] to [363].ValidateSubNetwork: logPrior._ Mean operation changed, from [132 x 1] to [132].Parallel training (2 workers) using BlockMomentumSGD with block momentum = 0.5000, block momentum time constant (per worker) = 2954.6394, block learning rate = 1.0000, block size per worker = 2048 samples, using Nesterov-style block momentum, resetting SGD momentum after sync.
MPI Rank 0: 
MPI Rank 0: 12/21/2016 05:28:15: Starting minibatch loop, distributed reading is ENABLED.
MPI Rank 0: 12/21/2016 05:28:15:   BaseAdaptiveLearnRateSearch: Epoch[ 5 of 5]-Minibatch[   1-   1, 5.00%]: CE = 1.95736438 * 512; Err = 0.52343750 * 512; time = 0.0733s; samplesPerSecond = 6986.7
MPI Rank 0: 12/21/2016 05:28:15:   BaseAdaptiveLearnRateSearch: Epoch[ 5 of 5]-Minibatch[   2-   2, 10.00%]: CE = 1.86304613 * 512; Err = 0.55468750 * 512; time = 0.0560s; samplesPerSecond = 9146.5
MPI Rank 0: 12/21/2016 05:28:15:   BaseAdaptiveLearnRateSearch: Epoch[ 5 of 5]-Minibatch[   3-   3, 15.00%]: CE = 0.00000000 * 0; Err = 0.00000000 * 0; time = 9.1000e-005s; samplesPerSecond = 0.0
MPI Rank 0: 		(model aggregation stats): 1-th sync point was hit, introducing a 0.00-seconds latency this time; accumulated time on sync point = 0.00 seconds , average latency = 0.00 seconds
MPI Rank 0: 		(model aggregation stats) 1-th sync:     0.24 seconds since last report (0.07 seconds on comm.); 1536 samples processed by 2 workers (1024 by me);
MPI Rank 0: 		(model aggregation stats) 1-th sync: totalThroughput = 6.31k samplesPerSecond , throughputPerWorker = 3.16k samplesPerSecond
MPI Rank 0: 12/21/2016 05:28:15:  Finished Mini-Epoch[5]: CE = 1.84766700 * 1536; Err = 0.51627604 * 1536; learningRatePerSample = 0; minibatchSize = 1024
MPI Rank 0: ValidateSubNetwork: featNorm.invStdDev InvStdDev operation changed, from [363 x 1] to [363].ValidateSubNetwork: featNorm.mean Mean operation changed, from [363 x 1] to [363].ValidateSubNetwork: logPrior._ Mean operation changed, from [132 x 1] to [132].Parallel training (2 workers) using BlockMomentumSGD with block momentum = 0.5000, block momentum time constant (per worker) = 2954.6394, block learning rate = 1.0000, block size per worker = 2048 samples, using Nesterov-style block momentum, resetting SGD momentum after sync.
MPI Rank 0: 
MPI Rank 0: 12/21/2016 05:28:15: Starting minibatch loop, distributed reading is ENABLED.
MPI Rank 0: 12/21/2016 05:28:16:   AdaptiveLearnRateSearch: Epoch[ 5 of 5]-Minibatch[   1-   1, 5.00%]: CE = 1.95736438 * 512; Err = 0.52343750 * 512; time = 0.1561s; samplesPerSecond = 3280.3
MPI Rank 0: 12/21/2016 05:28:16:   AdaptiveLearnRateSearch: Epoch[ 5 of 5]-Minibatch[   2-   2, 10.00%]: CE = 1.80265075 * 512; Err = 0.49609375 * 512; time = 0.0832s; samplesPerSecond = 6154.4
MPI Rank 0: 12/21/2016 05:28:16:   AdaptiveLearnRateSearch: Epoch[ 5 of 5]-Minibatch[   3-   3, 15.00%]: CE = 0.00000000 * 0; Err = 0.00000000 * 0; time = 8.8000e-005s; samplesPerSecond = 0.0
MPI Rank 0: 		(model aggregation stats): 1-th sync point was hit, introducing a 0.00-seconds latency this time; accumulated time on sync point = 0.00 seconds , average latency = 0.00 seconds
MPI Rank 0: 		(model aggregation stats) 1-th sync:     0.37 seconds since last report (0.10 seconds on comm.); 1536 samples processed by 2 workers (1024 by me);
MPI Rank 0: 		(model aggregation stats) 1-th sync: totalThroughput = 4.18k samplesPerSecond , throughputPerWorker = 2.09k samplesPerSecond
MPI Rank 0: 12/21/2016 05:28:16:  Finished Mini-Epoch[5]: CE = 1.82753521 * 1536; Err = 0.49674479 * 1536; learningRatePerSample = 0.0063207938; minibatchSize = 1024
MPI Rank 0: ValidateSubNetwork: featNorm.invStdDev InvStdDev operation changed, from [363 x 1] to [363].ValidateSubNetwork: featNorm.mean Mean operation changed, from [363 x 1] to [363].ValidateSubNetwork: logPrior._ Mean operation changed, from [132 x 1] to [132].12/21/2016 05:28:16:  SearchForBestLearnRate Epoch[5]: Best learningRatePerSample = 0.006320793777, baseCriterion=1.85275755
MPI Rank 0: 
MPI Rank 0: 12/21/2016 05:28:16: Starting Epoch 5: learning rate per sample = 0.006321  effective momentum = 0.656119  momentum as time constant = 2429.9 samples
MPI Rank 0: Parallel training (2 workers) using BlockMomentumSGD with block momentum = 0.5000, block momentum time constant (per worker) = 2954.6394, block learning rate = 1.0000, block size per worker = 2048 samples, using Nesterov-style block momentum, resetting SGD momentum after sync.
MPI Rank 0: 
MPI Rank 0: 12/21/2016 05:28:16: Starting minibatch loop, distributed reading is ENABLED.
MPI Rank 0: 12/21/2016 05:28:16:  Epoch[ 5 of 5]-Minibatch[   1-   1, 5.00%]: CE = 1.95736438 * 512; Err = 0.52343750 * 512; time = 0.1400s; samplesPerSecond = 3658.0
MPI Rank 0: 12/21/2016 05:28:16:  Epoch[ 5 of 5]-Minibatch[   2-   2, 10.00%]: CE = 1.80265075 * 512; Err = 0.49609375 * 512; time = 0.1346s; samplesPerSecond = 3804.8
MPI Rank 0: 12/21/2016 05:28:17:  Epoch[ 5 of 5]-Minibatch[   3-   3, 15.00%]: CE = 1.75549697 * 512; Err = 0.51171875 * 512; time = 0.1655s; samplesPerSecond = 3093.4
MPI Rank 0: 		(model aggregation stats): 1-th sync point was hit, introducing a 0.15-seconds latency this time; accumulated time on sync point = 0.15 seconds , average latency = 0.15 seconds
MPI Rank 0: 		(model aggregation stats) 1-th sync:     0.93 seconds since last report (0.10 seconds on comm.); 4096 samples processed by 2 workers (2048 by me);
MPI Rank 0: 		(model aggregation stats) 1-th sync: totalThroughput = 4.40k samplesPerSecond , throughputPerWorker = 2.20k samplesPerSecond
MPI Rank 0: 12/21/2016 05:28:17:  Epoch[ 5 of 5]-Minibatch[   4-   4, 20.00%]: CE = 1.90051737 * 512; Err = 0.54492188 * 512; time = 0.4800s; samplesPerSecond = 1066.7
MPI Rank 0: 12/21/2016 05:28:17:  Epoch[ 5 of 5]-Minibatch[   5-   5, 25.00%]: CE = 1.86454994 * 512; Err = 0.53515625 * 512; time = 0.1962s; samplesPerSecond = 2609.5
MPI Rank 0: 12/21/2016 05:28:17:  Epoch[ 5 of 5]-Minibatch[   6-   6, 30.00%]: CE = 1.81624071 * 512; Err = 0.52343750 * 512; time = 0.1689s; samplesPerSecond = 3030.6
MPI Rank 0: 12/21/2016 05:28:18:  Epoch[ 5 of 5]-Minibatch[   7-   7, 35.00%]: CE = 1.76991108 * 512; Err = 0.52539063 * 512; time = 0.1359s; samplesPerSecond = 3767.2
MPI Rank 0: 		(model aggregation stats): 2-th sync point was hit, introducing a 0.00-seconds latency this time; accumulated time on sync point = 0.15 seconds , average latency = 0.07 seconds
MPI Rank 0: 		(model aggregation stats) 2-th sync:     0.86 seconds since last report (0.08 seconds on comm.); 4096 samples processed by 2 workers (2048 by me);
MPI Rank 0: 		(model aggregation stats) 2-th sync: totalThroughput = 4.75k samplesPerSecond , throughputPerWorker = 2.38k samplesPerSecond
MPI Rank 0: 12/21/2016 05:28:18:  Epoch[ 5 of 5]-Minibatch[   8-   8, 40.00%]: CE = 1.88699772 * 512; Err = 0.54687500 * 512; time = 0.3591s; samplesPerSecond = 1425.7
MPI Rank 0: 12/21/2016 05:28:18:  Epoch[ 5 of 5]-Minibatch[   9-   9, 45.00%]: CE = 1.74602775 * 512; Err = 0.51171875 * 512; time = 0.1409s; samplesPerSecond = 3634.2
MPI Rank 0: 12/21/2016 05:28:18:  Epoch[ 5 of 5]-Minibatch[  10-  10, 50.00%]: CE = 1.66755622 * 512; Err = 0.49609375 * 512; time = 0.1552s; samplesPerSecond = 3299.7
MPI Rank 0: 12/21/2016 05:28:18:  Epoch[ 5 of 5]-Minibatch[  11-  11, 55.00%]: CE = 1.74110886 * 512; Err = 0.50781250 * 512; time = 0.1982s; samplesPerSecond = 2583.4
MPI Rank 0: 		(model aggregation stats): 3-th sync point was hit, introducing a 0.04-seconds latency this time; accumulated time on sync point = 0.19 seconds , average latency = 0.06 seconds
MPI Rank 0: 		(model aggregation stats) 3-th sync:     0.71 seconds since last report (0.00 seconds on comm.); 4096 samples processed by 2 workers (2048 by me);
MPI Rank 0: 		(model aggregation stats) 3-th sync: totalThroughput = 5.79k samplesPerSecond , throughputPerWorker = 2.89k samplesPerSecond
MPI Rank 0: 12/21/2016 05:28:19:  Epoch[ 5 of 5]-Minibatch[  12-  12, 60.00%]: CE = 1.82518528 * 512; Err = 0.51953125 * 512; time = 0.2122s; samplesPerSecond = 2413.4
MPI Rank 0: 12/21/2016 05:28:19:  Epoch[ 5 of 5]-Minibatch[  13-  13, 65.00%]: CE = 1.72221058 * 512; Err = 0.48046875 * 512; time = 0.1355s; samplesPerSecond = 3777.4
MPI Rank 0: 12/21/2016 05:28:19:  Epoch[ 5 of 5]-Minibatch[  14-  14, 70.00%]: CE = 1.76486948 * 512; Err = 0.53320313 * 512; time = 0.2960s; samplesPerSecond = 1729.9
MPI Rank 0: 12/21/2016 05:28:19:  Epoch[ 5 of 5]-Minibatch[  15-  15, 75.00%]: CE = 1.72811109 * 512; Err = 0.50781250 * 512; time = 0.1174s; samplesPerSecond = 4362.6
MPI Rank 0: 12/21/2016 05:28:19:  Epoch[ 5 of 5]-Minibatch[  16-  16, 80.00%]: CE = 1.73720486 * 512; Err = 0.48437500 * 512; time = 0.0764s; samplesPerSecond = 6704.0
MPI Rank 0: 12/21/2016 05:28:19:  Epoch[ 5 of 5]-Minibatch[  17-  17, 85.00%]: CE = 1.64690249 * 512; Err = 0.47265625 * 512; time = 0.0736s; samplesPerSecond = 6960.7
MPI Rank 0: 12/21/2016 05:28:19:  Epoch[ 5 of 5]-Minibatch[  18-  18, 90.00%]: CE = 1.62069217 * 512; Err = 0.45507813 * 512; time = 0.0756s; samplesPerSecond = 6774.1
MPI Rank 0: 12/21/2016 05:28:20:  Epoch[ 5 of 5]-Minibatch[  19-  19, 95.00%]: CE = 1.68369876 * 512; Err = 0.46289063 * 512; time = 0.1490s; samplesPerSecond = 3435.1
MPI Rank 0: 12/21/2016 05:28:20:  Epoch[ 5 of 5]-Minibatch[  20-  20, 100.00%]: CE = 1.70660558 * 512; Err = 0.51171875 * 512; time = 0.0852s; samplesPerSecond = 6007.1
MPI Rank 0: 12/21/2016 05:28:20:  Epoch[ 5 of 5]-Minibatch[  21-  21, 105.00%]: CE = 1.62135182 * 512; Err = 0.50000000 * 512; time = 0.0768s; samplesPerSecond = 6663.7
MPI Rank 0: 12/21/2016 05:28:20:  Epoch[ 5 of 5]-Minibatch[  22-  22, 110.00%]: CE = 1.75380156 * 512; Err = 0.50976563 * 512; time = 0.0740s; samplesPerSecond = 6914.7
MPI Rank 0: 12/21/2016 05:28:20:  Epoch[ 5 of 5]-Minibatch[  23-  23, 115.00%]: CE = 1.79493194 * 512; Err = 0.50976563 * 512; time = 0.0724s; samplesPerSecond = 7075.1
MPI Rank 0: 12/21/2016 05:28:20:  Epoch[ 5 of 5]-Minibatch[  24-  24, 120.00%]: CE = 1.58781091 * 512; Err = 0.47656250 * 512; time = 0.0716s; samplesPerSecond = 7155.5
MPI Rank 0: 12/21/2016 05:28:20:  Epoch[ 5 of 5]-Minibatch[  25-  25, 125.00%]: CE = 1.58255106 * 179; Err = 0.46927374 * 179; time = 0.0290s; samplesPerSecond = 6170.5
MPI Rank 0: 12/21/2016 05:28:20:  Epoch[ 5 of 5]-Minibatch[  26-  26, 130.00%]: CE = 0.00000000 * 0; Err = 0.00000000 * 0; time = 3.9000e-005s; samplesPerSecond = 0.0
MPI Rank 0: 		(model aggregation stats): 4-th sync point was hit, introducing a 0.00-seconds latency this time; accumulated time on sync point = 0.19 seconds , average latency = 0.05 seconds
MPI Rank 0: 		(model aggregation stats) 4-th sync:     1.36 seconds since last report (0.01 seconds on comm.); 8192 samples processed by 2 workers (6323 by me);
MPI Rank 0: 		(model aggregation stats) 4-th sync: totalThroughput = 6.01k samplesPerSecond , throughputPerWorker = 3.00k samplesPerSecond
MPI Rank 0: 12/21/2016 05:28:20: Finished Epoch[ 5 of 5]: [Training] CE = 1.78255566 * 20480; Err = 0.51186523 * 20480; totalSamplesSeen = 102400; learningRatePerSample = 0.0063207938; epochTime=5.48442s
MPI Rank 0: 12/21/2016 05:28:20: SGD: Saving checkpoint model 'C:\Users\svcphil\AppData\Local\Temp\cntk-test-20161221052142.365503\Speech\HTKDeserializers\DNN_ParallelBMWithAdjustLR@release_cpu/models/cntkSpeech.dnn'
MPI Rank 0: 
MPI Rank 0: 12/21/2016 05:28:20: Action "train" complete.
MPI Rank 0: 
MPI Rank 0: 12/21/2016 05:28:20: __COMPLETED__
MPI Rank 0: ~MPIWrapper
MPI Rank 1: 12/21/2016 05:27:45: Redirecting stderr to file C:\Users\svcphil\AppData\Local\Temp\cntk-test-20161221052142.365503\Speech\HTKDeserializers\DNN_ParallelBMWithAdjustLR@release_cpu/stderr_SpeechTrain.logrank1
MPI Rank 1: 12/21/2016 05:27:45: -------------------------------------------------------------------
MPI Rank 1: 12/21/2016 05:27:45: Build info: 
MPI Rank 1: 
MPI Rank 1: 12/21/2016 05:27:45: 		Built time: Dec 21 2016 04:21:26
MPI Rank 1: 12/21/2016 05:27:45: 		Last modified date: Tue Dec 20 18:55:12 2016
MPI Rank 1: 12/21/2016 05:27:45: 		Build type: Release
MPI Rank 1: 12/21/2016 05:27:45: 		Build target: GPU
MPI Rank 1: 12/21/2016 05:27:45: 		With ASGD: yes
MPI Rank 1: 12/21/2016 05:27:45: 		Math lib: mkl
MPI Rank 1: 12/21/2016 05:27:45: 		CUDA_PATH: C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v8.0
MPI Rank 1: 12/21/2016 05:27:45: 		CUB_PATH: c:\src\cub-1.4.1
MPI Rank 1: 12/21/2016 05:27:45: 		CUDNN_PATH: C:\local\cudnn-8.0-windows10-x64-v5.1
MPI Rank 1: 12/21/2016 05:27:45: 		Build Branch: HEAD
MPI Rank 1: 12/21/2016 05:27:45: 		Build SHA1: 0a2e20ddce32ca3cd458ef0358757e1489d9afe3 (modified)
MPI Rank 1: 12/21/2016 05:27:45: 		Built by svcphil on LIANA-09-w
MPI Rank 1: 12/21/2016 05:27:45: 		Build Path: C:\jenkins\workspace\CNTK-Build-Windows\Source\CNTK\
MPI Rank 1: 12/21/2016 05:27:45: -------------------------------------------------------------------
MPI Rank 1: 12/21/2016 05:27:45: -------------------------------------------------------------------
MPI Rank 1: 12/21/2016 05:27:45: GPU info:
MPI Rank 1: 
MPI Rank 1: 12/21/2016 05:27:45: 		Device[0]: cores = 2496; computeCapability = 5.2; type = "Quadro M4000"; memory = 8192 MB
MPI Rank 1: 12/21/2016 05:27:45: -------------------------------------------------------------------
MPI Rank 1: 
MPI Rank 1: Configuration After Processing and Variable Resolution:
MPI Rank 1: 
MPI Rank 1: configparameters: cntk.cntk:command=SpeechTrain
MPI Rank 1: configparameters: cntk.cntk:ConfigDir=C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\HTKDeserializers\DNN\ParallelBMWithAdjustLR
MPI Rank 1: configparameters: cntk.cntk:currentDirectory=C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\Data
MPI Rank 1: configparameters: cntk.cntk:DataDir=C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\Data
MPI Rank 1: configparameters: cntk.cntk:deviceId=-1
MPI Rank 1: configparameters: cntk.cntk:framemode=true
MPI Rank 1: configparameters: cntk.cntk:modelPath=C:\Users\svcphil\AppData\Local\Temp\cntk-test-20161221052142.365503\Speech\HTKDeserializers\DNN_ParallelBMWithAdjustLR@release_cpu/models/cntkSpeech.dnn
MPI Rank 1: configparameters: cntk.cntk:numCPUThreads=4
MPI Rank 1: configparameters: cntk.cntk:OutputDir=C:\Users\svcphil\AppData\Local\Temp\cntk-test-20161221052142.365503\Speech\HTKDeserializers\DNN_ParallelBMWithAdjustLR@release_cpu
MPI Rank 1: configparameters: cntk.cntk:parallelTrain=true
MPI Rank 1: configparameters: cntk.cntk:precision=double
MPI Rank 1: configparameters: cntk.cntk:RunDir=C:\Users\svcphil\AppData\Local\Temp\cntk-test-20161221052142.365503\Speech\HTKDeserializers\DNN_ParallelBMWithAdjustLR@release_cpu
MPI Rank 1: configparameters: cntk.cntk:SpeechTrain={
MPI Rank 1:     action = "train"
MPI Rank 1:     BrainScriptNetworkBuilder = {
MPI Rank 1:         layerSizes = 363:512:512:132
MPI Rank 1:         trainingCriterion = 'CE'
MPI Rank 1:         evalCriterion = 'Err'
MPI Rank 1:         applyMeanVarNorm = true
MPI Rank 1:         L = Length(layerSizes)-1    // number of model layers
MPI Rank 1:         features = Input { layerSizes[0] } // 1, tag='feature' 
MPI Rank 1:         labels = Input { layerSizes[L] } // 1, tag='label' 
MPI Rank 1:         featNorm = if applyMeanVarNorm
MPI Rank 1:                    then MeanVarNorm(features)
MPI Rank 1:                    else features
MPI Rank 1:         layers[layer:1..L-1] = if layer > 1
MPI Rank 1:                                then SBFF(layers[layer-1].Eh, layerSizes[layer], layerSizes[layer-1])
MPI Rank 1:                                else SBFF(featNorm, layerSizes[layer], layerSizes[layer-1])
MPI Rank 1:         outLayer = BFF(layers[L-1].Eh, layerSizes[L], layerSizes[L-1])
MPI Rank 1:         outZ = outLayer.z        // + PastValue(layerSizes[L], 1, outLayer.z)
MPI Rank 1:         CE = if trainingCriterion == 'CE'
MPI Rank 1:              then CrossEntropyWithSoftmax(labels, outZ, tag='criterion')
MPI Rank 1:              else Fail('unknown trainingCriterion ' + trainingCriterion)
MPI Rank 1:         Err = if evalCriterion == 'Err' then
MPI Rank 1:               ClassificationError(labels, outZ, tag='evaluation')
MPI Rank 1:               else Fail('unknown evalCriterion ' + evalCriterion)
MPI Rank 1:         logPrior = LogPrior(labels)
MPI Rank 1:         // TODO: how to add a tag to an infix operation?
MPI Rank 1:         ScaledLogLikelihood = Minus (outZ, logPrior, tag='output')
MPI Rank 1:     }
MPI Rank 1:     SGD = {
MPI Rank 1:         epochSize = 20480
MPI Rank 1:         minibatchSize = 64:256:1024
MPI Rank 1:         learningRatesPerMB = 1.0:0.5:0.1
MPI Rank 1:         numMBsToShowResult = 1
MPI Rank 1:         momentumPerMB = 0.9:0.656119
MPI Rank 1:         dropoutRate = 0.0
MPI Rank 1:         maxEpochs = 5
MPI Rank 1:         keepCheckPointFiles = true
MPI Rank 1:         clippingThresholdPerSample = 1#INF
MPI Rank 1:         ParallelTrain = {
MPI Rank 1:             parallelizationMethod = "BlockMomentumSGD"
MPI Rank 1:             distributedMBReading = true
MPI Rank 1:             syncPerfStats=1
MPI Rank 1:             BlockMomentumSGD = {
MPI Rank 1:                 blockSizePerWorker=2048
MPI Rank 1:                 resetSGDMomentum=true
MPI Rank 1:                 useNesterovMomentum=true
MPI Rank 1:             }
MPI Rank 1:         }
MPI Rank 1:         AutoAdjust = {
MPI Rank 1:             reduceLearnRateIfImproveLessThan = 0
MPI Rank 1:             loadBestModel = true
MPI Rank 1:             increaseLearnRateIfImproveMoreThan = 1000000000
MPI Rank 1:             learnRateDecreaseFactor = 0.5
MPI Rank 1:             learnRateIncreaseFactor = 1.382
MPI Rank 1:             autoAdjustLR = "searchBeforeEpoch"
MPI Rank 1:             numMiniBatch4LRSearch = 20
MPI Rank 1:             numPrevLearnRates = 3            
MPI Rank 1:         }
MPI Rank 1:     }
MPI Rank 1:     reader = {
MPI Rank 1:         verbosity = 0 ; randomize = true
MPI Rank 1:         deserializers = ({
MPI Rank 1:             type = "HTKFeatureDeserializer"
MPI Rank 1:             module = "HTKDeserializers"                
MPI Rank 1:             input = { features = { dim = 363 ; scpFile = "glob_0000.scp" } }
MPI Rank 1:         }:{
MPI Rank 1:             type = "HTKMLFDeserializer"
MPI Rank 1:             module = "HTKDeserializers"
MPI Rank 1:             input = { labels = { mlfFile = "C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\Data/glob_0000.mlf" ; labelMappingFile = "C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\Data/state.list" ; labelDim = 132 } }
MPI Rank 1:         })
MPI Rank 1:     }
MPI Rank 1: } [SGD=[ParallelTrain=[parallelizationStartEpoch=2]]]
MPI Rank 1: 
MPI Rank 1: configparameters: cntk.cntk:stderr=C:\Users\svcphil\AppData\Local\Temp\cntk-test-20161221052142.365503\Speech\HTKDeserializers\DNN_ParallelBMWithAdjustLR@release_cpu/stderr
MPI Rank 1: configparameters: cntk.cntk:timestamping=true
MPI Rank 1: configparameters: cntk.cntk:traceLevel=1
MPI Rank 1: 12/21/2016 05:27:45: Commands: SpeechTrain
MPI Rank 1: 12/21/2016 05:27:45: precision = "double"
MPI Rank 1: 12/21/2016 05:27:45: Using 4 CPU threads.
MPI Rank 1: 
MPI Rank 1: 12/21/2016 05:27:45: ##############################################################################
MPI Rank 1: 12/21/2016 05:27:45: #                                                                            #
MPI Rank 1: 12/21/2016 05:27:45: # SpeechTrain command (train action)                                         #
MPI Rank 1: 12/21/2016 05:27:45: #                                                                            #
MPI Rank 1: 12/21/2016 05:27:45: ##############################################################################
MPI Rank 1: 
MPI Rank 1: 12/21/2016 05:27:45: WARNING: 'numMiniBatch4LRSearch' is deprecated, please remove it and use 'numSamples4Search' instead.
MPI Rank 1: 12/21/2016 05:27:45: 
MPI Rank 1: Creating virgin network.
MPI Rank 1: 
MPI Rank 1: Post-processing network...
MPI Rank 1: 
MPI Rank 1: 6 roots:
MPI Rank 1: 	CE = CrossEntropyWithSoftmax()
MPI Rank 1: 	Err = ClassificationError()
MPI Rank 1: 	ScaledLogLikelihood = Minus()
MPI Rank 1: 	featNorm.invStdDev = InvStdDev()
MPI Rank 1: 	featNorm.mean = Mean()
MPI Rank 1: 	logPrior._ = Mean()
MPI Rank 1: 
MPI Rank 1: Validating network. 25 nodes to process in pass 1.
MPI Rank 1: 
MPI Rank 1: Validating --> labels = InputValue() :  -> [132 x *]
MPI Rank 1: Validating --> outLayer.W = LearnableParameter() :  -> [132 x 512]
MPI Rank 1: Validating --> layers[2].Eh._.W = LearnableParameter() :  -> [512 x 512]
MPI Rank 1: Validating --> layers[1].Eh._.W = LearnableParameter() :  -> [512 x 363]
MPI Rank 1: Validating --> features = InputValue() :  -> [363 x *]
MPI Rank 1: Validating --> featNorm.mean = Mean (features) : [363 x *] -> [363]
MPI Rank 1: Validating --> featNorm.ElementTimesArgs[0] = Minus (features, featNorm.mean) : [363 x *], [363] -> [363 x *]
MPI Rank 1: Validating --> featNorm.invStdDev = InvStdDev (features) : [363 x *] -> [363]
MPI Rank 1: Validating --> featNorm = ElementTimes (featNorm.ElementTimesArgs[0], featNorm.invStdDev) : [363 x *], [363] -> [363 x *]
MPI Rank 1: Validating --> layers[1].Eh._.z.PlusArgs[0] = Times (layers[1].Eh._.W, featNorm) : [512 x 363], [363 x *] -> [512 x *]
MPI Rank 1: Validating --> layers[1].Eh._.B = LearnableParameter() :  -> [512 x 1]
MPI Rank 1: Validating --> layers[1].Eh._.z = Plus (layers[1].Eh._.z.PlusArgs[0], layers[1].Eh._.B) : [512 x *], [512 x 1] -> [512 x 1 x *]
MPI Rank 1: Validating --> layers[1].Eh = Sigmoid (layers[1].Eh._.z) : [512 x 1 x *] -> [512 x 1 x *]
MPI Rank 1: Validating --> layers[2].Eh._.z.PlusArgs[0] = Times (layers[2].Eh._.W, layers[1].Eh) : [512 x 512], [512 x 1 x *] -> [512 x 1 x *]
MPI Rank 1: Validating --> layers[2].Eh._.B = LearnableParameter() :  -> [512 x 1]
MPI Rank 1: Validating --> layers[2].Eh._.z = Plus (layers[2].Eh._.z.PlusArgs[0], layers[2].Eh._.B) : [512 x 1 x *], [512 x 1] -> [512 x 1 x *]
MPI Rank 1: Validating --> layers[2].Eh = Sigmoid (layers[2].Eh._.z) : [512 x 1 x *] -> [512 x 1 x *]
MPI Rank 1: Validating --> outLayer.z.PlusArgs[0] = Times (outLayer.W, layers[2].Eh) : [132 x 512], [512 x 1 x *] -> [132 x 1 x *]
MPI Rank 1: Validating --> outLayer.B = LearnableParameter() :  -> [132 x 1]
MPI Rank 1: Validating --> outZ = Plus (outLayer.z.PlusArgs[0], outLayer.B) : [132 x 1 x *], [132 x 1] -> [132 x 1 x *]
MPI Rank 1: Validating --> CE = CrossEntropyWithSoftmax (labels, outZ) : [132 x *], [132 x 1 x *] -> [1]
MPI Rank 1: Validating --> Err = ClassificationError (labels, outZ) : [132 x *], [132 x 1 x *] -> [1]
MPI Rank 1: Validating --> logPrior._ = Mean (labels) : [132 x *] -> [132]
MPI Rank 1: Validating --> logPrior = Log (logPrior._) : [132] -> [132]
MPI Rank 1: Validating --> ScaledLogLikelihood = Minus (outZ, logPrior) : [132 x 1 x *], [132] -> [132 x 1 x *]
MPI Rank 1: 
MPI Rank 1: Validating network. 17 nodes to process in pass 2.
MPI Rank 1: 
MPI Rank 1: 
MPI Rank 1: Validating network, final pass.
MPI Rank 1: 
MPI Rank 1: 
MPI Rank 1: 
MPI Rank 1: 
MPI Rank 1: Post-processing network complete.
MPI Rank 1: 
MPI Rank 1: Reading script file glob_0000.scp ... 948 entries
MPI Rank 1: HTKDataDeserializer::HTKDataDeserializer: selected 948 utterances grouped into 3 chunks, average chunk size: 316.0 utterances, 84244.7 frames (for I/O: 316.0 utterances, 84244.7 frames)
MPI Rank 1: HTKDataDeserializer::HTKDataDeserializer: determined feature kind as 33-dimensional 'USER' with frame shift 10.0 ms
MPI Rank 1: total 132 state names in state list C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\Data/state.list
MPI Rank 1: htkmlfreader: reading MLF file C:\jenkins\workspace\CNTK-Test-Windows-W1\Tests\EndToEndTests\Speech\Data/glob_0000.mlf ... total 948 entries
MPI Rank 1: MLFDataDeserializer::MLFDataDeserializer: 948 utterances with 252734 frames in 129 classes
MPI Rank 1: 12/21/2016 05:27:46: 
MPI Rank 1: Model has 25 nodes. Using CPU.
MPI Rank 1: 
MPI Rank 1: 12/21/2016 05:27:46: Training criterion:   CE = CrossEntropyWithSoftmax
MPI Rank 1: 12/21/2016 05:27:46: Evaluation criterion: Err = ClassificationError
MPI Rank 1: 
MPI Rank 1: 
MPI Rank 1: Allocating matrices for forward and/or backward propagation.
MPI Rank 1: 
MPI Rank 1: Memory Sharing: Out of 40 matrices, 19 are shared as 8, and 21 are not shared.
MPI Rank 1: 
MPI Rank 1: 	{ layers[1].Eh._.W : [512 x 363] (gradient)
MPI Rank 1: 	  layers[1].Eh._.z : [512 x 1 x *] }
MPI Rank 1: 	{ layers[1].Eh._.z : [512 x 1 x *] (gradient)
MPI Rank 1: 	  layers[2].Eh._.z.PlusArgs[0] : [512 x 1 x *] }
MPI Rank 1: 	{ layers[2].Eh._.W : [512 x 512] (gradient)
MPI Rank 1: 	  layers[2].Eh._.z : [512 x 1 x *] }
MPI Rank 1: 	{ layers[1].Eh : [512 x 1 x *] (gradient)
MPI Rank 1: 	  layers[1].Eh._.B : [512 x 1] (gradient)
MPI Rank 1: 	  layers[2].Eh._.z : [512 x 1 x *] (gradient)
MPI Rank 1: 	  outLayer.z.PlusArgs[0] : [132 x 1 x *] }
MPI Rank 1: 	{ outLayer.W : [132 x 512] (gradient)
MPI Rank 1: 	  outZ : [132 x 1 x *] }
MPI Rank 1: 	{ layers[1].Eh : [512 x 1 x *]
MPI Rank 1: 	  layers[1].Eh._.z.PlusArgs[0] : [512 x *] (gradient) }
MPI Rank 1: 	{ layers[2].Eh : [512 x 1 x *] (gradient)
MPI Rank 1: 	  layers[2].Eh._.B : [512 x 1] (gradient)
MPI Rank 1: 	  outZ : [132 x 1 x *] (gradient) }
MPI Rank 1: 	{ layers[2].Eh : [512 x 1 x *]
MPI Rank 1: 	  layers[2].Eh._.z.PlusArgs[0] : [512 x 1 x *] (gradient) }
MPI Rank 1: 
MPI Rank 1: 
MPI Rank 1: 12/21/2016 05:27:46: Training 516740 parameters in 6 out of 6 parameter tensors and 15 nodes with gradient:
MPI Rank 1: 
MPI Rank 1: 12/21/2016 05:27:46: 	Node 'layers[1].Eh._.B' (LearnableParameter operation) : [512 x 1]
MPI Rank 1: 12/21/2016 05:27:46: 	Node 'layers[1].Eh._.W' (LearnableParameter operation) : [512 x 363]
MPI Rank 1: 12/21/2016 05:27:46: 	Node 'layers[2].Eh._.B' (LearnableParameter operation) : [512 x 1]
MPI Rank 1: 12/21/2016 05:27:46: 	Node 'layers[2].Eh._.W' (LearnableParameter operation) : [512 x 512]
MPI Rank 1: 12/21/2016 05:27:46: 	Node 'outLayer.B' (LearnableParameter operation) : [132 x 1]
MPI Rank 1: 12/21/2016 05:27:46: 	Node 'outLayer.W' (LearnableParameter operation) : [132 x 512]
MPI Rank 1: 
MPI Rank 1: 
MPI Rank 1: 12/21/2016 05:27:46: Precomputing --> 3 PreCompute nodes found.
MPI Rank 1: 
MPI Rank 1: 12/21/2016 05:27:46: 	featNorm.mean = Mean()
MPI Rank 1: 12/21/2016 05:27:46: 	featNorm.invStdDev = InvStdDev()
MPI Rank 1: 12/21/2016 05:27:46: 	logPrior._ = Mean()
MPI Rank 1: 
MPI Rank 1: 12/21/2016 05:27:49: Precomputing --> Completed.
MPI Rank 1: 
MPI Rank 1: 
MPI Rank 1: 12/21/2016 05:27:49: Starting Epoch 1: learning rate per sample = 0.015625  effective momentum = 0.900000  momentum as time constant = 607.4 samples
MPI Rank 1: 
MPI Rank 1: 12/21/2016 05:27:49: Starting minibatch loop.
MPI Rank 1: 12/21/2016 05:27:49:  Epoch[ 1 of 5]-Minibatch[   1-   1, 0.31%]: CE = 4.91295596 * 64; Err = 0.96875000 * 64; time = 0.0323s; samplesPerSecond = 1981.2
MPI Rank 1: 12/21/2016 05:27:49:  Epoch[ 1 of 5]-Minibatch[   2-   2, 0.63%]: CE = 4.78498529 * 64; Err = 1.00000000 * 64; time = 0.0194s; samplesPerSecond = 3299.1
MPI Rank 1: 12/21/2016 05:27:49:  Epoch[ 1 of 5]-Minibatch[   3-   3, 0.94%]: CE = 4.19018696 * 64; Err = 0.81250000 * 64; time = 0.0242s; samplesPerSecond = 2645.4
MPI Rank 1: 12/21/2016 05:27:49:  Epoch[ 1 of 5]-Minibatch[   4-   4, 1.25%]: CE = 4.46135476 * 64; Err = 0.82812500 * 64; time = 0.0249s; samplesPerSecond = 2567.4
MPI Rank 1: 12/21/2016 05:27:49:  Epoch[ 1 of 5]-Minibatch[   5-   5, 1.56%]: CE = 4.72788003 * 64; Err = 0.92187500 * 64; time = 0.0227s; samplesPerSecond = 2824.0
MPI Rank 1: 12/21/2016 05:27:49:  Epoch[ 1 of 5]-Minibatch[   6-   6, 1.88%]: CE = 4.07654096 * 64; Err = 0.89062500 * 64; time = 0.0227s; samplesPerSecond = 2825.1
MPI Rank 1: 12/21/2016 05:27:49:  Epoch[ 1 of 5]-Minibatch[   7-   7, 2.19%]: CE = 4.50165607 * 64; Err = 0.96875000 * 64; time = 0.0235s; samplesPerSecond = 2729.1
MPI Rank 1: 12/21/2016 05:27:49:  Epoch[ 1 of 5]-Minibatch[   8-   8, 2.50%]: CE = 4.93153999 * 64; Err = 0.89062500 * 64; time = 0.0230s; samplesPerSecond = 2788.4
MPI Rank 1: 12/21/2016 05:27:49:  Epoch[ 1 of 5]-Minibatch[   9-   9, 2.81%]: CE = 4.79817443 * 64; Err = 0.93750000 * 64; time = 0.0260s; samplesPerSecond = 2465.9
MPI Rank 1: 12/21/2016 05:27:49:  Epoch[ 1 of 5]-Minibatch[  10-  10, 3.13%]: CE = 4.46089875 * 64; Err = 0.96875000 * 64; time = 0.0178s; samplesPerSecond = 3596.5
MPI Rank 1: 12/21/2016 05:27:49:  Epoch[ 1 of 5]-Minibatch[  11-  11, 3.44%]: CE = 4.34462020 * 64; Err = 0.90625000 * 64; time = 0.0263s; samplesPerSecond = 2431.8
MPI Rank 1: 12/21/2016 05:27:49:  Epoch[ 1 of 5]-Minibatch[  12-  12, 3.75%]: CE = 3.91243070 * 64; Err = 0.87500000 * 64; time = 0.0232s; samplesPerSecond = 2758.4
MPI Rank 1: 12/21/2016 05:27:49:  Epoch[ 1 of 5]-Minibatch[  13-  13, 4.06%]: CE = 4.73715179 * 64; Err = 0.92187500 * 64; time = 0.0248s; samplesPerSecond = 2581.0
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  14-  14, 4.38%]: CE = 4.42160986 * 64; Err = 0.93750000 * 64; time = 0.1507s; samplesPerSecond = 424.7
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  15-  15, 4.69%]: CE = 4.14675744 * 64; Err = 0.85937500 * 64; time = 0.0250s; samplesPerSecond = 2560.5
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  16-  16, 5.00%]: CE = 4.50951186 * 64; Err = 0.95312500 * 64; time = 0.0215s; samplesPerSecond = 2981.6
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  17-  17, 5.31%]: CE = 4.30758210 * 64; Err = 0.85937500 * 64; time = 0.0213s; samplesPerSecond = 3009.9
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  18-  18, 5.63%]: CE = 4.34534841 * 64; Err = 1.00000000 * 64; time = 0.0218s; samplesPerSecond = 2935.9
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  19-  19, 5.94%]: CE = 4.19517128 * 64; Err = 0.96875000 * 64; time = 0.0238s; samplesPerSecond = 2688.8
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  20-  20, 6.25%]: CE = 4.41248710 * 64; Err = 0.98437500 * 64; time = 0.0200s; samplesPerSecond = 3198.1
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  21-  21, 6.56%]: CE = 4.10891079 * 64; Err = 0.92187500 * 64; time = 0.0240s; samplesPerSecond = 2670.8
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  22-  22, 6.88%]: CE = 4.16379766 * 64; Err = 0.85937500 * 64; time = 0.0201s; samplesPerSecond = 3185.0
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  23-  23, 7.19%]: CE = 4.09455579 * 64; Err = 0.92187500 * 64; time = 0.0249s; samplesPerSecond = 2574.7
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  24-  24, 7.50%]: CE = 3.95980469 * 64; Err = 0.89062500 * 64; time = 0.0212s; samplesPerSecond = 3011.9
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  25-  25, 7.81%]: CE = 4.05428109 * 64; Err = 0.87500000 * 64; time = 0.0248s; samplesPerSecond = 2575.6
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  26-  26, 8.13%]: CE = 4.16245451 * 64; Err = 0.84375000 * 64; time = 0.0253s; samplesPerSecond = 2527.0
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  27-  27, 8.44%]: CE = 3.71756327 * 64; Err = 0.84375000 * 64; time = 0.0214s; samplesPerSecond = 2992.9
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  28-  28, 8.75%]: CE = 3.80779138 * 64; Err = 0.87500000 * 64; time = 0.1099s; samplesPerSecond = 582.4
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  29-  29, 9.06%]: CE = 3.72564857 * 64; Err = 0.81250000 * 64; time = 0.0241s; samplesPerSecond = 2656.7
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  30-  30, 9.38%]: CE = 4.01963243 * 64; Err = 0.87500000 * 64; time = 0.0221s; samplesPerSecond = 2895.9
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  31-  31, 9.69%]: CE = 3.68590709 * 64; Err = 0.89062500 * 64; time = 0.0183s; samplesPerSecond = 3492.1
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  32-  32, 10.00%]: CE = 3.81516754 * 64; Err = 0.78125000 * 64; time = 0.1400s; samplesPerSecond = 457.1
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  33-  33, 10.31%]: CE = 3.93685037 * 64; Err = 0.87500000 * 64; time = 0.0240s; samplesPerSecond = 2667.7
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  34-  34, 10.63%]: CE = 3.96481462 * 64; Err = 0.90625000 * 64; time = 0.0235s; samplesPerSecond = 2720.0
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  35-  35, 10.94%]: CE = 3.57865409 * 64; Err = 0.84375000 * 64; time = 0.0241s; samplesPerSecond = 2657.0
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  36-  36, 11.25%]: CE = 3.72265528 * 64; Err = 0.85937500 * 64; time = 0.0225s; samplesPerSecond = 2849.4
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  37-  37, 11.56%]: CE = 3.71485627 * 64; Err = 0.84375000 * 64; time = 0.0268s; samplesPerSecond = 2388.1
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  38-  38, 11.88%]: CE = 4.04042687 * 64; Err = 0.87500000 * 64; time = 0.0246s; samplesPerSecond = 2605.1
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  39-  39, 12.19%]: CE = 3.48663283 * 64; Err = 0.76562500 * 64; time = 0.0245s; samplesPerSecond = 2613.3
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  40-  40, 12.50%]: CE = 3.48828968 * 64; Err = 0.81250000 * 64; time = 0.0243s; samplesPerSecond = 2629.7
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  41-  41, 12.81%]: CE = 3.46883616 * 64; Err = 0.78125000 * 64; time = 0.0243s; samplesPerSecond = 2630.0
MPI Rank 1: 12/21/2016 05:27:50:  Epoch[ 1 of 5]-Minibatch[  42-  42, 13.13%]: CE = 4.12832965 * 64; Err = 0.90625000 * 64; time = 0.0248s; samplesPerSecond = 2582.8
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  43-  43, 13.44%]: CE = 3.82286476 * 64; Err = 0.90625000 * 64; time = 0.0242s; samplesPerSecond = 2646.5
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  44-  44, 13.75%]: CE = 3.99396471 * 64; Err = 0.90625000 * 64; time = 0.0249s; samplesPerSecond = 2573.1
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  45-  45, 14.06%]: CE = 3.84953256 * 64; Err = 0.89062500 * 64; time = 0.0242s; samplesPerSecond = 2650.0
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  46-  46, 14.37%]: CE = 3.57917953 * 64; Err = 0.79687500 * 64; time = 0.0255s; samplesPerSecond = 2513.4
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  47-  47, 14.69%]: CE = 3.86079148 * 64; Err = 0.84375000 * 64; time = 0.0235s; samplesPerSecond = 2719.7
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  48-  48, 15.00%]: CE = 3.88891763 * 64; Err = 0.85937500 * 64; time = 0.0247s; samplesPerSecond = 2590.9
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  49-  49, 15.31%]: CE = 3.94662742 * 64; Err = 0.89062500 * 64; time = 0.0244s; samplesPerSecond = 2620.5
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  50-  50, 15.63%]: CE = 3.83644301 * 64; Err = 0.87500000 * 64; time = 0.0258s; samplesPerSecond = 2483.6
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  51-  51, 15.94%]: CE = 3.66716866 * 64; Err = 0.89062500 * 64; time = 0.0232s; samplesPerSecond = 2760.6
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  52-  52, 16.25%]: CE = 4.00651571 * 64; Err = 0.90625000 * 64; time = 0.1412s; samplesPerSecond = 453.3
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  53-  53, 16.56%]: CE = 3.80511656 * 64; Err = 0.81250000 * 64; time = 0.0227s; samplesPerSecond = 2820.5
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  54-  54, 16.88%]: CE = 3.93380989 * 64; Err = 0.85937500 * 64; time = 0.0251s; samplesPerSecond = 2550.1
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  55-  55, 17.19%]: CE = 3.49394937 * 64; Err = 0.84375000 * 64; time = 0.0244s; samplesPerSecond = 2624.8
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  56-  56, 17.50%]: CE = 3.42224075 * 64; Err = 0.84375000 * 64; time = 0.0245s; samplesPerSecond = 2609.8
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  57-  57, 17.81%]: CE = 3.76078536 * 64; Err = 0.85937500 * 64; time = 0.0240s; samplesPerSecond = 2668.8
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  58-  58, 18.13%]: CE = 3.80639497 * 64; Err = 0.87500000 * 64; time = 0.0245s; samplesPerSecond = 2608.0
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  59-  59, 18.44%]: CE = 3.55543971 * 64; Err = 0.89062500 * 64; time = 0.0242s; samplesPerSecond = 2643.8
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  60-  60, 18.75%]: CE = 3.55947249 * 64; Err = 0.82812500 * 64; time = 0.0260s; samplesPerSecond = 2458.7
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  61-  61, 19.06%]: CE = 3.21133907 * 64; Err = 0.79687500 * 64; time = 0.0270s; samplesPerSecond = 2369.1
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  62-  62, 19.38%]: CE = 3.30807309 * 64; Err = 0.68750000 * 64; time = 0.0245s; samplesPerSecond = 2607.9
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  63-  63, 19.69%]: CE = 3.54643060 * 64; Err = 0.78125000 * 64; time = 0.0244s; samplesPerSecond = 2626.0
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  64-  64, 20.00%]: CE = 3.48819921 * 64; Err = 0.85937500 * 64; time = 0.0249s; samplesPerSecond = 2573.6
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  65-  65, 20.31%]: CE = 3.53098379 * 64; Err = 0.81250000 * 64; time = 0.0241s; samplesPerSecond = 2654.4
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  66-  66, 20.63%]: CE = 3.18218574 * 64; Err = 0.70312500 * 64; time = 0.0245s; samplesPerSecond = 2612.6
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  67-  67, 20.94%]: CE = 3.62919777 * 64; Err = 0.79687500 * 64; time = 0.0241s; samplesPerSecond = 2651.6
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  68-  68, 21.25%]: CE = 3.30344749 * 64; Err = 0.76562500 * 64; time = 0.0249s; samplesPerSecond = 2566.4
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  69-  69, 21.56%]: CE = 3.11192070 * 64; Err = 0.75000000 * 64; time = 0.0246s; samplesPerSecond = 2605.9
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  70-  70, 21.88%]: CE = 3.70063691 * 64; Err = 0.79687500 * 64; time = 0.0243s; samplesPerSecond = 2630.3
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  71-  71, 22.19%]: CE = 3.76244503 * 64; Err = 0.84375000 * 64; time = 0.0231s; samplesPerSecond = 2774.9
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  72-  72, 22.50%]: CE = 3.52103388 * 64; Err = 0.81250000 * 64; time = 0.1380s; samplesPerSecond = 463.6
MPI Rank 1: 12/21/2016 05:27:51:  Epoch[ 1 of 5]-Minibatch[  73-  73, 22.81%]: CE = 3.73227550 * 64; Err = 0.87500000 * 64; time = 0.0231s; samplesPerSecond = 2768.4
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  74-  74, 23.13%]: CE = 3.28056294 * 64; Err = 0.75000000 * 64; time = 0.0247s; samplesPerSecond = 2587.5
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  75-  75, 23.44%]: CE = 3.88497398 * 64; Err = 0.89062500 * 64; time = 0.0242s; samplesPerSecond = 2641.9
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  76-  76, 23.75%]: CE = 3.62146548 * 64; Err = 0.85937500 * 64; time = 0.0244s; samplesPerSecond = 2626.0
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  77-  77, 24.06%]: CE = 3.11930348 * 64; Err = 0.73437500 * 64; time = 0.0242s; samplesPerSecond = 2645.6
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  78-  78, 24.38%]: CE = 3.34530218 * 64; Err = 0.87500000 * 64; time = 0.0245s; samplesPerSecond = 2610.2
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  79-  79, 24.69%]: CE = 3.51426589 * 64; Err = 0.84375000 * 64; time = 0.0242s; samplesPerSecond = 2647.0
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  80-  80, 25.00%]: CE = 3.40713594 * 64; Err = 0.81250000 * 64; time = 0.0251s; samplesPerSecond = 2553.9
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  81-  81, 25.31%]: CE = 3.59134827 * 64; Err = 0.82812500 * 64; time = 0.0240s; samplesPerSecond = 2667.0
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  82-  82, 25.62%]: CE = 3.52703040 * 64; Err = 0.82812500 * 64; time = 0.0245s; samplesPerSecond = 2607.8
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  83-  83, 25.94%]: CE = 3.22259624 * 64; Err = 0.79687500 * 64; time = 0.0242s; samplesPerSecond = 2646.3
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  84-  84, 26.25%]: CE = 3.64961943 * 64; Err = 0.82812500 * 64; time = 0.0248s; samplesPerSecond = 2577.4
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  85-  85, 26.56%]: CE = 3.70782192 * 64; Err = 0.76562500 * 64; time = 0.0241s; samplesPerSecond = 2653.4
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  86-  86, 26.88%]: CE = 3.53921564 * 64; Err = 0.89062500 * 64; time = 0.0247s; samplesPerSecond = 2589.4
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  87-  87, 27.19%]: CE = 3.38712792 * 64; Err = 0.81250000 * 64; time = 0.0243s; samplesPerSecond = 2634.6
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  88-  88, 27.50%]: CE = 3.66470493 * 64; Err = 0.78125000 * 64; time = 0.0248s; samplesPerSecond = 2583.7
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  89-  89, 27.81%]: CE = 3.12758734 * 64; Err = 0.84375000 * 64; time = 0.0241s; samplesPerSecond = 2655.6
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  90-  90, 28.13%]: CE = 3.52072988 * 64; Err = 0.82812500 * 64; time = 0.0248s; samplesPerSecond = 2577.4
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  91-  91, 28.44%]: CE = 3.45630741 * 64; Err = 0.76562500 * 64; time = 0.0243s; samplesPerSecond = 2634.9
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  92-  92, 28.75%]: CE = 3.19535282 * 64; Err = 0.78125000 * 64; time = 0.1455s; samplesPerSecond = 439.7
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  93-  93, 29.06%]: CE = 3.40545723 * 64; Err = 0.81250000 * 64; time = 0.0232s; samplesPerSecond = 2763.0
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  94-  94, 29.38%]: CE = 3.47518793 * 64; Err = 0.70312500 * 64; time = 0.0257s; samplesPerSecond = 2488.8
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  95-  95, 29.69%]: CE = 3.32919398 * 64; Err = 0.78125000 * 64; time = 0.0231s; samplesPerSecond = 2770.1
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  96-  96, 30.00%]: CE = 3.86499937 * 64; Err = 0.93750000 * 64; time = 0.0256s; samplesPerSecond = 2496.4
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  97-  97, 30.31%]: CE = 3.42288014 * 64; Err = 0.84375000 * 64; time = 0.0219s; samplesPerSecond = 2928.4
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  98-  98, 30.63%]: CE = 3.31506114 * 64; Err = 0.82812500 * 64; time = 0.0257s; samplesPerSecond = 2485.7
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[  99-  99, 30.94%]: CE = 3.28863365 * 64; Err = 0.76562500 * 64; time = 0.0243s; samplesPerSecond = 2638.8
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[ 100- 100, 31.25%]: CE = 3.20182099 * 64; Err = 0.76562500 * 64; time = 0.0257s; samplesPerSecond = 2494.2
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[ 101- 101, 31.56%]: CE = 3.75128437 * 64; Err = 0.89062500 * 64; time = 0.0233s; samplesPerSecond = 2745.4
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[ 102- 102, 31.87%]: CE = 3.57333316 * 64; Err = 0.84375000 * 64; time = 0.0246s; samplesPerSecond = 2602.7
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[ 103- 103, 32.19%]: CE = 3.65041879 * 64; Err = 0.81250000 * 64; time = 0.0243s; samplesPerSecond = 2633.3
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[ 104- 104, 32.50%]: CE = 3.45052191 * 64; Err = 0.82812500 * 64; time = 0.0253s; samplesPerSecond = 2534.2
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[ 105- 105, 32.81%]: CE = 3.57278549 * 64; Err = 0.85937500 * 64; time = 0.0244s; samplesPerSecond = 2626.8
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[ 106- 106, 33.13%]: CE = 3.35244169 * 64; Err = 0.87500000 * 64; time = 0.0253s; samplesPerSecond = 2529.6
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[ 107- 107, 33.44%]: CE = 3.29949185 * 64; Err = 0.76562500 * 64; time = 0.0233s; samplesPerSecond = 2751.4
MPI Rank 1: 12/21/2016 05:27:52:  Epoch[ 1 of 5]-Minibatch[ 108- 108, 33.75%]: CE = 3.78609758 * 64; Err = 0.82812500 * 64; time = 0.0250s; samplesPerSecond = 2562.0
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 109- 109, 34.06%]: CE = 3.22622650 * 64; Err = 0.78125000 * 64; time = 0.0237s; samplesPerSecond = 2705.1
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 110- 110, 34.38%]: CE = 3.29821989 * 64; Err = 0.79687500 * 64; time = 0.0253s; samplesPerSecond = 2529.1
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 111- 111, 34.69%]: CE = 3.44143907 * 64; Err = 0.82812500 * 64; time = 0.0239s; samplesPerSecond = 2679.5
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 112- 112, 35.00%]: CE = 3.44276929 * 64; Err = 0.85937500 * 64; time = 0.1666s; samplesPerSecond = 384.2
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 113- 113, 35.31%]: CE = 3.18216790 * 64; Err = 0.76562500 * 64; time = 0.0252s; samplesPerSecond = 2538.8
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 114- 114, 35.63%]: CE = 3.18609709 * 64; Err = 0.78125000 * 64; time = 0.0225s; samplesPerSecond = 2845.6
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 115- 115, 35.94%]: CE = 3.06550821 * 64; Err = 0.73437500 * 64; time = 0.0207s; samplesPerSecond = 3095.5
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 116- 116, 36.25%]: CE = 3.43583629 * 64; Err = 0.79687500 * 64; time = 0.0389s; samplesPerSecond = 1645.3
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 117- 117, 36.56%]: CE = 3.10193105 * 64; Err = 0.78125000 * 64; time = 0.0271s; samplesPerSecond = 2365.4
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 118- 118, 36.88%]: CE = 3.42968418 * 64; Err = 0.81250000 * 64; time = 0.0235s; samplesPerSecond = 2719.8
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 119- 119, 37.19%]: CE = 2.85043824 * 64; Err = 0.60937500 * 64; time = 0.0187s; samplesPerSecond = 3425.8
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 120- 120, 37.50%]: CE = 3.50428373 * 64; Err = 0.85937500 * 64; time = 0.0230s; samplesPerSecond = 2782.6
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 121- 121, 37.81%]: CE = 3.28751701 * 64; Err = 0.82812500 * 64; time = 0.0249s; samplesPerSecond = 2569.2
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 122- 122, 38.13%]: CE = 3.79916343 * 64; Err = 0.89062500 * 64; time = 0.0226s; samplesPerSecond = 2833.5
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 123- 123, 38.44%]: CE = 3.55702537 * 64; Err = 0.82812500 * 64; time = 0.0233s; samplesPerSecond = 2743.0
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 124- 124, 38.75%]: CE = 3.00217445 * 64; Err = 0.71875000 * 64; time = 0.0261s; samplesPerSecond = 2453.8
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 125- 125, 39.06%]: CE = 3.07327108 * 64; Err = 0.75000000 * 64; time = 0.0234s; samplesPerSecond = 2738.1
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 126- 126, 39.38%]: CE = 2.88353063 * 64; Err = 0.59375000 * 64; time = 0.0252s; samplesPerSecond = 2541.8
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 127- 127, 39.69%]: CE = 3.13059468 * 64; Err = 0.79687500 * 64; time = 0.0244s; samplesPerSecond = 2620.2
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 128- 128, 40.00%]: CE = 3.21732650 * 64; Err = 0.85937500 * 64; time = 0.1669s; samplesPerSecond = 383.4
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 129- 129, 40.31%]: CE = 2.97299345 * 64; Err = 0.71875000 * 64; time = 0.0245s; samplesPerSecond = 2608.9
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 130- 130, 40.63%]: CE = 2.93691495 * 64; Err = 0.79687500 * 64; time = 0.0252s; samplesPerSecond = 2543.9
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 131- 131, 40.94%]: CE = 3.31837783 * 64; Err = 0.68750000 * 64; time = 0.0243s; samplesPerSecond = 2631.1
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 132- 132, 41.25%]: CE = 2.91929775 * 64; Err = 0.78125000 * 64; time = 0.0249s; samplesPerSecond = 2565.2
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 133- 133, 41.56%]: CE = 3.07940161 * 64; Err = 0.68750000 * 64; time = 0.0244s; samplesPerSecond = 2623.4
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 134- 134, 41.88%]: CE = 3.28344492 * 64; Err = 0.75000000 * 64; time = 0.0251s; samplesPerSecond = 2545.0
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 135- 135, 42.19%]: CE = 3.18447176 * 64; Err = 0.78125000 * 64; time = 0.0246s; samplesPerSecond = 2602.4
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 136- 136, 42.50%]: CE = 2.79093256 * 64; Err = 0.71875000 * 64; time = 0.0250s; samplesPerSecond = 2562.9
MPI Rank 1: 12/21/2016 05:27:53:  Epoch[ 1 of 5]-Minibatch[ 137- 137, 42.81%]: CE = 2.87937588 * 64; Err = 0.70312500 * 64; time = 0.0245s; samplesPerSecond = 2614.2
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 138- 138, 43.13%]: CE = 2.64594163 * 64; Err = 0.68750000 * 64; time = 0.0251s; samplesPerSecond = 2550.6
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 139- 139, 43.44%]: CE = 2.94206439 * 64; Err = 0.84375000 * 64; time = 0.0245s; samplesPerSecond = 2612.8
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 140- 140, 43.75%]: CE = 3.51285987 * 64; Err = 0.82812500 * 64; time = 0.0250s; samplesPerSecond = 2561.5
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 141- 141, 44.06%]: CE = 3.04888687 * 64; Err = 0.81250000 * 64; time = 0.0247s; samplesPerSecond = 2595.8
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 142- 142, 44.38%]: CE = 3.13123367 * 64; Err = 0.76562500 * 64; time = 0.0250s; samplesPerSecond = 2556.7
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 143- 143, 44.69%]: CE = 2.92926400 * 64; Err = 0.71875000 * 64; time = 0.0248s; samplesPerSecond = 2577.1
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 144- 144, 45.00%]: CE = 3.00144780 * 64; Err = 0.71875000 * 64; time = 0.0248s; samplesPerSecond = 2583.1
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 145- 145, 45.31%]: CE = 2.90962694 * 64; Err = 0.67187500 * 64; time = 0.0245s; samplesPerSecond = 2615.1
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 146- 146, 45.63%]: CE = 3.03283171 * 64; Err = 0.79687500 * 64; time = 0.0252s; samplesPerSecond = 2535.1
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 147- 147, 45.94%]: CE = 3.06942741 * 64; Err = 0.73437500 * 64; time = 0.0256s; samplesPerSecond = 2503.5
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 148- 148, 46.25%]: CE = 2.86661978 * 64; Err = 0.65625000 * 64; time = 0.1578s; samplesPerSecond = 405.6
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 149- 149, 46.56%]: CE = 2.76894440 * 64; Err = 0.68750000 * 64; time = 0.0244s; samplesPerSecond = 2620.4
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 150- 150, 46.88%]: CE = 2.71313692 * 64; Err = 0.59375000 * 64; time = 0.0251s; samplesPerSecond = 2546.0
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 151- 151, 47.19%]: CE = 2.74131048 * 64; Err = 0.65625000 * 64; time = 0.0246s; samplesPerSecond = 2599.6
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 152- 152, 47.50%]: CE = 3.28257238 * 64; Err = 0.71875000 * 64; time = 0.0250s; samplesPerSecond = 2556.1
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 153- 153, 47.81%]: CE = 3.08491448 * 64; Err = 0.76562500 * 64; time = 0.0244s; samplesPerSecond = 2621.4
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 154- 154, 48.13%]: CE = 2.98917665 * 64; Err = 0.71875000 * 64; time = 0.0251s; samplesPerSecond = 2545.6
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 155- 155, 48.44%]: CE = 2.90881148 * 64; Err = 0.81250000 * 64; time = 0.0247s; samplesPerSecond = 2593.6
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 156- 156, 48.75%]: CE = 3.58531995 * 64; Err = 0.78125000 * 64; time = 0.0254s; samplesPerSecond = 2518.7
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 157- 157, 49.06%]: CE = 3.28706069 * 64; Err = 0.75000000 * 64; time = 0.0259s; samplesPerSecond = 2472.1
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 158- 158, 49.38%]: CE = 3.06029676 * 64; Err = 0.81250000 * 64; time = 0.0239s; samplesPerSecond = 2683.0
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 159- 159, 49.69%]: CE = 2.95483403 * 64; Err = 0.68750000 * 64; time = 0.0245s; samplesPerSecond = 2614.0
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 160- 160, 50.00%]: CE = 3.07409648 * 64; Err = 0.76562500 * 64; time = 0.0249s; samplesPerSecond = 2568.6
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 161- 161, 50.31%]: CE = 2.69786051 * 64; Err = 0.67187500 * 64; time = 0.0242s; samplesPerSecond = 2641.5
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 162- 162, 50.63%]: CE = 2.80402381 * 64; Err = 0.70312500 * 64; time = 0.0248s; samplesPerSecond = 2580.4
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 163- 163, 50.94%]: CE = 2.62768914 * 64; Err = 0.62500000 * 64; time = 0.0245s; samplesPerSecond = 2610.3
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 164- 164, 51.25%]: CE = 2.64449167 * 64; Err = 0.67187500 * 64; time = 0.0248s; samplesPerSecond = 2577.6
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 165- 165, 51.56%]: CE = 3.08919011 * 64; Err = 0.79687500 * 64; time = 0.0247s; samplesPerSecond = 2591.9
MPI Rank 1: 12/21/2016 05:27:54:  Epoch[ 1 of 5]-Minibatch[ 166- 166, 51.88%]: CE = 3.07122141 * 64; Err = 0.70312500 * 64; time = 0.0254s; samplesPerSecond = 2516.1
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 167- 167, 52.19%]: CE = 3.05111668 * 64; Err = 0.73437500 * 64; time = 0.1810s; samplesPerSecond = 353.6
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 168- 168, 52.50%]: CE = 2.90345804 * 64; Err = 0.73437500 * 64; time = 0.0236s; samplesPerSecond = 2717.5
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 169- 169, 52.81%]: CE = 2.58801822 * 64; Err = 0.62500000 * 64; time = 0.0225s; samplesPerSecond = 2850.7
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 170- 170, 53.13%]: CE = 2.68278033 * 64; Err = 0.68750000 * 64; time = 0.0156s; samplesPerSecond = 4113.9
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 171- 171, 53.44%]: CE = 2.89664835 * 64; Err = 0.70312500 * 64; time = 0.0178s; samplesPerSecond = 3588.2
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 172- 172, 53.75%]: CE = 2.61913736 * 64; Err = 0.64062500 * 64; time = 0.0207s; samplesPerSecond = 3093.6
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 173- 173, 54.06%]: CE = 2.68386883 * 64; Err = 0.65625000 * 64; time = 0.0204s; samplesPerSecond = 3138.9
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 174- 174, 54.37%]: CE = 2.63044619 * 64; Err = 0.65625000 * 64; time = 0.0223s; samplesPerSecond = 2869.4
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 175- 175, 54.69%]: CE = 2.39899721 * 64; Err = 0.60937500 * 64; time = 0.0239s; samplesPerSecond = 2675.4
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 176- 176, 55.00%]: CE = 2.88430255 * 64; Err = 0.67187500 * 64; time = 0.0204s; samplesPerSecond = 3143.3
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 177- 177, 55.31%]: CE = 2.83595866 * 64; Err = 0.70312500 * 64; time = 0.0211s; samplesPerSecond = 3026.9
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 178- 178, 55.63%]: CE = 2.79519571 * 64; Err = 0.64062500 * 64; time = 0.0231s; samplesPerSecond = 2764.7
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 179- 179, 55.94%]: CE = 2.76600024 * 64; Err = 0.67187500 * 64; time = 0.0205s; samplesPerSecond = 3120.3
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 180- 180, 56.25%]: CE = 2.59895511 * 64; Err = 0.54687500 * 64; time = 0.0216s; samplesPerSecond = 2958.4
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 181- 181, 56.56%]: CE = 2.93763654 * 64; Err = 0.75000000 * 64; time = 0.0214s; samplesPerSecond = 2987.3
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 182- 182, 56.88%]: CE = 2.93634742 * 64; Err = 0.73437500 * 64; time = 0.0230s; samplesPerSecond = 2780.4
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 183- 183, 57.19%]: CE = 2.59901571 * 64; Err = 0.68750000 * 64; time = 0.0219s; samplesPerSecond = 2920.6
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 184- 184, 57.50%]: CE = 2.81753002 * 64; Err = 0.73437500 * 64; time = 0.0234s; samplesPerSecond = 2733.8
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 185- 185, 57.81%]: CE = 3.04424260 * 64; Err = 0.73437500 * 64; time = 0.0226s; samplesPerSecond = 2833.4
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 186- 186, 58.13%]: CE = 2.49622625 * 64; Err = 0.64062500 * 64; time = 0.1602s; samplesPerSecond = 399.5
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 187- 187, 58.44%]: CE = 2.94745408 * 64; Err = 0.71875000 * 64; time = 0.0274s; samplesPerSecond = 2338.0
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 188- 188, 58.75%]: CE = 2.80802583 * 64; Err = 0.71875000 * 64; time = 0.0244s; samplesPerSecond = 2624.0
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 189- 189, 59.06%]: CE = 2.54977638 * 64; Err = 0.67187500 * 64; time = 0.0204s; samplesPerSecond = 3133.0
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 190- 190, 59.38%]: CE = 2.90849909 * 64; Err = 0.68750000 * 64; time = 0.0208s; samplesPerSecond = 3079.6
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 191- 191, 59.69%]: CE = 2.89470021 * 64; Err = 0.71875000 * 64; time = 0.0232s; samplesPerSecond = 2760.4
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 192- 192, 60.00%]: CE = 2.55056761 * 64; Err = 0.64062500 * 64; time = 0.0247s; samplesPerSecond = 2595.5
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 193- 193, 60.31%]: CE = 2.39014720 * 64; Err = 0.59375000 * 64; time = 0.0242s; samplesPerSecond = 2650.1
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 194- 194, 60.62%]: CE = 2.61720826 * 64; Err = 0.65625000 * 64; time = 0.0250s; samplesPerSecond = 2557.3
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 195- 195, 60.94%]: CE = 2.59802571 * 64; Err = 0.65625000 * 64; time = 0.0242s; samplesPerSecond = 2647.9
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 196- 196, 61.25%]: CE = 2.94597696 * 64; Err = 0.75000000 * 64; time = 0.0247s; samplesPerSecond = 2586.3
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 197- 197, 61.56%]: CE = 2.79771307 * 64; Err = 0.75000000 * 64; time = 0.0244s; samplesPerSecond = 2619.9
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 198- 198, 61.88%]: CE = 3.20417932 * 64; Err = 0.71875000 * 64; time = 0.0247s; samplesPerSecond = 2591.9
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 199- 199, 62.19%]: CE = 2.27155558 * 64; Err = 0.53125000 * 64; time = 0.0242s; samplesPerSecond = 2645.5
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 200- 200, 62.50%]: CE = 2.87449908 * 64; Err = 0.68750000 * 64; time = 0.0250s; samplesPerSecond = 2559.5
MPI Rank 1: 12/21/2016 05:27:55:  Epoch[ 1 of 5]-Minibatch[ 201- 201, 62.81%]: CE = 2.71210245 * 64; Err = 0.65625000 * 64; time = 0.0244s; samplesPerSecond = 2623.3
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 202- 202, 63.13%]: CE = 2.44766371 * 64; Err = 0.57812500 * 64; time = 0.1761s; samplesPerSecond = 363.4
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 203- 203, 63.44%]: CE = 2.68243088 * 64; Err = 0.70312500 * 64; time = 0.0213s; samplesPerSecond = 3004.8
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 204- 204, 63.75%]: CE = 2.40962202 * 64; Err = 0.54687500 * 64; time = 0.0159s; samplesPerSecond = 4018.6
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 205- 205, 64.06%]: CE = 2.48400547 * 64; Err = 0.59375000 * 64; time = 0.0152s; samplesPerSecond = 4219.4
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 206- 206, 64.38%]: CE = 2.49121254 * 64; Err = 0.60937500 * 64; time = 0.0255s; samplesPerSecond = 2505.5
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 207- 207, 64.69%]: CE = 2.84691899 * 64; Err = 0.75000000 * 64; time = 0.0166s; samplesPerSecond = 3844.5
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 208- 208, 65.00%]: CE = 2.45273493 * 64; Err = 0.59375000 * 64; time = 0.0264s; samplesPerSecond = 2427.0
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 209- 209, 65.31%]: CE = 2.75036440 * 64; Err = 0.68750000 * 64; time = 0.0224s; samplesPerSecond = 2853.7
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 210- 210, 65.63%]: CE = 2.49555051 * 64; Err = 0.71875000 * 64; time = 0.0240s; samplesPerSecond = 2672.0
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 211- 211, 65.94%]: CE = 2.71109113 * 64; Err = 0.68750000 * 64; time = 0.0220s; samplesPerSecond = 2915.2
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 212- 212, 66.25%]: CE = 2.38218216 * 64; Err = 0.59375000 * 64; time = 0.0265s; samplesPerSecond = 2411.5
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 213- 213, 66.56%]: CE = 2.60308722 * 64; Err = 0.57812500 * 64; time = 0.0222s; samplesPerSecond = 2887.7
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 214- 214, 66.88%]: CE = 2.65611547 * 64; Err = 0.68750000 * 64; time = 0.0265s; samplesPerSecond = 2418.7
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 215- 215, 67.19%]: CE = 2.49633370 * 64; Err = 0.57812500 * 64; time = 0.0222s; samplesPerSecond = 2887.8
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 216- 216, 67.50%]: CE = 2.23315412 * 64; Err = 0.60937500 * 64; time = 0.0263s; samplesPerSecond = 2429.9
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 217- 217, 67.81%]: CE = 2.94093183 * 64; Err = 0.73437500 * 64; time = 0.0222s; samplesPerSecond = 2879.3
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 218- 218, 68.13%]: CE = 2.69840742 * 64; Err = 0.65625000 * 64; time = 0.0245s; samplesPerSecond = 2607.9
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 219- 219, 68.44%]: CE = 2.57215231 * 64; Err = 0.60937500 * 64; time = 0.0230s; samplesPerSecond = 2786.1
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 220- 220, 68.75%]: CE = 2.57160696 * 64; Err = 0.68750000 * 64; time = 0.1523s; samplesPerSecond = 420.1
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 221- 221, 69.06%]: CE = 2.57776681 * 64; Err = 0.65625000 * 64; time = 0.0259s; samplesPerSecond = 2470.6
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 222- 222, 69.38%]: CE = 2.32289644 * 64; Err = 0.57812500 * 64; time = 0.0202s; samplesPerSecond = 3162.2
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 223- 223, 69.69%]: CE = 2.66432343 * 64; Err = 0.70312500 * 64; time = 0.0245s; samplesPerSecond = 2610.1
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 224- 224, 70.00%]: CE = 2.20387606 * 64; Err = 0.65625000 * 64; time = 0.0251s; samplesPerSecond = 2554.5
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 225- 225, 70.31%]: CE = 2.39888933 * 64; Err = 0.59375000 * 64; time = 0.0257s; samplesPerSecond = 2485.5
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 226- 226, 70.63%]: CE = 2.80393339 * 64; Err = 0.70312500 * 64; time = 0.0236s; samplesPerSecond = 2715.0
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 227- 227, 70.94%]: CE = 2.71082242 * 64; Err = 0.68750000 * 64; time = 0.0243s; samplesPerSecond = 2631.5
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 228- 228, 71.25%]: CE = 2.62244612 * 64; Err = 0.70312500 * 64; time = 0.0249s; samplesPerSecond = 2574.3
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 229- 229, 71.56%]: CE = 2.29777087 * 64; Err = 0.62500000 * 64; time = 0.0238s; samplesPerSecond = 2689.9
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 230- 230, 71.88%]: CE = 2.51121239 * 64; Err = 0.65625000 * 64; time = 0.0234s; samplesPerSecond = 2731.9
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 231- 231, 72.19%]: CE = 2.76103008 * 64; Err = 0.64062500 * 64; time = 0.0243s; samplesPerSecond = 2635.6
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 232- 232, 72.50%]: CE = 3.01432561 * 64; Err = 0.76562500 * 64; time = 0.0246s; samplesPerSecond = 2601.4
MPI Rank 1: 12/21/2016 05:27:56:  Epoch[ 1 of 5]-Minibatch[ 233- 233, 72.81%]: CE = 2.99024474 * 64; Err = 0.76562500 * 64; time = 0.0240s; samplesPerSecond = 2663.1
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 234- 234, 73.13%]: CE = 2.86664042 * 64; Err = 0.81250000 * 64; time = 0.0246s; samplesPerSecond = 2598.0
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 235- 235, 73.44%]: CE = 2.60998588 * 64; Err = 0.67187500 * 64; time = 0.0242s; samplesPerSecond = 2646.4
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 236- 236, 73.75%]: CE = 2.18201917 * 64; Err = 0.53125000 * 64; time = 0.0254s; samplesPerSecond = 2523.9
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 237- 237, 74.06%]: CE = 2.17418609 * 64; Err = 0.57812500 * 64; time = 0.0235s; samplesPerSecond = 2721.8
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 238- 238, 74.38%]: CE = 2.25759717 * 64; Err = 0.64062500 * 64; time = 0.1694s; samplesPerSecond = 377.9
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 239- 239, 74.69%]: CE = 2.17788677 * 64; Err = 0.60937500 * 64; time = 0.0200s; samplesPerSecond = 3201.3
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 240- 240, 75.00%]: CE = 2.20328249 * 64; Err = 0.54687500 * 64; time = 0.0246s; samplesPerSecond = 2601.4
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 241- 241, 75.31%]: CE = 2.60590014 * 64; Err = 0.60937500 * 64; time = 0.0220s; samplesPerSecond = 2914.4
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 242- 242, 75.63%]: CE = 2.09884739 * 64; Err = 0.56250000 * 64; time = 0.0265s; samplesPerSecond = 2419.0
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 243- 243, 75.94%]: CE = 2.10587746 * 64; Err = 0.54687500 * 64; time = 0.0221s; samplesPerSecond = 2895.0
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 244- 244, 76.25%]: CE = 2.64457627 * 64; Err = 0.73437500 * 64; time = 0.0249s; samplesPerSecond = 2571.7
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 245- 245, 76.56%]: CE = 2.47600990 * 64; Err = 0.64062500 * 64; time = 0.0230s; samplesPerSecond = 2784.4
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 246- 246, 76.88%]: CE = 2.88789135 * 64; Err = 0.68750000 * 64; time = 0.0220s; samplesPerSecond = 2906.6
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 247- 247, 77.19%]: CE = 2.51823068 * 64; Err = 0.56250000 * 64; time = 0.0183s; samplesPerSecond = 3502.2
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 248- 248, 77.50%]: CE = 2.24877264 * 64; Err = 0.62500000 * 64; time = 0.0297s; samplesPerSecond = 2151.8
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 249- 249, 77.81%]: CE = 2.51043156 * 64; Err = 0.71875000 * 64; time = 0.0182s; samplesPerSecond = 3523.3
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 250- 250, 78.13%]: CE = 2.54234511 * 64; Err = 0.70312500 * 64; time = 0.0259s; samplesPerSecond = 2473.9
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 251- 251, 78.44%]: CE = 2.68548933 * 64; Err = 0.70312500 * 64; time = 0.0215s; samplesPerSecond = 2976.5
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 252- 252, 78.75%]: CE = 2.23175466 * 64; Err = 0.57812500 * 64; time = 0.0254s; samplesPerSecond = 2521.9
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 253- 253, 79.06%]: CE = 2.24553589 * 64; Err = 0.60937500 * 64; time = 0.0209s; samplesPerSecond = 3068.1
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 254- 254, 79.38%]: CE = 2.28765068 * 64; Err = 0.62500000 * 64; time = 0.0254s; samplesPerSecond = 2516.1
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 255- 255, 79.69%]: CE = 2.54161451 * 64; Err = 0.62500000 * 64; time = 0.0248s; samplesPerSecond = 2576.2
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 256- 256, 80.00%]: CE = 2.35401834 * 64; Err = 0.59375000 * 64; time = 0.1580s; samplesPerSecond = 405.1
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 257- 257, 80.31%]: CE = 2.18137731 * 64; Err = 0.56250000 * 64; time = 0.0245s; samplesPerSecond = 2610.5
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 258- 258, 80.63%]: CE = 2.51499174 * 64; Err = 0.59375000 * 64; time = 0.0254s; samplesPerSecond = 2523.2
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 259- 259, 80.94%]: CE = 2.12242410 * 64; Err = 0.65625000 * 64; time = 0.0246s; samplesPerSecond = 2597.7
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 260- 260, 81.25%]: CE = 2.57230724 * 64; Err = 0.68750000 * 64; time = 0.0256s; samplesPerSecond = 2501.1
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 261- 261, 81.56%]: CE = 2.24717210 * 64; Err = 0.57812500 * 64; time = 0.0249s; samplesPerSecond = 2570.5
MPI Rank 1: 12/21/2016 05:27:57:  Epoch[ 1 of 5]-Minibatch[ 262- 262, 81.88%]: CE = 2.46805084 * 64; Err = 0.60937500 * 64; time = 0.0267s; samplesPerSecond = 2394.1
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 263- 263, 82.19%]: CE = 1.94672270 * 64; Err = 0.48437500 * 64; time = 0.0229s; samplesPerSecond = 2793.4
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 264- 264, 82.50%]: CE = 2.34898229 * 64; Err = 0.67187500 * 64; time = 0.0254s; samplesPerSecond = 2521.0
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 265- 265, 82.81%]: CE = 2.19361248 * 64; Err = 0.57812500 * 64; time = 0.0236s; samplesPerSecond = 2711.1
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 266- 266, 83.13%]: CE = 1.96058399 * 64; Err = 0.46875000 * 64; time = 0.0239s; samplesPerSecond = 2675.9
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 267- 267, 83.44%]: CE = 2.02827934 * 64; Err = 0.53125000 * 64; time = 0.0176s; samplesPerSecond = 3644.0
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 268- 268, 83.75%]: CE = 2.16395773 * 64; Err = 0.56250000 * 64; time = 0.0250s; samplesPerSecond = 2555.3
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 269- 269, 84.06%]: CE = 2.42837196 * 64; Err = 0.64062500 * 64; time = 0.0231s; samplesPerSecond = 2773.4
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 270- 270, 84.38%]: CE = 2.56277231 * 64; Err = 0.75000000 * 64; time = 0.0246s; samplesPerSecond = 2606.6
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 271- 271, 84.69%]: CE = 2.35831855 * 64; Err = 0.59375000 * 64; time = 0.0220s; samplesPerSecond = 2907.0
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 272- 272, 85.00%]: CE = 2.48323539 * 64; Err = 0.70312500 * 64; time = 0.0230s; samplesPerSecond = 2780.8
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 273- 273, 85.31%]: CE = 2.66412354 * 64; Err = 0.67187500 * 64; time = 0.0243s; samplesPerSecond = 2632.1
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 274- 274, 85.63%]: CE = 2.35827343 * 64; Err = 0.65625000 * 64; time = 0.0222s; samplesPerSecond = 2885.5
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 275- 275, 85.94%]: CE = 2.35993611 * 64; Err = 0.59375000 * 64; time = 0.0197s; samplesPerSecond = 3245.6
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 276- 276, 86.25%]: CE = 2.27682017 * 64; Err = 0.59375000 * 64; time = 0.1781s; samplesPerSecond = 359.4
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 277- 277, 86.56%]: CE = 2.58742110 * 64; Err = 0.70312500 * 64; time = 0.0211s; samplesPerSecond = 3030.6
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 278- 278, 86.88%]: CE = 2.59364573 * 64; Err = 0.70312500 * 64; time = 0.0235s; samplesPerSecond = 2726.3
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 279- 279, 87.19%]: CE = 2.58154982 * 64; Err = 0.67187500 * 64; time = 0.0318s; samplesPerSecond = 2014.0
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 280- 280, 87.50%]: CE = 2.65251947 * 64; Err = 0.71875000 * 64; time = 0.0227s; samplesPerSecond = 2823.5
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 281- 281, 87.81%]: CE = 2.42794113 * 64; Err = 0.56250000 * 64; time = 0.0247s; samplesPerSecond = 2592.4
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 282- 282, 88.13%]: CE = 2.31306675 * 64; Err = 0.56250000 * 64; time = 0.0240s; samplesPerSecond = 2662.6
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 283- 283, 88.44%]: CE = 2.30780317 * 64; Err = 0.57812500 * 64; time = 0.0254s; samplesPerSecond = 2515.5
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 284- 284, 88.75%]: CE = 2.20092907 * 64; Err = 0.71875000 * 64; time = 0.0225s; samplesPerSecond = 2846.8
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 285- 285, 89.06%]: CE = 2.37127008 * 64; Err = 0.60937500 * 64; time = 0.0238s; samplesPerSecond = 2686.7
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 286- 286, 89.38%]: CE = 1.96581596 * 64; Err = 0.51562500 * 64; time = 0.0188s; samplesPerSecond = 3399.0
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 287- 287, 89.69%]: CE = 2.38139796 * 64; Err = 0.68750000 * 64; time = 0.0252s; samplesPerSecond = 2540.5
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 288- 288, 90.00%]: CE = 2.17378766 * 64; Err = 0.56250000 * 64; time = 0.0229s; samplesPerSecond = 2793.1
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 289- 289, 90.31%]: CE = 2.43769870 * 64; Err = 0.62500000 * 64; time = 0.0252s; samplesPerSecond = 2544.2
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 290- 290, 90.63%]: CE = 1.92877315 * 64; Err = 0.48437500 * 64; time = 0.0225s; samplesPerSecond = 2841.8
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 291- 291, 90.94%]: CE = 2.40592700 * 64; Err = 0.62500000 * 64; time = 0.0234s; samplesPerSecond = 2734.1
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 292- 292, 91.25%]: CE = 2.08578061 * 64; Err = 0.59375000 * 64; time = 0.0262s; samplesPerSecond = 2441.4
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 293- 293, 91.56%]: CE = 2.00803832 * 64; Err = 0.51562500 * 64; time = 0.0234s; samplesPerSecond = 2733.1
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 294- 294, 91.88%]: CE = 2.17692353 * 64; Err = 0.57812500 * 64; time = 0.0249s; samplesPerSecond = 2565.5
MPI Rank 1: 12/21/2016 05:27:58:  Epoch[ 1 of 5]-Minibatch[ 295- 295, 92.19%]: CE = 2.50142509 * 64; Err = 0.70312500 * 64; time = 0.0258s; samplesPerSecond = 2481.0
MPI Rank 1: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 296- 296, 92.50%]: CE = 2.23106504 * 64; Err = 0.60937500 * 64; time = 0.1611s; samplesPerSecond = 397.2
MPI Rank 1: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 297- 297, 92.81%]: CE = 2.15600594 * 64; Err = 0.59375000 * 64; time = 0.0244s; samplesPerSecond = 2620.6
MPI Rank 1: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 298- 298, 93.13%]: CE = 2.57861376 * 64; Err = 0.68750000 * 64; time = 0.0255s; samplesPerSecond = 2505.7
MPI Rank 1: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 299- 299, 93.44%]: CE = 2.07193617 * 64; Err = 0.56250000 * 64; time = 0.0246s; samplesPerSecond = 2604.9
MPI Rank 1: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 300- 300, 93.75%]: CE = 2.16370481 * 64; Err = 0.60937500 * 64; time = 0.0257s; samplesPerSecond = 2492.2
MPI Rank 1: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 301- 301, 94.06%]: CE = 2.24899831 * 64; Err = 0.56250000 * 64; time = 0.0244s; samplesPerSecond = 2624.0
MPI Rank 1: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 302- 302, 94.38%]: CE = 1.87617314 * 64; Err = 0.54687500 * 64; time = 0.0258s; samplesPerSecond = 2476.6
MPI Rank 1: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 303- 303, 94.69%]: CE = 2.22035878 * 64; Err = 0.56250000 * 64; time = 0.0245s; samplesPerSecond = 2610.6
MPI Rank 1: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 304- 304, 95.00%]: CE = 2.23859583 * 64; Err = 0.65625000 * 64; time = 0.0258s; samplesPerSecond = 2482.6
MPI Rank 1: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 305- 305, 95.31%]: CE = 2.36221656 * 64; Err = 0.59375000 * 64; time = 0.0242s; samplesPerSecond = 2648.5
MPI Rank 1: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 306- 306, 95.63%]: CE = 2.11637634 * 64; Err = 0.54687500 * 64; time = 0.0253s; samplesPerSecond = 2528.6
MPI Rank 1: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 307- 307, 95.94%]: CE = 2.32528810 * 64; Err = 0.57812500 * 64; time = 0.0245s; samplesPerSecond = 2607.9
MPI Rank 1: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 308- 308, 96.25%]: CE = 2.06869602 * 64; Err = 0.50000000 * 64; time = 0.0256s; samplesPerSecond = 2497.7
MPI Rank 1: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 309- 309, 96.56%]: CE = 2.10471025 * 64; Err = 0.56250000 * 64; time = 0.0249s; samplesPerSecond = 2565.4
MPI Rank 1: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 310- 310, 96.88%]: CE = 2.69881704 * 64; Err = 0.71875000 * 64; time = 0.0255s; samplesPerSecond = 2513.8
MPI Rank 1: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 311- 311, 97.19%]: CE = 2.21301732 * 64; Err = 0.65625000 * 64; time = 0.0247s; samplesPerSecond = 2586.2
MPI Rank 1: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 312- 312, 97.50%]: CE = 2.34597297 * 64; Err = 0.60937500 * 64; time = 0.0255s; samplesPerSecond = 2512.5
MPI Rank 1: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 313- 313, 97.81%]: CE = 2.08858265 * 64; Err = 0.57812500 * 64; time = 0.0246s; samplesPerSecond = 2596.5
MPI Rank 1: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 314- 314, 98.13%]: CE = 2.10805385 * 64; Err = 0.54687500 * 64; time = 0.0257s; samplesPerSecond = 2491.6
MPI Rank 1: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 315- 315, 98.44%]: CE = 2.29975623 * 64; Err = 0.60937500 * 64; time = 0.1621s; samplesPerSecond = 394.9
MPI Rank 1: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 316- 316, 98.75%]: CE = 2.29188916 * 64; Err = 0.60937500 * 64; time = 0.0188s; samplesPerSecond = 3401.7
MPI Rank 1: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 317- 317, 99.06%]: CE = 2.03062764 * 64; Err = 0.50000000 * 64; time = 0.0238s; samplesPerSecond = 2690.9
MPI Rank 1: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 318- 318, 99.38%]: CE = 2.29874982 * 64; Err = 0.59375000 * 64; time = 0.0250s; samplesPerSecond = 2559.8
MPI Rank 1: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 319- 319, 99.69%]: CE = 2.22342700 * 64; Err = 0.68750000 * 64; time = 0.0244s; samplesPerSecond = 2625.2
MPI Rank 1: 12/21/2016 05:27:59:  Epoch[ 1 of 5]-Minibatch[ 320- 320, 100.00%]: CE = 2.32233814 * 64; Err = 0.59375000 * 64; time = 0.0249s; samplesPerSecond = 2567.5
MPI Rank 1: 12/21/2016 05:27:59: Finished Epoch[ 1 of 5]: [Training] CE = 3.02444900 * 20480; Err = 0.72885742 * 20480; totalSamplesSeen = 20480; learningRatePerSample = 0.015625; epochTime=10.165s
MPI Rank 1: 
MPI Rank 1: 12/21/2016 05:28:00: Starting Epoch 2: learning rate per sample = 0.001953  effective momentum = 0.656119  momentum as time constant = 607.5 samples
MPI Rank 1: Parallel training (2 workers) using BlockMomentumSGD with block momentum = 0.5000, block momentum time constant (per worker) = 2954.6394, block learning rate = 1.0000, block size per worker = 2048 samples, using Nesterov-style block momentum, resetting SGD momentum after sync.
MPI Rank 1: 
MPI Rank 1: 12/21/2016 05:28:00: Starting minibatch loop, distributed reading is ENABLED.
MPI Rank 1: 12/21/2016 05:28:00:  Epoch[ 2 of 5]-Minibatch[   1-   1, 1.25%]: CE = 2.09278351 * 128; Err = 0.57812500 * 128; time = 0.1133s; samplesPerSecond = 1129.3
MPI Rank 1: 12/21/2016 05:28:00:  Epoch[ 2 of 5]-Minibatch[   2-   2, 2.50%]: CE = 2.18078929 * 128; Err = 0.55468750 * 128; time = 0.0808s; samplesPerSecond = 1584.1
MPI Rank 1: 12/21/2016 05:28:00:  Epoch[ 2 of 5]-Minibatch[   3-   3, 3.75%]: CE = 2.16895392 * 128; Err = 0.60937500 * 128; time = 0.0417s; samplesPerSecond = 3067.5
MPI Rank 1: 12/21/2016 05:28:00:  Epoch[ 2 of 5]-Minibatch[   4-   4, 5.00%]: CE = 2.24801623 * 128; Err = 0.60937500 * 128; time = 0.0476s; samplesPerSecond = 2689.0
MPI Rank 1: 12/21/2016 05:28:00:  Epoch[ 2 of 5]-Minibatch[   5-   5, 6.25%]: CE = 2.18628039 * 128; Err = 0.67968750 * 128; time = 0.0459s; samplesPerSecond = 2788.5
MPI Rank 1: 12/21/2016 05:28:00:  Epoch[ 2 of 5]-Minibatch[   6-   6, 7.50%]: CE = 2.36950291 * 128; Err = 0.64062500 * 128; time = 0.0475s; samplesPerSecond = 2692.8
MPI Rank 1: 12/21/2016 05:28:00:  Epoch[ 2 of 5]-Minibatch[   7-   7, 8.75%]: CE = 1.98754400 * 128; Err = 0.59375000 * 128; time = 0.0460s; samplesPerSecond = 2781.6
MPI Rank 1: 12/21/2016 05:28:00:  Epoch[ 2 of 5]-Minibatch[   8-   8, 10.00%]: CE = 2.29285147 * 128; Err = 0.66406250 * 128; time = 0.0475s; samplesPerSecond = 2696.6
MPI Rank 1: 12/21/2016 05:28:00:  Epoch[ 2 of 5]-Minibatch[   9-   9, 11.25%]: CE = 1.98005135 * 128; Err = 0.53906250 * 128; time = 0.0457s; samplesPerSecond = 2799.8
MPI Rank 1: 12/21/2016 05:28:00:  Epoch[ 2 of 5]-Minibatch[  10-  10, 12.50%]: CE = 1.98169796 * 128; Err = 0.51562500 * 128; time = 0.0494s; samplesPerSecond = 2588.7
MPI Rank 1: 12/21/2016 05:28:00:  Epoch[ 2 of 5]-Minibatch[  11-  11, 13.75%]: CE = 2.36498402 * 128; Err = 0.67968750 * 128; time = 0.0438s; samplesPerSecond = 2922.0
MPI Rank 1: 12/21/2016 05:28:00:  Epoch[ 2 of 5]-Minibatch[  12-  12, 15.00%]: CE = 2.19219951 * 128; Err = 0.61718750 * 128; time = 0.1672s; samplesPerSecond = 765.6
MPI Rank 1: 12/21/2016 05:28:00:  Epoch[ 2 of 5]-Minibatch[  13-  13, 16.25%]: CE = 2.06452491 * 128; Err = 0.53906250 * 128; time = 0.0509s; samplesPerSecond = 2514.7
MPI Rank 1: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  14-  14, 17.50%]: CE = 1.94180051 * 128; Err = 0.56250000 * 128; time = 0.0453s; samplesPerSecond = 2824.7
MPI Rank 1: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  15-  15, 18.75%]: CE = 2.15015004 * 128; Err = 0.57031250 * 128; time = 0.0463s; samplesPerSecond = 2766.8
MPI Rank 1: 		(model aggregation stats): 1-th sync point was hit, introducing a 0.01-seconds latency this time; accumulated time on sync point = 0.01 seconds , average latency = 0.01 seconds
MPI Rank 1: 		(model aggregation stats) 1-th sync:     1.03 seconds since last report (0.02 seconds on comm.); 4096 samples processed by 2 workers (2048 by me);
MPI Rank 1: 		(model aggregation stats) 1-th sync: totalThroughput = 3.98k samplesPerSecond , throughputPerWorker = 1.99k samplesPerSecond
MPI Rank 1: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  16-  16, 20.00%]: CE = 1.99832595 * 128; Err = 0.46875000 * 128; time = 0.0955s; samplesPerSecond = 1340.3
MPI Rank 1: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  17-  17, 21.25%]: CE = 2.07071668 * 128; Err = 0.52343750 * 128; time = 0.0442s; samplesPerSecond = 2895.0
MPI Rank 1: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  18-  18, 22.50%]: CE = 1.96928991 * 128; Err = 0.60937500 * 128; time = 0.0311s; samplesPerSecond = 4116.9
MPI Rank 1: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  19-  19, 23.75%]: CE = 2.34985747 * 128; Err = 0.65625000 * 128; time = 0.0456s; samplesPerSecond = 2809.4
MPI Rank 1: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  20-  20, 25.00%]: CE = 1.92885006 * 128; Err = 0.49218750 * 128; time = 0.0451s; samplesPerSecond = 2838.8
MPI Rank 1: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  21-  21, 26.25%]: CE = 2.03928767 * 128; Err = 0.50781250 * 128; time = 0.0458s; samplesPerSecond = 2796.9
MPI Rank 1: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  22-  22, 27.50%]: CE = 2.02655787 * 128; Err = 0.54687500 * 128; time = 0.1803s; samplesPerSecond = 709.8
MPI Rank 1: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  23-  23, 28.75%]: CE = 1.87866448 * 128; Err = 0.52343750 * 128; time = 0.0335s; samplesPerSecond = 3816.6
MPI Rank 1: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  24-  24, 30.00%]: CE = 1.91108869 * 128; Err = 0.57031250 * 128; time = 0.0457s; samplesPerSecond = 2800.1
MPI Rank 1: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  25-  25, 31.25%]: CE = 2.11885017 * 128; Err = 0.59375000 * 128; time = 0.0389s; samplesPerSecond = 3287.3
MPI Rank 1: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  26-  26, 32.50%]: CE = 2.12920514 * 128; Err = 0.57812500 * 128; time = 0.0435s; samplesPerSecond = 2940.8
MPI Rank 1: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  27-  27, 33.75%]: CE = 2.00735878 * 128; Err = 0.57812500 * 128; time = 0.0439s; samplesPerSecond = 2915.9
MPI Rank 1: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  28-  28, 35.00%]: CE = 2.34481434 * 128; Err = 0.62500000 * 128; time = 0.0394s; samplesPerSecond = 3244.7
MPI Rank 1: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  29-  29, 36.25%]: CE = 2.20659008 * 128; Err = 0.57812500 * 128; time = 0.0371s; samplesPerSecond = 3450.2
MPI Rank 1: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  30-  30, 37.50%]: CE = 2.13527297 * 128; Err = 0.57031250 * 128; time = 0.0454s; samplesPerSecond = 2820.0
MPI Rank 1: 12/21/2016 05:28:01:  Epoch[ 2 of 5]-Minibatch[  31-  31, 38.75%]: CE = 2.13058781 * 128; Err = 0.57812500 * 128; time = 0.0457s; samplesPerSecond = 2800.4
MPI Rank 1: 		(model aggregation stats): 2-th sync point was hit, introducing a 0.01-seconds latency this time; accumulated time on sync point = 0.03 seconds , average latency = 0.01 seconds
MPI Rank 1: 		(model aggregation stats) 2-th sync:     1.01 seconds since last report (0.06 seconds on comm.); 4096 samples processed by 2 workers (2048 by me);
MPI Rank 1: 		(model aggregation stats) 2-th sync: totalThroughput = 4.06k samplesPerSecond , throughputPerWorker = 2.03k samplesPerSecond
MPI Rank 1: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  32-  32, 40.00%]: CE = 1.87502810 * 128; Err = 0.56250000 * 128; time = 0.2367s; samplesPerSecond = 540.9
MPI Rank 1: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  33-  33, 41.25%]: CE = 1.91173305 * 128; Err = 0.52343750 * 128; time = 0.0469s; samplesPerSecond = 2729.6
MPI Rank 1: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  34-  34, 42.50%]: CE = 2.24952705 * 128; Err = 0.61718750 * 128; time = 0.0460s; samplesPerSecond = 2780.0
MPI Rank 1: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  35-  35, 43.75%]: CE = 1.95090577 * 128; Err = 0.53906250 * 128; time = 0.0440s; samplesPerSecond = 2909.4
MPI Rank 1: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  36-  36, 45.00%]: CE = 2.25694794 * 128; Err = 0.61718750 * 128; time = 0.0397s; samplesPerSecond = 3228.2
MPI Rank 1: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  37-  37, 46.25%]: CE = 1.93963435 * 128; Err = 0.52343750 * 128; time = 0.0357s; samplesPerSecond = 3586.4
MPI Rank 1: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  38-  38, 47.50%]: CE = 1.97756156 * 128; Err = 0.47656250 * 128; time = 0.0388s; samplesPerSecond = 3296.7
MPI Rank 1: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  39-  39, 48.75%]: CE = 1.95177783 * 128; Err = 0.51562500 * 128; time = 0.0369s; samplesPerSecond = 3465.9
MPI Rank 1: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  40-  40, 50.00%]: CE = 1.92424985 * 128; Err = 0.55468750 * 128; time = 0.0369s; samplesPerSecond = 3473.3
MPI Rank 1: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  41-  41, 51.25%]: CE = 1.86182077 * 128; Err = 0.50000000 * 128; time = 0.0368s; samplesPerSecond = 3479.3
MPI Rank 1: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  42-  42, 52.50%]: CE = 1.87138664 * 128; Err = 0.57812500 * 128; time = 0.0393s; samplesPerSecond = 3260.1
MPI Rank 1: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  43-  43, 53.75%]: CE = 2.10051544 * 128; Err = 0.58593750 * 128; time = 0.0332s; samplesPerSecond = 3851.5
MPI Rank 1: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  44-  44, 55.00%]: CE = 1.73668662 * 128; Err = 0.43750000 * 128; time = 0.1602s; samplesPerSecond = 799.2
MPI Rank 1: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  45-  45, 56.25%]: CE = 2.25252786 * 128; Err = 0.58593750 * 128; time = 0.0362s; samplesPerSecond = 3534.5
MPI Rank 1: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  46-  46, 57.50%]: CE = 1.93265905 * 128; Err = 0.53906250 * 128; time = 0.0367s; samplesPerSecond = 3483.0
MPI Rank 1: 12/21/2016 05:28:02:  Epoch[ 2 of 5]-Minibatch[  47-  47, 58.75%]: CE = 1.92026864 * 128; Err = 0.53125000 * 128; time = 0.0371s; samplesPerSecond = 3446.4
MPI Rank 1: 		(model aggregation stats): 3-th sync point was hit, introducing a 0.02-seconds latency this time; accumulated time on sync point = 0.05 seconds , average latency = 0.02 seconds
MPI Rank 1: 		(model aggregation stats) 3-th sync:     0.85 seconds since last report (0.07 seconds on comm.); 4096 samples processed by 2 workers (2048 by me);
MPI Rank 1: 		(model aggregation stats) 3-th sync: totalThroughput = 4.81k samplesPerSecond , throughputPerWorker = 2.40k samplesPerSecond
MPI Rank 1: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  48-  48, 60.00%]: CE = 1.96897429 * 128; Err = 0.50781250 * 128; time = 0.1407s; samplesPerSecond = 909.5
MPI Rank 1: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  49-  49, 61.25%]: CE = 1.62916168 * 128; Err = 0.43750000 * 128; time = 0.0449s; samplesPerSecond = 2847.9
MPI Rank 1: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  50-  50, 62.50%]: CE = 2.08670592 * 128; Err = 0.64843750 * 128; time = 0.0425s; samplesPerSecond = 3013.5
MPI Rank 1: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  51-  51, 63.75%]: CE = 1.92556235 * 128; Err = 0.49218750 * 128; time = 0.0470s; samplesPerSecond = 2725.8
MPI Rank 1: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  52-  52, 65.00%]: CE = 2.00072664 * 128; Err = 0.53125000 * 128; time = 0.0434s; samplesPerSecond = 2946.9
MPI Rank 1: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  53-  53, 66.25%]: CE = 2.06829027 * 128; Err = 0.56250000 * 128; time = 0.1667s; samplesPerSecond = 767.6
MPI Rank 1: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  54-  54, 67.50%]: CE = 1.83341656 * 128; Err = 0.50000000 * 128; time = 0.0421s; samplesPerSecond = 3037.1
MPI Rank 1: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  55-  55, 68.75%]: CE = 1.89431569 * 128; Err = 0.57812500 * 128; time = 0.0550s; samplesPerSecond = 2327.5
MPI Rank 1: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  56-  56, 70.00%]: CE = 1.98866240 * 128; Err = 0.53125000 * 128; time = 0.0450s; samplesPerSecond = 2843.2
MPI Rank 1: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  57-  57, 71.25%]: CE = 1.83148964 * 37; Err = 0.56756757 * 37; time = 0.0132s; samplesPerSecond = 2793.3
MPI Rank 1: 12/21/2016 05:28:03:  Epoch[ 2 of 5]-Minibatch[  58-  58, 72.50%]: CE = 0.00000000 * 0; Err = 0.00000000 * 0; time = 3.9000e-005s; samplesPerSecond = 0.0
MPI Rank 1: 		(model aggregation stats): 4-th sync point was hit, introducing a 0.19-seconds latency this time; accumulated time on sync point = 0.24 seconds , average latency = 0.06 seconds
MPI Rank 1: 		(model aggregation stats) 4-th sync:     1.92 seconds since last report (1.20 seconds on comm.); 8192 samples processed by 2 workers (1061 by me);
MPI Rank 1: 		(model aggregation stats) 4-th sync: totalThroughput = 4.28k samplesPerSecond , throughputPerWorker = 2.14k samplesPerSecond
MPI Rank 1: 12/21/2016 05:28:04: Finished Epoch[ 2 of 5]: [Training] CE = 2.03791679 * 20480; Err = 0.55712891 * 20480; totalSamplesSeen = 40960; learningRatePerSample = 0.001953125; epochTime=4.81589s
MPI Rank 1: 
MPI Rank 1: 12/21/2016 05:28:05: Starting Epoch 3: learning rate per sample = 0.000098  effective momentum = 0.656119  momentum as time constant = 2429.9 samples
MPI Rank 1: Parallel training (2 workers) using BlockMomentumSGD with block momentum = 0.5000, block momentum time constant (per worker) = 2954.6394, block learning rate = 1.0000, block size per worker = 2048 samples, using Nesterov-style block momentum, resetting SGD momentum after sync.
MPI Rank 1: 
MPI Rank 1: 12/21/2016 05:28:05: Starting minibatch loop, distributed reading is ENABLED.
MPI Rank 1: 12/21/2016 05:28:05:  Epoch[ 3 of 5]-Minibatch[   1-   1, 5.00%]: CE = 1.94395242 * 512; Err = 0.50585938 * 512; time = 0.1732s; samplesPerSecond = 2955.9
MPI Rank 1: 12/21/2016 05:28:05:  Epoch[ 3 of 5]-Minibatch[   2-   2, 10.00%]: CE = 1.91829301 * 512; Err = 0.54101563 * 512; time = 0.1115s; samplesPerSecond = 4591.9
MPI Rank 1: 12/21/2016 05:28:05:  Epoch[ 3 of 5]-Minibatch[   3-   3, 15.00%]: CE = 1.87474368 * 512; Err = 0.53125000 * 512; time = 0.1332s; samplesPerSecond = 3845.1
MPI Rank 1: 		(model aggregation stats): 1-th sync point was hit, introducing a 0.17-seconds latency this time; accumulated time on sync point = 0.17 seconds , average latency = 0.17 seconds
MPI Rank 1: 		(model aggregation stats) 1-th sync:     0.84 seconds since last report (0.07 seconds on comm.); 4096 samples processed by 2 workers (2048 by me);
MPI Rank 1: 		(model aggregation stats) 1-th sync: totalThroughput = 4.88k samplesPerSecond , throughputPerWorker = 2.44k samplesPerSecond
MPI Rank 1: 12/21/2016 05:28:05:  Epoch[ 3 of 5]-Minibatch[   4-   4, 20.00%]: CE = 1.86118062 * 512; Err = 0.52148438 * 512; time = 0.4133s; samplesPerSecond = 1238.7
MPI Rank 1: 12/21/2016 05:28:06:  Epoch[ 3 of 5]-Minibatch[   5-   5, 25.00%]: CE = 1.89327895 * 512; Err = 0.51562500 * 512; time = 0.1524s; samplesPerSecond = 3360.3
MPI Rank 1: 12/21/2016 05:28:06:  Epoch[ 3 of 5]-Minibatch[   6-   6, 30.00%]: CE = 1.95084124 * 512; Err = 0.53320313 * 512; time = 0.1461s; samplesPerSecond = 3505.1
MPI Rank 1: 12/21/2016 05:28:06:  Epoch[ 3 of 5]-Minibatch[   7-   7, 35.00%]: CE = 2.01212462 * 512; Err = 0.55859375 * 512; time = 0.2568s; samplesPerSecond = 1993.7
MPI Rank 1: 		(model aggregation stats): 2-th sync point was hit, introducing a 0.00-seconds latency this time; accumulated time on sync point = 0.17 seconds , average latency = 0.08 seconds
MPI Rank 1: 		(model aggregation stats) 2-th sync:     0.80 seconds since last report (0.06 seconds on comm.); 4096 samples processed by 2 workers (2048 by me);
MPI Rank 1: 		(model aggregation stats) 2-th sync: totalThroughput = 5.15k samplesPerSecond , throughputPerWorker = 2.57k samplesPerSecond
MPI Rank 1: 12/21/2016 05:28:06:  Epoch[ 3 of 5]-Minibatch[   8-   8, 40.00%]: CE = 1.92556904 * 512; Err = 0.58007813 * 512; time = 0.2391s; samplesPerSecond = 2141.4
MPI Rank 1: 12/21/2016 05:28:06:  Epoch[ 3 of 5]-Minibatch[   9-   9, 45.00%]: CE = 1.91264292 * 512; Err = 0.51171875 * 512; time = 0.1259s; samplesPerSecond = 4065.6
MPI Rank 1: 12/21/2016 05:28:07:  Epoch[ 3 of 5]-Minibatch[  10-  10, 50.00%]: CE = 1.89071222 * 512; Err = 0.52929688 * 512; time = 0.2591s; samplesPerSecond = 1975.9
MPI Rank 1: 12/21/2016 05:28:07:  Epoch[ 3 of 5]-Minibatch[  11-  11, 55.00%]: CE = 2.10267317 * 512; Err = 0.60546875 * 512; time = 0.1370s; samplesPerSecond = 3737.2
MPI Rank 1: 		(model aggregation stats): 3-th sync point was hit, introducing a 0.05-seconds latency this time; accumulated time on sync point = 0.21 seconds , average latency = 0.07 seconds
MPI Rank 1: 		(model aggregation stats) 3-th sync:     0.83 seconds since last report (0.07 seconds on comm.); 4096 samples processed by 2 workers (2048 by me);
MPI Rank 1: 		(model aggregation stats) 3-th sync: totalThroughput = 4.93k samplesPerSecond , throughputPerWorker = 2.47k samplesPerSecond
MPI Rank 1: 12/21/2016 05:28:07:  Epoch[ 3 of 5]-Minibatch[  12-  12, 60.00%]: CE = 1.94130255 * 512; Err = 0.53515625 * 512; time = 0.3070s; samplesPerSecond = 1667.9
MPI Rank 1: 12/21/2016 05:28:07:  Epoch[ 3 of 5]-Minibatch[  13-  13, 65.00%]: CE = 1.93636276 * 512; Err = 0.54687500 * 512; time = 0.2009s; samplesPerSecond = 2549.0
MPI Rank 1: 12/21/2016 05:28:07:  Epoch[ 3 of 5]-Minibatch[  14-  14, 70.00%]: CE = 1.94802346 * 512; Err = 0.56250000 * 512; time = 0.1523s; samplesPerSecond = 3360.7
MPI Rank 1: 12/21/2016 05:28:08:  Epoch[ 3 of 5]-Minibatch[  15-  15, 75.00%]: CE = 1.98562867 * 512; Err = 0.54101563 * 512; time = 0.2472s; samplesPerSecond = 2071.2
MPI Rank 1: 12/21/2016 05:28:08:  Epoch[ 3 of 5]-Minibatch[  16-  16, 80.00%]: CE = 1.79595464 * 11; Err = 0.54545455 * 11; time = 0.0405s; samplesPerSecond = 271.9
MPI Rank 1: 12/21/2016 05:28:08:  Epoch[ 3 of 5]-Minibatch[  17-  17, 85.00%]: CE = 0.00000000 * 0; Err = 0.00000000 * 0; time = 4.0000e-005s; samplesPerSecond = 0.0
MPI Rank 1: 		(model aggregation stats): 4-th sync point was hit, introducing a 0.00-seconds latency this time; accumulated time on sync point = 0.21 seconds , average latency = 0.05 seconds
MPI Rank 1: 		(model aggregation stats) 4-th sync:     1.47 seconds since last report (0.80 seconds on comm.); 8192 samples processed by 2 workers (1547 by me);
MPI Rank 1: 		(model aggregation stats) 4-th sync: totalThroughput = 5.59k samplesPerSecond , throughputPerWorker = 2.79k samplesPerSecond
MPI Rank 1: 12/21/2016 05:28:09: Finished Epoch[ 3 of 5]: [Training] CE = 1.94906734 * 20480; Err = 0.54541016 * 20480; totalSamplesSeen = 61440; learningRatePerSample = 9.7656251e-005; epochTime=3.9332s
MPI Rank 1: ValidateSubNetwork: featNorm.invStdDev InvStdDev operation changed, from [363 x 1] to [363].ValidateSubNetwork: featNorm.mean Mean operation changed, from [363 x 1] to [363].ValidateSubNetwork: logPrior._ Mean operation changed, from [132 x 1] to [132].Parallel training (2 workers) using BlockMomentumSGD with block momentum = 0.5000, block momentum time constant (per worker) = 2954.6394, block learning rate = 1.0000, block size per worker = 2048 samples, using Nesterov-style block momentum, resetting SGD momentum after sync.
MPI Rank 1: 
MPI Rank 1: 12/21/2016 05:28:09: Starting minibatch loop, distributed reading is ENABLED.
MPI Rank 1: 12/21/2016 05:28:09:   BaseAdaptiveLearnRateSearch: Epoch[ 4 of 5]-Minibatch[   1-   1, 5.00%]: CE = 1.95468900 * 512; Err = 0.52929688 * 512; time = 0.1134s; samplesPerSecond = 4515.2
MPI Rank 1: 12/21/2016 05:28:09:   BaseAdaptiveLearnRateSearch: Epoch[ 4 of 5]-Minibatch[   2-   2, 10.00%]: CE = 0.00000000 * 0; Err = 0.00000000 * 0; time = 8.3000e-005s; samplesPerSecond = 0.0
MPI Rank 1: 		(model aggregation stats): 1-th sync point was hit, introducing a 0.04-seconds latency this time; accumulated time on sync point = 0.04 seconds , average latency = 0.04 seconds
MPI Rank 1: 		(model aggregation stats) 1-th sync:     0.25 seconds since last report (0.06 seconds on comm.); 1536 samples processed by 2 workers (512 by me);
MPI Rank 1: 		(model aggregation stats) 1-th sync: totalThroughput = 6.22k samplesPerSecond , throughputPerWorker = 3.11k samplesPerSecond
MPI Rank 1: 12/21/2016 05:28:09:  Finished Mini-Epoch[4]: CE = 1.91381689 * 1536; Err = 0.52148438 * 1536; learningRatePerSample = 0; minibatchSize = 1024
MPI Rank 1: ValidateSubNetwork: featNorm.invStdDev InvStdDev operation changed, from [363 x 1] to [363].ValidateSubNetwork: featNorm.mean Mean operation changed, from [363 x 1] to [363].ValidateSubNetwork: logPrior._ Mean operation changed, from [132 x 1] to [132].Parallel training (2 workers) using BlockMomentumSGD with block momentum = 0.5000, block momentum time constant (per worker) = 2954.6394, block learning rate = 1.0000, block size per worker = 2048 samples, using Nesterov-style block momentum, resetting SGD momentum after sync.
MPI Rank 1: 
MPI Rank 1: 12/21/2016 05:28:09: Starting minibatch loop, distributed reading is ENABLED.
MPI Rank 1: 12/21/2016 05:28:10:   AdaptiveLearnRateSearch: Epoch[ 4 of 5]-Minibatch[   1-   1, 5.00%]: CE = 1.95468900 * 512; Err = 0.52929688 * 512; time = 0.1622s; samplesPerSecond = 3156.7
MPI Rank 1: 12/21/2016 05:28:10:   AdaptiveLearnRateSearch: Epoch[ 4 of 5]-Minibatch[   2-   2, 10.00%]: CE = 0.00000000 * 0; Err = 0.00000000 * 0; time = 9.0000e-005s; samplesPerSecond = 0.0
MPI Rank 1: 		(model aggregation stats): 1-th sync point was hit, introducing a 0.09-seconds latency this time; accumulated time on sync point = 0.09 seconds , average latency = 0.09 seconds
MPI Rank 1: 		(model aggregation stats) 1-th sync:     0.37 seconds since last report (0.07 seconds on comm.); 1536 samples processed by 2 workers (512 by me);
MPI Rank 1: 		(model aggregation stats) 1-th sync: totalThroughput = 4.20k samplesPerSecond , throughputPerWorker = 2.10k samplesPerSecond
MPI Rank 1: 12/21/2016 05:28:10:  Finished Mini-Epoch[4]: CE = 1.91308005 * 1536; Err = 0.52083333 * 1536; learningRatePerSample = 0.0039062502; minibatchSize = 1024
MPI Rank 1: ValidateSubNetwork: featNorm.invStdDev InvStdDev operation changed, from [363 x 1] to [363].ValidateSubNetwork: featNorm.mean Mean operation changed, from [363 x 1] to [363].ValidateSubNetwork: logPrior._ Mean operation changed, from [132 x 1] to [132].12/21/2016 05:28:10:  SearchForBestLearnRate Epoch[4]: Best learningRatePerSample = 0.003906250186, baseCriterion=1.922629502
MPI Rank 1: 
MPI Rank 1: 12/21/2016 05:28:10: Starting Epoch 4: learning rate per sample = 0.003906  effective momentum = 0.656119  momentum as time constant = 2429.9 samples
MPI Rank 1: Parallel training (2 workers) using BlockMomentumSGD with block momentum = 0.5000, block momentum time constant (per worker) = 2954.6394, block learning rate = 1.0000, block size per worker = 2048 samples, using Nesterov-style block momentum, resetting SGD momentum after sync.
MPI Rank 1: 
MPI Rank 1: 12/21/2016 05:28:10: Starting minibatch loop, distributed reading is ENABLED.
MPI Rank 1: 12/21/2016 05:28:10:  Epoch[ 4 of 5]-Minibatch[   1-   1, 5.00%]: CE = 1.95468900 * 512; Err = 0.52929688 * 512; time = 0.1454s; samplesPerSecond = 3522.0
MPI Rank 1: 12/21/2016 05:28:11:  Epoch[ 4 of 5]-Minibatch[   2-   2, 10.00%]: CE = 1.87167107 * 512; Err = 0.53515625 * 512; time = 0.1782s; samplesPerSecond = 2872.6
MPI Rank 1: 12/21/2016 05:28:11:  Epoch[ 4 of 5]-Minibatch[   3-   3, 15.00%]: CE = 1.88950637 * 512; Err = 0.52734375 * 512; time = 0.1927s; samplesPerSecond = 2656.6
MPI Rank 1: 		(model aggregation stats): 1-th sync point was hit, introducing a 0.08-seconds latency this time; accumulated time on sync point = 0.08 seconds , average latency = 0.08 seconds
MPI Rank 1: 		(model aggregation stats) 1-th sync:     0.82 seconds since last report (0.08 seconds on comm.); 4096 samples processed by 2 workers (2048 by me);
MPI Rank 1: 		(model aggregation stats) 1-th sync: totalThroughput = 4.97k samplesPerSecond , throughputPerWorker = 2.49k samplesPerSecond
MPI Rank 1: 12/21/2016 05:28:11:  Epoch[ 4 of 5]-Minibatch[   4-   4, 20.00%]: CE = 1.88732813 * 512; Err = 0.52148438 * 512; time = 0.2991s; samplesPerSecond = 1711.8
MPI Rank 1: 12/21/2016 05:28:11:  Epoch[ 4 of 5]-Minibatch[   5-   5, 25.00%]: CE = 1.97210123 * 512; Err = 0.54296875 * 512; time = 0.3571s; samplesPerSecond = 1433.8
MPI Rank 1: 12/21/2016 05:28:12:  Epoch[ 4 of 5]-Minibatch[   6-   6, 30.00%]: CE = 1.85454102 * 512; Err = 0.50390625 * 512; time = 0.1888s; samplesPerSecond = 2712.3
MPI Rank 1: 12/21/2016 05:28:12:  Epoch[ 4 of 5]-Minibatch[   7-   7, 35.00%]: CE = 1.92811151 * 512; Err = 0.52148438 * 512; time = 0.1986s; samplesPerSecond = 2577.8
MPI Rank 1: 		(model aggregation stats): 2-th sync point was hit, introducing a 0.00-seconds latency this time; accumulated time on sync point = 0.08 seconds , average latency = 0.04 seconds
MPI Rank 1: 		(model aggregation stats) 2-th sync:     0.92 seconds since last report (0.07 seconds on comm.); 4096 samples processed by 2 workers (2048 by me);
MPI Rank 1: 		(model aggregation stats) 2-th sync: totalThroughput = 4.48k samplesPerSecond , throughputPerWorker = 2.24k samplesPerSecond
MPI Rank 1: 12/21/2016 05:28:12:  Epoch[ 4 of 5]-Minibatch[   8-   8, 40.00%]: CE = 1.82157074 * 512; Err = 0.50195313 * 512; time = 0.1689s; samplesPerSecond = 3032.2
MPI Rank 1: 12/21/2016 05:28:12:  Epoch[ 4 of 5]-Minibatch[   9-   9, 45.00%]: CE = 1.78976739 * 512; Err = 0.48046875 * 512; time = 0.1129s; samplesPerSecond = 4534.8
MPI Rank 1: 12/21/2016 05:28:12:  Epoch[ 4 of 5]-Minibatch[  10-  10, 50.00%]: CE = 1.88394408 * 512; Err = 0.54296875 * 512; time = 0.1528s; samplesPerSecond = 3350.2
MPI Rank 1: 12/21/2016 05:28:13:  Epoch[ 4 of 5]-Minibatch[  11-  11, 55.00%]: CE = 2.00101366 * 512; Err = 0.54492188 * 512; time = 0.2656s; samplesPerSecond = 1928.0
MPI Rank 1: 		(model aggregation stats): 3-th sync point was hit, introducing a 0.05-seconds latency this time; accumulated time on sync point = 0.13 seconds , average latency = 0.04 seconds
MPI Rank 1: 		(model aggregation stats) 3-th sync:     0.82 seconds since last report (0.06 seconds on comm.); 4096 samples processed by 2 workers (2048 by me);
MPI Rank 1: 		(model aggregation stats) 3-th sync: totalThroughput = 4.97k samplesPerSecond , throughputPerWorker = 2.49k samplesPerSecond
MPI Rank 1: 12/21/2016 05:28:13:  Epoch[ 4 of 5]-Minibatch[  12-  12, 60.00%]: CE = 1.78858571 * 512; Err = 0.51953125 * 512; time = 0.2906s; samplesPerSecond = 1762.1
MPI Rank 1: 12/21/2016 05:28:13:  Epoch[ 4 of 5]-Minibatch[  13-  13, 65.00%]: CE = 1.85860025 * 512; Err = 0.53515625 * 512; time = 0.2479s; samplesPerSecond = 2065.4
MPI Rank 1: 12/21/2016 05:28:13:  Epoch[ 4 of 5]-Minibatch[  14-  14, 70.00%]: CE = 1.91177316 * 512; Err = 0.54492188 * 512; time = 0.1516s; samplesPerSecond = 3377.3
MPI Rank 1: 12/21/2016 05:28:13:  Epoch[ 4 of 5]-Minibatch[  15-  15, 75.00%]: CE = 1.74248464 * 512; Err = 0.50000000 * 512; time = 0.1269s; samplesPerSecond = 4033.8
MPI Rank 1: 12/21/2016 05:28:13:  Epoch[ 4 of 5]-Minibatch[  16-  16, 80.00%]: CE = 1.98287530 * 206; Err = 0.56796117 * 206; time = 0.0483s; samplesPerSecond = 4268.6
MPI Rank 1: 12/21/2016 05:28:13:  Epoch[ 4 of 5]-Minibatch[  17-  17, 85.00%]: CE = 0.00000000 * 0; Err = 0.00000000 * 0; time = 3.8000e-005s; samplesPerSecond = 0.0
MPI Rank 1: 		(model aggregation stats): 4-th sync point was hit, introducing a 0.25-seconds latency this time; accumulated time on sync point = 0.38 seconds , average latency = 0.10 seconds
MPI Rank 1: 		(model aggregation stats) 4-th sync:     1.59 seconds since last report (0.75 seconds on comm.); 8192 samples processed by 2 workers (1742 by me);
MPI Rank 1: 		(model aggregation stats) 4-th sync: totalThroughput = 5.14k samplesPerSecond , throughputPerWorker = 2.57k samplesPerSecond
MPI Rank 1: 12/21/2016 05:28:14: Finished Epoch[ 4 of 5]: [Training] CE = 1.86802921 * 20480; Err = 0.52246094 * 20480; totalSamplesSeen = 81920; learningRatePerSample = 0.0039062502; epochTime=5.64752s
MPI Rank 1: ValidateSubNetwork: featNorm.invStdDev InvStdDev operation changed, from [363 x 1] to [363].ValidateSubNetwork: featNorm.mean Mean operation changed, from [363 x 1] to [363].ValidateSubNetwork: logPrior._ Mean operation changed, from [132 x 1] to [132].Parallel training (2 workers) using BlockMomentumSGD with block momentum = 0.5000, block momentum time constant (per worker) = 2954.6394, block learning rate = 1.0000, block size per worker = 2048 samples, using Nesterov-style block momentum, resetting SGD momentum after sync.
MPI Rank 1: 
MPI Rank 1: 12/21/2016 05:28:15: Starting minibatch loop, distributed reading is ENABLED.
MPI Rank 1: 12/21/2016 05:28:15:   BaseAdaptiveLearnRateSearch: Epoch[ 5 of 5]-Minibatch[   1-   1, 5.00%]: CE = 1.72259049 * 512; Err = 0.47070313 * 512; time = 0.0701s; samplesPerSecond = 7306.2
MPI Rank 1: 12/21/2016 05:28:15:   BaseAdaptiveLearnRateSearch: Epoch[ 5 of 5]-Minibatch[   2-   2, 10.00%]: CE = 0.00000000 * 0; Err = 0.00000000 * 0; time = 8.8000e-005s; samplesPerSecond = 0.0
MPI Rank 1: 		(model aggregation stats): 1-th sync point was hit, introducing a 0.06-seconds latency this time; accumulated time on sync point = 0.06 seconds , average latency = 0.06 seconds
MPI Rank 1: 		(model aggregation stats) 1-th sync:     0.24 seconds since last report (0.07 seconds on comm.); 1536 samples processed by 2 workers (512 by me);
MPI Rank 1: 		(model aggregation stats) 1-th sync: totalThroughput = 6.40k samplesPerSecond , throughputPerWorker = 3.20k samplesPerSecond
MPI Rank 1: 12/21/2016 05:28:15:  Finished Mini-Epoch[5]: CE = 1.84766700 * 1536; Err = 0.51627604 * 1536; learningRatePerSample = 0; minibatchSize = 1024
MPI Rank 1: ValidateSubNetwork: featNorm.invStdDev InvStdDev operation changed, from [363 x 1] to [363].ValidateSubNetwork: featNorm.mean Mean operation changed, from [363 x 1] to [363].ValidateSubNetwork: logPrior._ Mean operation changed, from [132 x 1] to [132].Parallel training (2 workers) using BlockMomentumSGD with block momentum = 0.5000, block momentum time constant (per worker) = 2954.6394, block learning rate = 1.0000, block size per worker = 2048 samples, using Nesterov-style block momentum, resetting SGD momentum after sync.
MPI Rank 1: 
MPI Rank 1: 12/21/2016 05:28:15: Starting minibatch loop, distributed reading is ENABLED.
MPI Rank 1: 12/21/2016 05:28:16:   AdaptiveLearnRateSearch: Epoch[ 5 of 5]-Minibatch[   1-   1, 5.00%]: CE = 1.72259049 * 512; Err = 0.47070313 * 512; time = 0.0976s; samplesPerSecond = 5245.4
MPI Rank 1: 12/21/2016 05:28:16:   AdaptiveLearnRateSearch: Epoch[ 5 of 5]-Minibatch[   2-   2, 10.00%]: CE = 0.00000000 * 0; Err = 0.00000000 * 0; time = 0.0001s; samplesPerSecond = 0.0
MPI Rank 1: 		(model aggregation stats): 1-th sync point was hit, introducing a 0.14-seconds latency this time; accumulated time on sync point = 0.14 seconds , average latency = 0.14 seconds
MPI Rank 1: 		(model aggregation stats) 1-th sync:     0.37 seconds since last report (0.09 seconds on comm.); 1536 samples processed by 2 workers (512 by me);
MPI Rank 1: 		(model aggregation stats) 1-th sync: totalThroughput = 4.13k samplesPerSecond , throughputPerWorker = 2.07k samplesPerSecond
MPI Rank 1: 12/21/2016 05:28:16:  Finished Mini-Epoch[5]: CE = 1.82753521 * 1536; Err = 0.49674479 * 1536; learningRatePerSample = 0.0063207938; minibatchSize = 1024
MPI Rank 1: ValidateSubNetwork: featNorm.invStdDev InvStdDev operation changed, from [363 x 1] to [363].ValidateSubNetwork: featNorm.mean Mean operation changed, from [363 x 1] to [363].ValidateSubNetwork: logPrior._ Mean operation changed, from [132 x 1] to [132].12/21/2016 05:28:16:  SearchForBestLearnRate Epoch[5]: Best learningRatePerSample = 0.006320793777, baseCriterion=1.85275755
MPI Rank 1: 
MPI Rank 1: 12/21/2016 05:28:16: Starting Epoch 5: learning rate per sample = 0.006321  effective momentum = 0.656119  momentum as time constant = 2429.9 samples
MPI Rank 1: Parallel training (2 workers) using BlockMomentumSGD with block momentum = 0.5000, block momentum time constant (per worker) = 2954.6394, block learning rate = 1.0000, block size per worker = 2048 samples, using Nesterov-style block momentum, resetting SGD momentum after sync.
MPI Rank 1: 
MPI Rank 1: 12/21/2016 05:28:16: Starting minibatch loop, distributed reading is ENABLED.
MPI Rank 1: 12/21/2016 05:28:16:  Epoch[ 5 of 5]-Minibatch[   1-   1, 5.00%]: CE = 1.72259049 * 512; Err = 0.47070313 * 512; time = 0.1615s; samplesPerSecond = 3170.2
MPI Rank 1: 12/21/2016 05:28:17:  Epoch[ 5 of 5]-Minibatch[   2-   2, 10.00%]: CE = 1.81334940 * 512; Err = 0.51562500 * 512; time = 0.3473s; samplesPerSecond = 1474.1
MPI Rank 1: 12/21/2016 05:28:17:  Epoch[ 5 of 5]-Minibatch[   3-   3, 15.00%]: CE = 1.75469950 * 512; Err = 0.50195313 * 512; time = 0.1326s; samplesPerSecond = 3861.8
MPI Rank 1: 		(model aggregation stats): 1-th sync point was hit, introducing a 0.00-seconds latency this time; accumulated time on sync point = 0.00 seconds , average latency = 0.00 seconds
MPI Rank 1: 		(model aggregation stats) 1-th sync:     0.89 seconds since last report (0.08 seconds on comm.); 4096 samples processed by 2 workers (2048 by me);
MPI Rank 1: 		(model aggregation stats) 1-th sync: totalThroughput = 4.61k samplesPerSecond , throughputPerWorker = 2.31k samplesPerSecond
MPI Rank 1: 12/21/2016 05:28:17:  Epoch[ 5 of 5]-Minibatch[   4-   4, 20.00%]: CE = 1.94839844 * 512; Err = 0.56445313 * 512; time = 0.2353s; samplesPerSecond = 2176.0
MPI Rank 1: 12/21/2016 05:28:17:  Epoch[ 5 of 5]-Minibatch[   5-   5, 25.00%]: CE = 1.81340242 * 512; Err = 0.55078125 * 512; time = 0.1856s; samplesPerSecond = 2758.5
MPI Rank 1: 12/21/2016 05:28:17:  Epoch[ 5 of 5]-Minibatch[   6-   6, 30.00%]: CE = 1.76005194 * 512; Err = 0.51171875 * 512; time = 0.1261s; samplesPerSecond = 4059.1
MPI Rank 1: 12/21/2016 05:28:18:  Epoch[ 5 of 5]-Minibatch[   7-   7, 35.00%]: CE = 1.90019478 * 512; Err = 0.52539063 * 512; time = 0.1615s; samplesPerSecond = 3170.9
MPI Rank 1: 		(model aggregation stats): 2-th sync point was hit, introducing a 0.20-seconds latency this time; accumulated time on sync point = 0.20 seconds , average latency = 0.10 seconds
MPI Rank 1: 		(model aggregation stats) 2-th sync:     0.91 seconds since last report (0.09 seconds on comm.); 4096 samples processed by 2 workers (2048 by me);
MPI Rank 1: 		(model aggregation stats) 2-th sync: totalThroughput = 4.50k samplesPerSecond , throughputPerWorker = 2.25k samplesPerSecond
MPI Rank 1: 12/21/2016 05:28:18:  Epoch[ 5 of 5]-Minibatch[   8-   8, 40.00%]: CE = 1.84919225 * 512; Err = 0.52148438 * 512; time = 0.4345s; samplesPerSecond = 1178.3
MPI Rank 1: 12/21/2016 05:28:18:  Epoch[ 5 of 5]-Minibatch[   9-   9, 45.00%]: CE = 1.87396402 * 512; Err = 0.50390625 * 512; time = 0.1820s; samplesPerSecond = 2812.4
MPI Rank 1: 12/21/2016 05:28:18:  Epoch[ 5 of 5]-Minibatch[  10-  10, 50.00%]: CE = 1.76656360 * 512; Err = 0.50781250 * 512; time = 0.2841s; samplesPerSecond = 1802.1
MPI Rank 1: 12/21/2016 05:28:19:  Epoch[ 5 of 5]-Minibatch[  11-  11, 55.00%]: CE = 1.73350646 * 512; Err = 0.47265625 * 512; time = 0.1222s; samplesPerSecond = 4188.3
MPI Rank 1: 		(model aggregation stats): 3-th sync point was hit, introducing a 0.00-seconds latency this time; accumulated time on sync point = 0.20 seconds , average latency = 0.07 seconds
MPI Rank 1: 		(model aggregation stats) 3-th sync:     0.70 seconds since last report (0.00 seconds on comm.); 4096 samples processed by 2 workers (2048 by me);
MPI Rank 1: 		(model aggregation stats) 3-th sync: totalThroughput = 5.82k samplesPerSecond , throughputPerWorker = 2.91k samplesPerSecond
MPI Rank 1: 12/21/2016 05:28:19:  Epoch[ 5 of 5]-Minibatch[  12-  12, 60.00%]: CE = 1.82274869 * 512; Err = 0.54882813 * 512; time = 0.1135s; samplesPerSecond = 4512.3
MPI Rank 1: 12/21/2016 05:28:19:  Epoch[ 5 of 5]-Minibatch[  13-  13, 65.00%]: CE = 2.04859191 * 512; Err = 0.55078125 * 512; time = 0.1202s; samplesPerSecond = 4260.0
MPI Rank 1: 12/21/2016 05:28:19:  Epoch[ 5 of 5]-Minibatch[  14-  14, 70.00%]: CE = 1.86725595 * 512; Err = 0.55859375 * 512; time = 0.1173s; samplesPerSecond = 4366.0
MPI Rank 1: 12/21/2016 05:28:19:  Epoch[ 5 of 5]-Minibatch[  15-  15, 75.00%]: CE = 1.89531398 * 512; Err = 0.54492188 * 512; time = 0.1493s; samplesPerSecond = 3429.1
MPI Rank 1: 12/21/2016 05:28:19:  Epoch[ 5 of 5]-Minibatch[  16-  16, 80.00%]: CE = 1.65643464 * 333; Err = 0.48348348 * 333; time = 0.1392s; samplesPerSecond = 2391.6
MPI Rank 1: 12/21/2016 05:28:19:  Epoch[ 5 of 5]-Minibatch[  17-  17, 85.00%]: CE = 0.00000000 * 0; Err = 0.00000000 * 0; time = 8.8000e-005s; samplesPerSecond = 0.0
MPI Rank 1: 		(model aggregation stats): 4-th sync point was hit, introducing a 0.10-seconds latency this time; accumulated time on sync point = 0.29 seconds , average latency = 0.07 seconds
MPI Rank 1: 		(model aggregation stats) 4-th sync:     1.36 seconds since last report (0.71 seconds on comm.); 8192 samples processed by 2 workers (1869 by me);
MPI Rank 1: 		(model aggregation stats) 4-th sync: totalThroughput = 6.01k samplesPerSecond , throughputPerWorker = 3.00k samplesPerSecond
MPI Rank 1: 12/21/2016 05:28:20: Finished Epoch[ 5 of 5]: [Training] CE = 1.78255566 * 20480; Err = 0.51186523 * 20480; totalSamplesSeen = 102400; learningRatePerSample = 0.0063207938; epochTime=5.48441s
MPI Rank 1: 
MPI Rank 1: 12/21/2016 05:28:20: Action "train" complete.
MPI Rank 1: 
MPI Rank 1: 12/21/2016 05:28:20: __COMPLETED__
MPI Rank 1: ~MPIWrapper
