CPU info:
    CPU Model Name: Intel(R) Xeon(R) CPU E5-2690 v3 @ 2.60GHz
    Hardware threads: 12
    Total Memory: 57691188 kB
-------------------------------------------------------------------
=== Running /home/ubuntu/workspace/build/gpu/release/bin/cntk configFile=/home/ubuntu/workspace/Tests/EndToEndTests/Speech/Simple/cntk.cntk currentDirectory=/home/ubuntu/workspace/Tests/EndToEndTests/Speech/Data RunDir=/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu DataDir=/home/ubuntu/workspace/Tests/EndToEndTests/Speech/Data ConfigDir=/home/ubuntu/workspace/Tests/EndToEndTests/Speech/Simple OutputDir=/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu DeviceId=0 timestamping=true forceDeterministicAlgorithms=true makeMode=true
CNTK 2.3.1+ (HEAD b7b3e4, Jan 17 2018 02:42:45) at 2018/01/17 06:13:18

/home/ubuntu/workspace/build/gpu/release/bin/cntk  configFile=/home/ubuntu/workspace/Tests/EndToEndTests/Speech/Simple/cntk.cntk  currentDirectory=/home/ubuntu/workspace/Tests/EndToEndTests/Speech/Data  RunDir=/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu  DataDir=/home/ubuntu/workspace/Tests/EndToEndTests/Speech/Data  ConfigDir=/home/ubuntu/workspace/Tests/EndToEndTests/Speech/Simple  OutputDir=/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu  DeviceId=0  timestamping=true  forceDeterministicAlgorithms=true  makeMode=true
Changed current directory to /home/ubuntu/workspace/Tests/EndToEndTests/Speech/Data
01/17/2018 06:13:18: -------------------------------------------------------------------
01/17/2018 06:13:18: Build info: 

01/17/2018 06:13:18: 		Built time: Jan 17 2018 02:36:21
01/17/2018 06:13:18: 		Last modified date: Wed Jan 17 02:34:37 2018
01/17/2018 06:13:18: 		Build type: release
01/17/2018 06:13:18: 		Build target: GPU
01/17/2018 06:13:18: 		With ASGD: yes
01/17/2018 06:13:18: 		Math lib: mkl
01/17/2018 06:13:18: 		CUDA version: 9.0.0
01/17/2018 06:13:18: 		CUDNN version: 7.0.4
01/17/2018 06:13:18: 		Build Branch: HEAD
01/17/2018 06:13:18: 		Build SHA1: b7b3e4fb3ff0f69024ce19a19b8f2780fb63078b
01/17/2018 06:13:18: 		MPI distribution: Open MPI
01/17/2018 06:13:18: 		MPI version: 1.10.7
01/17/2018 06:13:18: -------------------------------------------------------------------
01/17/2018 06:13:18: -------------------------------------------------------------------
01/17/2018 06:13:18: GPU info:

01/17/2018 06:13:18: 		Device[0]: cores = 3072; computeCapability = 5.2; type = "Tesla M60"; total memory = 8123 MB; free memory = 8112 MB
01/17/2018 06:13:18: -------------------------------------------------------------------

Configuration After Processing and Variable Resolution:

configparameters: cntk.cntk:command=Simple_Demo:Simple_Demo_Output
configparameters: cntk.cntk:ConfigDir=/home/ubuntu/workspace/Tests/EndToEndTests/Speech/Simple
configparameters: cntk.cntk:currentDirectory=/home/ubuntu/workspace/Tests/EndToEndTests/Speech/Data
configparameters: cntk.cntk:DataDir=/home/ubuntu/workspace/Tests/EndToEndTests/Speech/Data
configparameters: cntk.cntk:deviceId=0
configparameters: cntk.cntk:DeviceNumber=-1
configparameters: cntk.cntk:forceDeterministicAlgorithms=true
configparameters: cntk.cntk:makeMode=true
configparameters: cntk.cntk:modelPath=/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn
configparameters: cntk.cntk:OutputDir=/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu
configparameters: cntk.cntk:outputNodeNames=ScaledLogLikelihood
configparameters: cntk.cntk:precision=float
configparameters: cntk.cntk:RunDir=/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu
configparameters: cntk.cntk:Simple_Demo=[
    action=train
    SimpleNetworkBuilder=[
        layerSizes=2:50*2:2
        trainingCriterion=CrossEntropyWithSoftmax
        evalCriterion=ClassificationError
        layerTypes=Sigmoid
        initValueScale=1.0
        applyMeanVarNorm=true
        uniformInit=true
        needPrior=true
    ]
    SGD=[
        epochSize=0 
        minibatchSize=128
        learningRatesPerSample=0.1
        momentumAsTimeConstant=2500
        dropoutRate=0.0
        maxEpochs=50
        keepCheckPointFiles = true
    ]
    reader=[
        readerType=CNTKTextFormatReader
        file=/home/ubuntu/workspace/Tests/EndToEndTests/Speech/Data/SimpleDataTrain_cntk_text.txt
        input = [
            features=[
dim = 2      
                format = "dense"
            ]
            labels=[
dim = 2 
                format = "dense"
            ]
        ]
    ]
]

configparameters: cntk.cntk:Simple_Demo_Output=[
    action=write
    reader=[
        readerType=CNTKTextFormatReader
        file=/home/ubuntu/workspace/Tests/EndToEndTests/Speech/Data/SimpleDataTest_cntk_text.txt
        input = [
            features=[
dim = 2 
                format = "dense" 
            ]
            labels=[
dim = 2 
                format = "dense"
            ]
        ]
    ]
outputPath=/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/SimpleOutput    
]

configparameters: cntk.cntk:timestamping=true
configparameters: cntk.cntk:traceLevel=1
01/17/2018 06:13:18: Commands: Simple_Demo Simple_Demo_Output
01/17/2018 06:13:18: precision = "float"
01/17/2018 06:13:18: WARNING: forceDeterministicAlgorithms flag is specified. Using 1 CPU thread for processing.

01/17/2018 06:13:18: ##############################################################################
01/17/2018 06:13:18: #                                                                            #
01/17/2018 06:13:18: # Simple_Demo command (train action)                                         #
01/17/2018 06:13:18: #                                                                            #
01/17/2018 06:13:18: ##############################################################################

01/17/2018 06:13:18: 
Creating virgin network.
SimpleNetworkBuilder Using GPU 0
SetUniformRandomValue (GPU): creating curand object with seed 1, sizeof(ElemType)==4
01/17/2018 06:13:18: 
Model has 25 nodes. Using GPU 0.

01/17/2018 06:13:18: Training criterion:   CrossEntropyWithSoftmax = CrossEntropyWithSoftmax
01/17/2018 06:13:18: Evaluation criterion: EvalClassificationError = ClassificationError


Allocating matrices for forward and/or backward propagation.

Gradient Memory Aliasing: 4 are aliased.
	W2*H1 (gradient) reuses HLast (gradient)
	W1*H1 (gradient) reuses W1*H1+B1 (gradient)

Memory Sharing: Out of 40 matrices, 21 are shared as 5, and 19 are not shared.

Here are the ones that share memory:
	{ PosteriorProb : [2 x 1 x *]
	  ScaledLogLikelihood : [2 x 1 x *] }
	{ H2 : [50 x 1 x *]
	  W0*features+B0 : [50 x 1 x *]
	  W1 : [50 x 50] (gradient)
	  W1*H1 : [50 x 1 x *] }
	{ B0 : [50 x 1] (gradient)
	  H1 : [50 x 1 x *] }
	{ HLast : [2 x 1 x *] (gradient)
	  W0 : [50 x 2] (gradient)
	  W0*features+B0 : [50 x 1 x *] (gradient)
	  W1*H1 : [50 x 1 x *] (gradient)
	  W1*H1+B1 : [50 x 1 x *]
	  W1*H1+B1 : [50 x 1 x *] (gradient)
	  W2*H1 : [2 x 1 x *]
	  W2*H1 : [2 x 1 x *] (gradient) }
	{ H1 : [50 x 1 x *] (gradient)
	  H2 : [50 x 1 x *] (gradient)
	  HLast : [2 x 1 x *]
	  W0*features : [50 x *]
	  W0*features : [50 x *] (gradient) }

Here are the ones that don't share memory:
	{B1 : [50 x 1]}
	{W2 : [2 x 50]}
	{B2 : [2 x 1]}
	{labels : [2 x *]}
	{Prior : [2]}
	{CrossEntropyWithSoftmax : [1]}
	{EvalClassificationError : [1]}
	{W2 : [2 x 50] (gradient)}
	{LogOfPrior : [2]}
	{MVNormalizedFeatures : [2 x *]}
	{B2 : [2 x 1] (gradient)}
	{B1 : [50 x 1] (gradient)}
	{CrossEntropyWithSoftmax : [1] (gradient)}
	{B0 : [50 x 1]}
	{W1 : [50 x 50]}
	{MeanOfFeatures : [2]}
	{InvStdOfFeatures : [2]}
	{W0 : [50 x 2]}
	{features : [2 x *]}


01/17/2018 06:13:18: Training 2802 parameters in 6 out of 6 parameter tensors and 15 nodes with gradient:

01/17/2018 06:13:18: 	Node 'B0' (LearnableParameter operation) : [50 x 1]
01/17/2018 06:13:18: 	Node 'B1' (LearnableParameter operation) : [50 x 1]
01/17/2018 06:13:18: 	Node 'B2' (LearnableParameter operation) : [2 x 1]
01/17/2018 06:13:18: 	Node 'W0' (LearnableParameter operation) : [50 x 2]
01/17/2018 06:13:18: 	Node 'W1' (LearnableParameter operation) : [50 x 50]
01/17/2018 06:13:18: 	Node 'W2' (LearnableParameter operation) : [2 x 50]


01/17/2018 06:13:18: Precomputing --> 3 PreCompute nodes found.

01/17/2018 06:13:18: 	MeanOfFeatures = Mean()
01/17/2018 06:13:18: 	InvStdOfFeatures = InvStdDev()
01/17/2018 06:13:18: 	Prior = Mean()

01/17/2018 06:13:18: Precomputing --> Completed.


01/17/2018 06:13:18: Starting Epoch 1: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:18: Starting minibatch loop.
01/17/2018 06:13:18:  Epoch[ 1 of 50]-Minibatch[   1-  10]: CrossEntropyWithSoftmax = 0.84310627 * 1280; EvalClassificationError = 0.51718750 * 1280; time = 0.0065s; samplesPerSecond = 197041.3
01/17/2018 06:13:18:  Epoch[ 1 of 50]-Minibatch[  11-  20]: CrossEntropyWithSoftmax = 0.79649162 * 1280; EvalClassificationError = 0.51562500 * 1280; time = 0.0053s; samplesPerSecond = 241327.3
01/17/2018 06:13:18:  Epoch[ 1 of 50]-Minibatch[  21-  30]: CrossEntropyWithSoftmax = 0.72212505 * 1280; EvalClassificationError = 0.48515625 * 1280; time = 0.0053s; samplesPerSecond = 239642.0
01/17/2018 06:13:18:  Epoch[ 1 of 50]-Minibatch[  31-  40]: CrossEntropyWithSoftmax = 0.71446991 * 1280; EvalClassificationError = 0.49765625 * 1280; time = 0.0057s; samplesPerSecond = 224105.3
01/17/2018 06:13:18:  Epoch[ 1 of 50]-Minibatch[  41-  50]: CrossEntropyWithSoftmax = 0.70253906 * 1280; EvalClassificationError = 0.50078125 * 1280; time = 0.0054s; samplesPerSecond = 238774.8
01/17/2018 06:13:18:  Epoch[ 1 of 50]-Minibatch[  51-  60]: CrossEntropyWithSoftmax = 0.70667038 * 1280; EvalClassificationError = 0.51718750 * 1280; time = 0.0053s; samplesPerSecond = 241988.8
01/17/2018 06:13:18:  Epoch[ 1 of 50]-Minibatch[  61-  70]: CrossEntropyWithSoftmax = 0.70342674 * 1280; EvalClassificationError = 0.50156250 * 1280; time = 0.0058s; samplesPerSecond = 220617.4
01/17/2018 06:13:18: Finished Epoch[ 1 of 50]: [Training] CrossEntropyWithSoftmax = 0.73692187 * 10000; EvalClassificationError = 0.50500000 * 10000; totalSamplesSeen = 10000; learningRatePerSample = 0.1; epochTime=0.0448005s
01/17/2018 06:13:18: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.1'

01/17/2018 06:13:18: Starting Epoch 2: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:18: Starting minibatch loop.
01/17/2018 06:13:18:  Epoch[ 2 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.70579815 * 1280; EvalClassificationError = 0.49296875 * 1280; time = 0.0059s; samplesPerSecond = 216758.1
01/17/2018 06:13:18:  Epoch[ 2 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.74300938 * 1280; EvalClassificationError = 0.47968750 * 1280; time = 0.0053s; samplesPerSecond = 240069.0
01/17/2018 06:13:18:  Epoch[ 2 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.74903164 * 1280; EvalClassificationError = 0.51171875 * 1280; time = 0.0054s; samplesPerSecond = 238467.8
01/17/2018 06:13:18:  Epoch[ 2 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.74969616 * 1280; EvalClassificationError = 0.51171875 * 1280; time = 0.0058s; samplesPerSecond = 221131.9
01/17/2018 06:13:18:  Epoch[ 2 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.74229851 * 1280; EvalClassificationError = 0.50234375 * 1280; time = 0.0054s; samplesPerSecond = 238748.1
01/17/2018 06:13:18:  Epoch[ 2 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.72703133 * 1280; EvalClassificationError = 0.50546875 * 1280; time = 0.0054s; samplesPerSecond = 237745.9
01/17/2018 06:13:18:  Epoch[ 2 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.73382492 * 1280; EvalClassificationError = 0.50078125 * 1280; time = 0.0056s; samplesPerSecond = 228229.1
01/17/2018 06:13:18: Finished Epoch[ 2 of 50]: [Training] CrossEntropyWithSoftmax = 0.73301997 * 10000; EvalClassificationError = 0.49860000 * 10000; totalSamplesSeen = 20000; learningRatePerSample = 0.1; epochTime=0.0451511s
01/17/2018 06:13:18: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.2'

01/17/2018 06:13:18: Starting Epoch 3: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:18: Starting minibatch loop.
01/17/2018 06:13:18:  Epoch[ 3 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.73711982 * 1280; EvalClassificationError = 0.50000000 * 1280; time = 0.0067s; samplesPerSecond = 191238.9
01/17/2018 06:13:18:  Epoch[ 3 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.73555017 * 1280; EvalClassificationError = 0.50625000 * 1280; time = 0.0056s; samplesPerSecond = 227924.3
01/17/2018 06:13:18:  Epoch[ 3 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.76823006 * 1280; EvalClassificationError = 0.52187500 * 1280; time = 0.0058s; samplesPerSecond = 222442.3
01/17/2018 06:13:18:  Epoch[ 3 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.75786629 * 1280; EvalClassificationError = 0.50234375 * 1280; time = 0.0055s; samplesPerSecond = 232968.7
01/17/2018 06:13:18:  Epoch[ 3 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.73280983 * 1280; EvalClassificationError = 0.50937500 * 1280; time = 0.0055s; samplesPerSecond = 232355.5
01/17/2018 06:13:18:  Epoch[ 3 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.71082916 * 1280; EvalClassificationError = 0.48984375 * 1280; time = 0.0056s; samplesPerSecond = 226898.1
01/17/2018 06:13:18:  Epoch[ 3 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.69264183 * 1280; EvalClassificationError = 0.50703125 * 1280; time = 0.0058s; samplesPerSecond = 222515.8
01/17/2018 06:13:18: Finished Epoch[ 3 of 50]: [Training] CrossEntropyWithSoftmax = 0.72925596 * 10000; EvalClassificationError = 0.49610000 * 10000; totalSamplesSeen = 30000; learningRatePerSample = 0.1; epochTime=0.0468858s
01/17/2018 06:13:18: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.3'

01/17/2018 06:13:18: Starting Epoch 4: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:18: Starting minibatch loop.
01/17/2018 06:13:18:  Epoch[ 4 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.69857616 * 1280; EvalClassificationError = 0.49296875 * 1280; time = 0.0060s; samplesPerSecond = 213653.8
01/17/2018 06:13:18:  Epoch[ 4 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.68717909 * 1280; EvalClassificationError = 0.47031250 * 1280; time = 0.0056s; samplesPerSecond = 230522.6
01/17/2018 06:13:18:  Epoch[ 4 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.44480124 * 1280; EvalClassificationError = 0.18593750 * 1280; time = 0.0054s; samplesPerSecond = 238144.0
01/17/2018 06:13:18:  Epoch[ 4 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.20336590 * 1280; EvalClassificationError = 0.08046875 * 1280; time = 0.0057s; samplesPerSecond = 225634.2
01/17/2018 06:13:19:  Epoch[ 4 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.16974277 * 1280; EvalClassificationError = 0.06953125 * 1280; time = 0.0055s; samplesPerSecond = 233781.4
01/17/2018 06:13:19:  Epoch[ 4 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.16295624 * 1280; EvalClassificationError = 0.07421875 * 1280; time = 0.0054s; samplesPerSecond = 236712.7
01/17/2018 06:13:19:  Epoch[ 4 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.16088600 * 1280; EvalClassificationError = 0.06718750 * 1280; time = 0.0058s; samplesPerSecond = 221771.7
01/17/2018 06:13:19: Finished Epoch[ 4 of 50]: [Training] CrossEntropyWithSoftmax = 0.34535120 * 10000; EvalClassificationError = 0.19360000 * 10000; totalSamplesSeen = 40000; learningRatePerSample = 0.1; epochTime=0.0462825s
01/17/2018 06:13:19: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.4'

01/17/2018 06:13:19: Starting Epoch 5: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:19: Starting minibatch loop.
01/17/2018 06:13:19:  Epoch[ 5 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.18401880 * 1280; EvalClassificationError = 0.07734375 * 1280; time = 0.0273s; samplesPerSecond = 46968.2
01/17/2018 06:13:19:  Epoch[ 5 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.17633183 * 1280; EvalClassificationError = 0.08125000 * 1280; time = 0.0087s; samplesPerSecond = 147101.1
01/17/2018 06:13:19:  Epoch[ 5 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.15577967 * 1280; EvalClassificationError = 0.06484375 * 1280; time = 0.0060s; samplesPerSecond = 211654.2
01/17/2018 06:13:19:  Epoch[ 5 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.15854549 * 1280; EvalClassificationError = 0.06953125 * 1280; time = 0.0053s; samplesPerSecond = 243735.2
01/17/2018 06:13:19:  Epoch[ 5 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.18194976 * 1280; EvalClassificationError = 0.08203125 * 1280; time = 0.0063s; samplesPerSecond = 202987.7
01/17/2018 06:13:19:  Epoch[ 5 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.16833982 * 1280; EvalClassificationError = 0.07812500 * 1280; time = 0.0052s; samplesPerSecond = 246566.4
01/17/2018 06:13:19:  Epoch[ 5 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.17359457 * 1280; EvalClassificationError = 0.07734375 * 1280; time = 0.0055s; samplesPerSecond = 231984.9
01/17/2018 06:13:19: Finished Epoch[ 5 of 50]: [Training] CrossEntropyWithSoftmax = 0.16894944 * 10000; EvalClassificationError = 0.07490000 * 10000; totalSamplesSeen = 50000; learningRatePerSample = 0.1; epochTime=0.0711467s
01/17/2018 06:13:19: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.5'

01/17/2018 06:13:19: Starting Epoch 6: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:19: Starting minibatch loop.
01/17/2018 06:13:19:  Epoch[ 6 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.19006202 * 1280; EvalClassificationError = 0.08437500 * 1280; time = 0.0055s; samplesPerSecond = 231582.0
01/17/2018 06:13:19:  Epoch[ 6 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.19627984 * 1280; EvalClassificationError = 0.08437500 * 1280; time = 0.0057s; samplesPerSecond = 226288.3
01/17/2018 06:13:19:  Epoch[ 6 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.16198635 * 1280; EvalClassificationError = 0.07656250 * 1280; time = 0.0052s; samplesPerSecond = 244204.9
01/17/2018 06:13:19:  Epoch[ 6 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.16606755 * 1280; EvalClassificationError = 0.07421875 * 1280; time = 0.0052s; samplesPerSecond = 248346.0
01/17/2018 06:13:19:  Epoch[ 6 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.18116536 * 1280; EvalClassificationError = 0.08828125 * 1280; time = 0.0056s; samplesPerSecond = 227612.2
01/17/2018 06:13:19:  Epoch[ 6 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.16550684 * 1280; EvalClassificationError = 0.06796875 * 1280; time = 0.0053s; samplesPerSecond = 242314.1
01/17/2018 06:13:19:  Epoch[ 6 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.16042614 * 1280; EvalClassificationError = 0.07500000 * 1280; time = 0.0054s; samplesPerSecond = 237094.1
01/17/2018 06:13:19: Finished Epoch[ 6 of 50]: [Training] CrossEntropyWithSoftmax = 0.17043514 * 10000; EvalClassificationError = 0.07720000 * 10000; totalSamplesSeen = 60000; learningRatePerSample = 0.1; epochTime=0.0443891s
01/17/2018 06:13:19: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.6'

01/17/2018 06:13:19: Starting Epoch 7: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:19: Starting minibatch loop.
01/17/2018 06:13:19:  Epoch[ 7 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.15693432 * 1280; EvalClassificationError = 0.07187500 * 1280; time = 0.0055s; samplesPerSecond = 232985.7
01/17/2018 06:13:19:  Epoch[ 7 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.16458679 * 1280; EvalClassificationError = 0.08359375 * 1280; time = 0.0056s; samplesPerSecond = 230099.9
01/17/2018 06:13:19:  Epoch[ 7 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.15007732 * 1280; EvalClassificationError = 0.07578125 * 1280; time = 0.0052s; samplesPerSecond = 244873.0
01/17/2018 06:13:19:  Epoch[ 7 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.13501539 * 1280; EvalClassificationError = 0.05937500 * 1280; time = 0.0053s; samplesPerSecond = 243027.2
01/17/2018 06:13:19:  Epoch[ 7 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.14193029 * 1280; EvalClassificationError = 0.06250000 * 1280; time = 0.0056s; samplesPerSecond = 226914.1
01/17/2018 06:13:19:  Epoch[ 7 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.17521372 * 1280; EvalClassificationError = 0.08281250 * 1280; time = 0.0054s; samplesPerSecond = 238223.7
01/17/2018 06:13:19:  Epoch[ 7 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.16884174 * 1280; EvalClassificationError = 0.08125000 * 1280; time = 0.0052s; samplesPerSecond = 247553.5
01/17/2018 06:13:19: Finished Epoch[ 7 of 50]: [Training] CrossEntropyWithSoftmax = 0.16109182 * 10000; EvalClassificationError = 0.07630000 * 10000; totalSamplesSeen = 70000; learningRatePerSample = 0.1; epochTime=0.0494294s
01/17/2018 06:13:19: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.7'

01/17/2018 06:13:19: Starting Epoch 8: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:19: Starting minibatch loop.
01/17/2018 06:13:19:  Epoch[ 8 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.19287047 * 1280; EvalClassificationError = 0.09375000 * 1280; time = 0.0082s; samplesPerSecond = 155704.5
01/17/2018 06:13:19:  Epoch[ 8 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.18293436 * 1280; EvalClassificationError = 0.08671875 * 1280; time = 0.0086s; samplesPerSecond = 148909.9
01/17/2018 06:13:19:  Epoch[ 8 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.18483684 * 1280; EvalClassificationError = 0.07109375 * 1280; time = 0.0063s; samplesPerSecond = 201816.3
01/17/2018 06:13:19:  Epoch[ 8 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.17957010 * 1280; EvalClassificationError = 0.07734375 * 1280; time = 0.0053s; samplesPerSecond = 243096.4
01/17/2018 06:13:19:  Epoch[ 8 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.17522516 * 1280; EvalClassificationError = 0.08046875 * 1280; time = 0.0055s; samplesPerSecond = 231888.3
01/17/2018 06:13:19:  Epoch[ 8 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.20335159 * 1280; EvalClassificationError = 0.08671875 * 1280; time = 0.0055s; samplesPerSecond = 233875.4
01/17/2018 06:13:19:  Epoch[ 8 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.15407667 * 1280; EvalClassificationError = 0.06718750 * 1280; time = 0.0051s; samplesPerSecond = 252874.5
01/17/2018 06:13:19: Finished Epoch[ 8 of 50]: [Training] CrossEntropyWithSoftmax = 0.18057247 * 10000; EvalClassificationError = 0.08020000 * 10000; totalSamplesSeen = 80000; learningRatePerSample = 0.1; epochTime=0.0512552s
01/17/2018 06:13:19: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.8'

01/17/2018 06:13:19: Starting Epoch 9: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:19: Starting minibatch loop.
01/17/2018 06:13:19:  Epoch[ 9 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.17599977 * 1280; EvalClassificationError = 0.08515625 * 1280; time = 0.0065s; samplesPerSecond = 198378.9
01/17/2018 06:13:19:  Epoch[ 9 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.17346860 * 1280; EvalClassificationError = 0.07578125 * 1280; time = 0.0053s; samplesPerSecond = 239871.1
01/17/2018 06:13:19:  Epoch[ 9 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.17637930 * 1280; EvalClassificationError = 0.07968750 * 1280; time = 0.0055s; samplesPerSecond = 232507.4
01/17/2018 06:13:19:  Epoch[ 9 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.17621803 * 1280; EvalClassificationError = 0.07500000 * 1280; time = 0.0053s; samplesPerSecond = 240873.2
01/17/2018 06:13:19:  Epoch[ 9 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.16279397 * 1280; EvalClassificationError = 0.07578125 * 1280; time = 0.0052s; samplesPerSecond = 244382.1
01/17/2018 06:13:19:  Epoch[ 9 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.14524059 * 1280; EvalClassificationError = 0.06328125 * 1280; time = 0.0056s; samplesPerSecond = 227507.0
01/17/2018 06:13:19:  Epoch[ 9 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.14306850 * 1280; EvalClassificationError = 0.06484375 * 1280; time = 0.0053s; samplesPerSecond = 243447.8
01/17/2018 06:13:19: Finished Epoch[ 9 of 50]: [Training] CrossEntropyWithSoftmax = 0.16553330 * 10000; EvalClassificationError = 0.07510000 * 10000; totalSamplesSeen = 90000; learningRatePerSample = 0.1; epochTime=0.0449216s
01/17/2018 06:13:19: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.9'

01/17/2018 06:13:19: Starting Epoch 10: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:19: Starting minibatch loop.
01/17/2018 06:13:19:  Epoch[10 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.16552613 * 1280; EvalClassificationError = 0.07890625 * 1280; time = 0.0056s; samplesPerSecond = 230269.7
01/17/2018 06:13:19:  Epoch[10 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.14848220 * 1280; EvalClassificationError = 0.07031250 * 1280; time = 0.0051s; samplesPerSecond = 249400.9
01/17/2018 06:13:19:  Epoch[10 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.16952171 * 1280; EvalClassificationError = 0.07812500 * 1280; time = 0.0052s; samplesPerSecond = 247567.8
01/17/2018 06:13:19:  Epoch[10 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.15449610 * 1280; EvalClassificationError = 0.07187500 * 1280; time = 0.0054s; samplesPerSecond = 236638.3
01/17/2018 06:13:19:  Epoch[10 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.21318636 * 1280; EvalClassificationError = 0.09453125 * 1280; time = 0.0054s; samplesPerSecond = 235017.6
01/17/2018 06:13:19:  Epoch[10 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.16077089 * 1280; EvalClassificationError = 0.06640625 * 1280; time = 0.0058s; samplesPerSecond = 222515.8
01/17/2018 06:13:19:  Epoch[10 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.17158928 * 1280; EvalClassificationError = 0.07890625 * 1280; time = 0.0072s; samplesPerSecond = 178481.8
01/17/2018 06:13:19: Finished Epoch[10 of 50]: [Training] CrossEntropyWithSoftmax = 0.16660631 * 10000; EvalClassificationError = 0.07680000 * 10000; totalSamplesSeen = 100000; learningRatePerSample = 0.1; epochTime=0.0460358s
01/17/2018 06:13:19: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.10'

01/17/2018 06:13:19: Starting Epoch 11: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:19: Starting minibatch loop.
01/17/2018 06:13:19:  Epoch[11 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.17469569 * 1280; EvalClassificationError = 0.07812500 * 1280; time = 0.0059s; samplesPerSecond = 218751.1
01/17/2018 06:13:19:  Epoch[11 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.16525284 * 1280; EvalClassificationError = 0.07343750 * 1280; time = 0.0052s; samplesPerSecond = 244386.7
01/17/2018 06:13:19:  Epoch[11 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.17623069 * 1280; EvalClassificationError = 0.08593750 * 1280; time = 0.0080s; samplesPerSecond = 159168.3
01/17/2018 06:13:19:  Epoch[11 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.16749387 * 1280; EvalClassificationError = 0.07812500 * 1280; time = 0.0068s; samplesPerSecond = 187699.8
01/17/2018 06:13:19:  Epoch[11 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.16025534 * 1280; EvalClassificationError = 0.07031250 * 1280; time = 0.0090s; samplesPerSecond = 142170.1
01/17/2018 06:13:19:  Epoch[11 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.15658598 * 1280; EvalClassificationError = 0.07109375 * 1280; time = 0.0069s; samplesPerSecond = 184981.8
01/17/2018 06:13:19:  Epoch[11 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.16311159 * 1280; EvalClassificationError = 0.07421875 * 1280; time = 0.0055s; samplesPerSecond = 234771.9
01/17/2018 06:13:19: Finished Epoch[11 of 50]: [Training] CrossEntropyWithSoftmax = 0.17151410 * 10000; EvalClassificationError = 0.07770000 * 10000; totalSamplesSeen = 110000; learningRatePerSample = 0.1; epochTime=0.0659085s
01/17/2018 06:13:19: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.11'

01/17/2018 06:13:19: Starting Epoch 12: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:19: Starting minibatch loop.
01/17/2018 06:13:19:  Epoch[12 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.16809069 * 1280; EvalClassificationError = 0.06562500 * 1280; time = 0.0062s; samplesPerSecond = 205708.4
01/17/2018 06:13:19:  Epoch[12 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.17254847 * 1280; EvalClassificationError = 0.08359375 * 1280; time = 0.0101s; samplesPerSecond = 127121.6
01/17/2018 06:13:19:  Epoch[12 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.17613344 * 1280; EvalClassificationError = 0.07890625 * 1280; time = 0.0065s; samplesPerSecond = 197949.4
01/17/2018 06:13:19:  Epoch[12 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.16155581 * 1280; EvalClassificationError = 0.07656250 * 1280; time = 0.0058s; samplesPerSecond = 220029.6
01/17/2018 06:13:19:  Epoch[12 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.15687809 * 1280; EvalClassificationError = 0.08125000 * 1280; time = 0.0053s; samplesPerSecond = 243359.9
01/17/2018 06:13:19:  Epoch[12 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.16276302 * 1280; EvalClassificationError = 0.07578125 * 1280; time = 0.0056s; samplesPerSecond = 228261.6
01/17/2018 06:13:19:  Epoch[12 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.14401045 * 1280; EvalClassificationError = 0.06718750 * 1280; time = 0.0055s; samplesPerSecond = 233866.8
01/17/2018 06:13:19: Finished Epoch[12 of 50]: [Training] CrossEntropyWithSoftmax = 0.16279980 * 10000; EvalClassificationError = 0.07560000 * 10000; totalSamplesSeen = 120000; learningRatePerSample = 0.1; epochTime=0.055577s
01/17/2018 06:13:19: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.12'

01/17/2018 06:13:19: Starting Epoch 13: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:19: Starting minibatch loop.
01/17/2018 06:13:19:  Epoch[13 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.14974458 * 1280; EvalClassificationError = 0.06953125 * 1280; time = 0.0056s; samplesPerSecond = 228347.2
01/17/2018 06:13:19:  Epoch[13 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.17683125 * 1280; EvalClassificationError = 0.08515625 * 1280; time = 0.0053s; samplesPerSecond = 242117.0
01/17/2018 06:13:19:  Epoch[13 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.17482150 * 1280; EvalClassificationError = 0.08984375 * 1280; time = 0.0070s; samplesPerSecond = 182959.1
01/17/2018 06:13:19:  Epoch[13 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.15381856 * 1280; EvalClassificationError = 0.06562500 * 1280; time = 0.0052s; samplesPerSecond = 248124.5
01/17/2018 06:13:19:  Epoch[13 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.14474854 * 1280; EvalClassificationError = 0.06796875 * 1280; time = 0.0053s; samplesPerSecond = 243244.3
01/17/2018 06:13:19:  Epoch[13 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.16736126 * 1280; EvalClassificationError = 0.08203125 * 1280; time = 0.0060s; samplesPerSecond = 215122.4
01/17/2018 06:13:19:  Epoch[13 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.15166674 * 1280; EvalClassificationError = 0.06796875 * 1280; time = 0.0052s; samplesPerSecond = 248220.8
01/17/2018 06:13:19: Finished Epoch[13 of 50]: [Training] CrossEntropyWithSoftmax = 0.16015929 * 10000; EvalClassificationError = 0.07590000 * 10000; totalSamplesSeen = 130000; learningRatePerSample = 0.1; epochTime=0.0462351s
01/17/2018 06:13:19: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.13'

01/17/2018 06:13:19: Starting Epoch 14: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:19: Starting minibatch loop.
01/17/2018 06:13:19:  Epoch[14 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.19459400 * 1280; EvalClassificationError = 0.08203125 * 1280; time = 0.0108s; samplesPerSecond = 118362.9
01/17/2018 06:13:19:  Epoch[14 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.20189953 * 1280; EvalClassificationError = 0.08125000 * 1280; time = 0.0075s; samplesPerSecond = 170949.3
01/17/2018 06:13:19:  Epoch[14 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.17680378 * 1280; EvalClassificationError = 0.08281250 * 1280; time = 0.0088s; samplesPerSecond = 144852.1
01/17/2018 06:13:19:  Epoch[14 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.20349355 * 1280; EvalClassificationError = 0.09218750 * 1280; time = 0.0073s; samplesPerSecond = 176478.7
01/17/2018 06:13:19:  Epoch[14 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.18833909 * 1280; EvalClassificationError = 0.07734375 * 1280; time = 0.0052s; samplesPerSecond = 245018.3
01/17/2018 06:13:19:  Epoch[14 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.16565113 * 1280; EvalClassificationError = 0.07187500 * 1280; time = 0.0052s; samplesPerSecond = 244508.1
01/17/2018 06:13:19:  Epoch[14 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.15764112 * 1280; EvalClassificationError = 0.08203125 * 1280; time = 0.0057s; samplesPerSecond = 225725.7
01/17/2018 06:13:19: Finished Epoch[14 of 50]: [Training] CrossEntropyWithSoftmax = 0.18298158 * 10000; EvalClassificationError = 0.08140000 * 10000; totalSamplesSeen = 140000; learningRatePerSample = 0.1; epochTime=0.057136s
01/17/2018 06:13:19: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.14'

01/17/2018 06:13:19: Starting Epoch 15: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:19: Starting minibatch loop.
01/17/2018 06:13:19:  Epoch[15 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.17048591 * 1280; EvalClassificationError = 0.08515625 * 1280; time = 0.0062s; samplesPerSecond = 207170.0
01/17/2018 06:13:19:  Epoch[15 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.14827458 * 1280; EvalClassificationError = 0.06484375 * 1280; time = 0.0059s; samplesPerSecond = 215364.9
01/17/2018 06:13:19:  Epoch[15 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.17491755 * 1280; EvalClassificationError = 0.07968750 * 1280; time = 0.0054s; samplesPerSecond = 237384.3
01/17/2018 06:13:19:  Epoch[15 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.17042289 * 1280; EvalClassificationError = 0.07812500 * 1280; time = 0.0054s; samplesPerSecond = 237710.6
01/17/2018 06:13:19:  Epoch[15 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.16541433 * 1280; EvalClassificationError = 0.08281250 * 1280; time = 0.0052s; samplesPerSecond = 244461.4
01/17/2018 06:13:19:  Epoch[15 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.16479816 * 1280; EvalClassificationError = 0.07656250 * 1280; time = 0.0051s; samplesPerSecond = 250039.1
01/17/2018 06:13:19:  Epoch[15 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.15720472 * 1280; EvalClassificationError = 0.07812500 * 1280; time = 0.0057s; samplesPerSecond = 225332.3
01/17/2018 06:13:19: Finished Epoch[15 of 50]: [Training] CrossEntropyWithSoftmax = 0.16389122 * 10000; EvalClassificationError = 0.07710000 * 10000; totalSamplesSeen = 150000; learningRatePerSample = 0.1; epochTime=0.0454236s
01/17/2018 06:13:19: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.15'

01/17/2018 06:13:19: Starting Epoch 16: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:19: Starting minibatch loop.
01/17/2018 06:13:19:  Epoch[16 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.17302433 * 1280; EvalClassificationError = 0.07265625 * 1280; time = 0.0060s; samplesPerSecond = 213910.9
01/17/2018 06:13:19:  Epoch[16 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.16493438 * 1280; EvalClassificationError = 0.07421875 * 1280; time = 0.0054s; samplesPerSecond = 238863.9
01/17/2018 06:13:19:  Epoch[16 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.17643499 * 1280; EvalClassificationError = 0.08281250 * 1280; time = 0.0054s; samplesPerSecond = 237807.7
01/17/2018 06:13:19:  Epoch[16 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.17659817 * 1280; EvalClassificationError = 0.08359375 * 1280; time = 0.0054s; samplesPerSecond = 235488.9
01/17/2018 06:13:19:  Epoch[16 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.18818760 * 1280; EvalClassificationError = 0.08828125 * 1280; time = 0.0052s; samplesPerSecond = 244093.1
01/17/2018 06:13:19:  Epoch[16 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.17494001 * 1280; EvalClassificationError = 0.07890625 * 1280; time = 0.0052s; samplesPerSecond = 244503.4
01/17/2018 06:13:19:  Epoch[16 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.17475004 * 1280; EvalClassificationError = 0.08359375 * 1280; time = 0.0055s; samplesPerSecond = 232896.7
01/17/2018 06:13:19: Finished Epoch[16 of 50]: [Training] CrossEntropyWithSoftmax = 0.17363783 * 10000; EvalClassificationError = 0.08020000 * 10000; totalSamplesSeen = 160000; learningRatePerSample = 0.1; epochTime=0.0445577s
01/17/2018 06:13:19: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.16'

01/17/2018 06:13:19: Starting Epoch 17: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:19: Starting minibatch loop.
01/17/2018 06:13:19:  Epoch[17 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.15547833 * 1280; EvalClassificationError = 0.07890625 * 1280; time = 0.0058s; samplesPerSecond = 220222.6
01/17/2018 06:13:19:  Epoch[17 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.16448460 * 1280; EvalClassificationError = 0.07578125 * 1280; time = 0.0055s; samplesPerSecond = 234711.7
01/17/2018 06:13:19:  Epoch[17 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.18492689 * 1280; EvalClassificationError = 0.09296875 * 1280; time = 0.0053s; samplesPerSecond = 242925.7
01/17/2018 06:13:19:  Epoch[17 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.15973330 * 1280; EvalClassificationError = 0.07265625 * 1280; time = 0.0074s; samplesPerSecond = 172098.5
01/17/2018 06:13:19:  Epoch[17 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.15462508 * 1280; EvalClassificationError = 0.08359375 * 1280; time = 0.0074s; samplesPerSecond = 172811.8
01/17/2018 06:13:19:  Epoch[17 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.15531902 * 1280; EvalClassificationError = 0.06875000 * 1280; time = 0.0068s; samplesPerSecond = 189511.7
01/17/2018 06:13:19:  Epoch[17 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.15558491 * 1280; EvalClassificationError = 0.07187500 * 1280; time = 0.0066s; samplesPerSecond = 194798.3
01/17/2018 06:13:19: Finished Epoch[17 of 50]: [Training] CrossEntropyWithSoftmax = 0.16203029 * 10000; EvalClassificationError = 0.07860000 * 10000; totalSamplesSeen = 170000; learningRatePerSample = 0.1; epochTime=0.0523749s
01/17/2018 06:13:19: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.17'

01/17/2018 06:13:19: Starting Epoch 18: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:19: Starting minibatch loop.
01/17/2018 06:13:19:  Epoch[18 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.15390018 * 1280; EvalClassificationError = 0.07421875 * 1280; time = 0.0080s; samplesPerSecond = 159551.3
01/17/2018 06:13:19:  Epoch[18 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.15887769 * 1280; EvalClassificationError = 0.07656250 * 1280; time = 0.0140s; samplesPerSecond = 91584.3
01/17/2018 06:13:19:  Epoch[18 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.16481850 * 1280; EvalClassificationError = 0.08281250 * 1280; time = 0.0063s; samplesPerSecond = 203779.5
01/17/2018 06:13:19:  Epoch[18 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.15366807 * 1280; EvalClassificationError = 0.07187500 * 1280; time = 0.0093s; samplesPerSecond = 137429.0
01/17/2018 06:13:19:  Epoch[18 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.14910550 * 1280; EvalClassificationError = 0.07265625 * 1280; time = 0.0063s; samplesPerSecond = 201778.2
01/17/2018 06:13:19:  Epoch[18 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.15585804 * 1280; EvalClassificationError = 0.07421875 * 1280; time = 0.0053s; samplesPerSecond = 241902.0
01/17/2018 06:13:19:  Epoch[18 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.15353146 * 1280; EvalClassificationError = 0.06796875 * 1280; time = 0.0053s; samplesPerSecond = 243341.4
01/17/2018 06:13:19: Finished Epoch[18 of 50]: [Training] CrossEntropyWithSoftmax = 0.15599238 * 10000; EvalClassificationError = 0.07440000 * 10000; totalSamplesSeen = 180000; learningRatePerSample = 0.1; epochTime=0.0614739s
01/17/2018 06:13:19: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.18'

01/17/2018 06:13:19: Starting Epoch 19: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:19: Starting minibatch loop.
01/17/2018 06:13:19:  Epoch[19 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.15185032 * 1280; EvalClassificationError = 0.07734375 * 1280; time = 0.0052s; samplesPerSecond = 245450.5
01/17/2018 06:13:19:  Epoch[19 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.14510214 * 1280; EvalClassificationError = 0.07031250 * 1280; time = 0.0054s; samplesPerSecond = 239140.6
01/17/2018 06:13:19:  Epoch[19 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.16338732 * 1280; EvalClassificationError = 0.08671875 * 1280; time = 0.0056s; samplesPerSecond = 230191.0
01/17/2018 06:13:19:  Epoch[19 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.14804358 * 1280; EvalClassificationError = 0.06875000 * 1280; time = 0.0052s; samplesPerSecond = 246381.3
01/17/2018 06:13:19:  Epoch[19 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.17848501 * 1280; EvalClassificationError = 0.08750000 * 1280; time = 0.0052s; samplesPerSecond = 246980.3
01/17/2018 06:13:19:  Epoch[19 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.16291170 * 1280; EvalClassificationError = 0.07109375 * 1280; time = 0.0054s; samplesPerSecond = 237129.3
01/17/2018 06:13:19:  Epoch[19 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.17254391 * 1280; EvalClassificationError = 0.09140625 * 1280; time = 0.0051s; samplesPerSecond = 248775.6
01/17/2018 06:13:19: Finished Epoch[19 of 50]: [Training] CrossEntropyWithSoftmax = 0.16105387 * 10000; EvalClassificationError = 0.07880000 * 10000; totalSamplesSeen = 190000; learningRatePerSample = 0.1; epochTime=0.0432995s
01/17/2018 06:13:19: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.19'

01/17/2018 06:13:19: Starting Epoch 20: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:19: Starting minibatch loop.
01/17/2018 06:13:19:  Epoch[20 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.16133311 * 1280; EvalClassificationError = 0.06953125 * 1280; time = 0.0060s; samplesPerSecond = 213885.9
01/17/2018 06:13:19:  Epoch[20 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.16010811 * 1280; EvalClassificationError = 0.07109375 * 1280; time = 0.0052s; samplesPerSecond = 244260.8
01/17/2018 06:13:19:  Epoch[20 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.15739422 * 1280; EvalClassificationError = 0.07656250 * 1280; time = 0.0052s; samplesPerSecond = 247467.3
01/17/2018 06:13:19:  Epoch[20 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.17021790 * 1280; EvalClassificationError = 0.08593750 * 1280; time = 0.0233s; samplesPerSecond = 54963.9
01/17/2018 06:13:19:  Epoch[20 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.16135745 * 1280; EvalClassificationError = 0.07968750 * 1280; time = 0.0067s; samplesPerSecond = 190808.4
01/17/2018 06:13:19:  Epoch[20 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.15534086 * 1280; EvalClassificationError = 0.07500000 * 1280; time = 0.0087s; samplesPerSecond = 147314.4
01/17/2018 06:13:19:  Epoch[20 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.16405239 * 1280; EvalClassificationError = 0.08046875 * 1280; time = 0.0083s; samplesPerSecond = 153372.4
01/17/2018 06:13:19: Finished Epoch[20 of 50]: [Training] CrossEntropyWithSoftmax = 0.15970442 * 10000; EvalClassificationError = 0.07630000 * 10000; totalSamplesSeen = 200000; learningRatePerSample = 0.1; epochTime=0.0711547s
01/17/2018 06:13:19: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.20'

01/17/2018 06:13:19: Starting Epoch 21: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:19: Starting minibatch loop.
01/17/2018 06:13:19:  Epoch[21 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.16151631 * 1280; EvalClassificationError = 0.08359375 * 1280; time = 0.0074s; samplesPerSecond = 173915.4
01/17/2018 06:13:19:  Epoch[21 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.14376364 * 1280; EvalClassificationError = 0.07421875 * 1280; time = 0.0073s; samplesPerSecond = 176318.3
01/17/2018 06:13:19:  Epoch[21 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.15459173 * 1280; EvalClassificationError = 0.07656250 * 1280; time = 0.0078s; samplesPerSecond = 163869.4
01/17/2018 06:13:19:  Epoch[21 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.14109597 * 1280; EvalClassificationError = 0.07031250 * 1280; time = 0.0089s; samplesPerSecond = 144160.4
01/17/2018 06:13:19:  Epoch[21 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.15917463 * 1280; EvalClassificationError = 0.06953125 * 1280; time = 0.0074s; samplesPerSecond = 173592.3
01/17/2018 06:13:19:  Epoch[21 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.16572986 * 1280; EvalClassificationError = 0.08359375 * 1280; time = 0.0055s; samplesPerSecond = 232630.0
01/17/2018 06:13:19:  Epoch[21 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.16026697 * 1280; EvalClassificationError = 0.08281250 * 1280; time = 0.0052s; samplesPerSecond = 248427.9
01/17/2018 06:13:19: Finished Epoch[21 of 50]: [Training] CrossEntropyWithSoftmax = 0.15594215 * 10000; EvalClassificationError = 0.07770000 * 10000; totalSamplesSeen = 210000; learningRatePerSample = 0.1; epochTime=0.0563767s
01/17/2018 06:13:19: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.21'

01/17/2018 06:13:19: Starting Epoch 22: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:19: Starting minibatch loop.
01/17/2018 06:13:19:  Epoch[22 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.15155299 * 1280; EvalClassificationError = 0.06953125 * 1280; time = 0.0058s; samplesPerSecond = 221875.5
01/17/2018 06:13:19:  Epoch[22 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.13562584 * 1280; EvalClassificationError = 0.06250000 * 1280; time = 0.0056s; samplesPerSecond = 227059.1
01/17/2018 06:13:19:  Epoch[22 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.16118867 * 1280; EvalClassificationError = 0.08125000 * 1280; time = 0.0051s; samplesPerSecond = 252874.5
01/17/2018 06:13:19:  Epoch[22 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.16034932 * 1280; EvalClassificationError = 0.07890625 * 1280; time = 0.0053s; samplesPerSecond = 240538.2
01/17/2018 06:13:19:  Epoch[22 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.15333858 * 1280; EvalClassificationError = 0.07109375 * 1280; time = 0.0054s; samplesPerSecond = 235862.1
01/17/2018 06:13:19:  Epoch[22 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.14866667 * 1280; EvalClassificationError = 0.07343750 * 1280; time = 0.0074s; samplesPerSecond = 171929.8
01/17/2018 06:13:19:  Epoch[22 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.18244123 * 1280; EvalClassificationError = 0.08671875 * 1280; time = 0.0051s; samplesPerSecond = 250141.7
01/17/2018 06:13:19: Finished Epoch[22 of 50]: [Training] CrossEntropyWithSoftmax = 0.15800615 * 10000; EvalClassificationError = 0.07550000 * 10000; totalSamplesSeen = 220000; learningRatePerSample = 0.1; epochTime=0.0464607s
01/17/2018 06:13:19: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.22'

01/17/2018 06:13:19: Starting Epoch 23: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:19: Starting minibatch loop.
01/17/2018 06:13:20:  Epoch[23 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.17648998 * 1280; EvalClassificationError = 0.08046875 * 1280; time = 0.0054s; samplesPerSecond = 235212.0
01/17/2018 06:13:20:  Epoch[23 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.17531897 * 1280; EvalClassificationError = 0.08593750 * 1280; time = 0.0054s; samplesPerSecond = 236010.0
01/17/2018 06:13:20:  Epoch[23 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.15463798 * 1280; EvalClassificationError = 0.07578125 * 1280; time = 0.0051s; samplesPerSecond = 249415.4
01/17/2018 06:13:20:  Epoch[23 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.17028651 * 1280; EvalClassificationError = 0.08359375 * 1280; time = 0.0059s; samplesPerSecond = 216604.1
01/17/2018 06:13:20:  Epoch[23 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.15742364 * 1280; EvalClassificationError = 0.07812500 * 1280; time = 0.0074s; samplesPerSecond = 172541.6
01/17/2018 06:13:20:  Epoch[23 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.16154852 * 1280; EvalClassificationError = 0.07500000 * 1280; time = 0.0068s; samplesPerSecond = 188310.1
01/17/2018 06:13:20:  Epoch[23 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.14952307 * 1280; EvalClassificationError = 0.07109375 * 1280; time = 0.0063s; samplesPerSecond = 202698.4
01/17/2018 06:13:20: Finished Epoch[23 of 50]: [Training] CrossEntropyWithSoftmax = 0.16315575 * 10000; EvalClassificationError = 0.07850000 * 10000; totalSamplesSeen = 230000; learningRatePerSample = 0.1; epochTime=0.0516804s
01/17/2018 06:13:20: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.23'

01/17/2018 06:13:20: Starting Epoch 24: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:20: Starting minibatch loop.
01/17/2018 06:13:20:  Epoch[24 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.16309845 * 1280; EvalClassificationError = 0.07890625 * 1280; time = 0.0070s; samplesPerSecond = 181686.6
01/17/2018 06:13:20:  Epoch[24 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.17734957 * 1280; EvalClassificationError = 0.08671875 * 1280; time = 0.0066s; samplesPerSecond = 193745.6
01/17/2018 06:13:20:  Epoch[24 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.14409983 * 1280; EvalClassificationError = 0.07031250 * 1280; time = 0.0073s; samplesPerSecond = 175607.1
01/17/2018 06:13:20:  Epoch[24 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.12802043 * 1280; EvalClassificationError = 0.05859375 * 1280; time = 0.0082s; samplesPerSecond = 155170.3
01/17/2018 06:13:20:  Epoch[24 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.17809982 * 1280; EvalClassificationError = 0.08281250 * 1280; time = 0.0060s; samplesPerSecond = 212999.6
01/17/2018 06:13:20:  Epoch[24 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.13289251 * 1280; EvalClassificationError = 0.05937500 * 1280; time = 0.0053s; samplesPerSecond = 241906.5
01/17/2018 06:13:20:  Epoch[24 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.15595989 * 1280; EvalClassificationError = 0.08281250 * 1280; time = 0.0052s; samplesPerSecond = 248452.0
01/17/2018 06:13:20: Finished Epoch[24 of 50]: [Training] CrossEntropyWithSoftmax = 0.15495480 * 10000; EvalClassificationError = 0.07510000 * 10000; totalSamplesSeen = 240000; learningRatePerSample = 0.1; epochTime=0.0529813s
01/17/2018 06:13:20: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.24'

01/17/2018 06:13:20: Starting Epoch 25: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:20: Starting minibatch loop.
01/17/2018 06:13:20:  Epoch[25 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.13681033 * 1280; EvalClassificationError = 0.06015625 * 1280; time = 0.0056s; samplesPerSecond = 229798.4
01/17/2018 06:13:20:  Epoch[25 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.17395720 * 1280; EvalClassificationError = 0.08750000 * 1280; time = 0.0052s; samplesPerSecond = 243907.1
01/17/2018 06:13:20:  Epoch[25 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.14068034 * 1280; EvalClassificationError = 0.06640625 * 1280; time = 0.0056s; samplesPerSecond = 226576.7
01/17/2018 06:13:20:  Epoch[25 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.15338879 * 1280; EvalClassificationError = 0.07578125 * 1280; time = 0.0051s; samplesPerSecond = 251978.4
01/17/2018 06:13:20:  Epoch[25 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.16744380 * 1280; EvalClassificationError = 0.08359375 * 1280; time = 0.0053s; samplesPerSecond = 243586.8
01/17/2018 06:13:20:  Epoch[25 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.16937728 * 1280; EvalClassificationError = 0.08828125 * 1280; time = 0.0054s; samplesPerSecond = 235649.3
01/17/2018 06:13:20:  Epoch[25 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.17972708 * 1280; EvalClassificationError = 0.09765625 * 1280; time = 0.0052s; samplesPerSecond = 245361.1
01/17/2018 06:13:20: Finished Epoch[25 of 50]: [Training] CrossEntropyWithSoftmax = 0.16098881 * 10000; EvalClassificationError = 0.07980000 * 10000; totalSamplesSeen = 250000; learningRatePerSample = 0.1; epochTime=0.043677s
01/17/2018 06:13:20: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.25'

01/17/2018 06:13:20: Starting Epoch 26: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:20: Starting minibatch loop.
01/17/2018 06:13:20:  Epoch[26 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.15768273 * 1280; EvalClassificationError = 0.06875000 * 1280; time = 0.0055s; samplesPerSecond = 233130.0
01/17/2018 06:13:20:  Epoch[26 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.19794986 * 1280; EvalClassificationError = 0.09218750 * 1280; time = 0.0052s; samplesPerSecond = 244919.8
01/17/2018 06:13:20:  Epoch[26 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.16209564 * 1280; EvalClassificationError = 0.08125000 * 1280; time = 0.0057s; samplesPerSecond = 223647.2
01/17/2018 06:13:20:  Epoch[26 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.14502654 * 1280; EvalClassificationError = 0.07343750 * 1280; time = 0.0052s; samplesPerSecond = 245285.9
01/17/2018 06:13:20:  Epoch[26 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.16414237 * 1280; EvalClassificationError = 0.08203125 * 1280; time = 0.0052s; samplesPerSecond = 244088.5
01/17/2018 06:13:20:  Epoch[26 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.12360010 * 1280; EvalClassificationError = 0.05312500 * 1280; time = 0.0054s; samplesPerSecond = 234871.0
01/17/2018 06:13:20:  Epoch[26 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.16247244 * 1280; EvalClassificationError = 0.07812500 * 1280; time = 0.0052s; samplesPerSecond = 247587.0
01/17/2018 06:13:20: Finished Epoch[26 of 50]: [Training] CrossEntropyWithSoftmax = 0.15834043 * 10000; EvalClassificationError = 0.07490000 * 10000; totalSamplesSeen = 260000; learningRatePerSample = 0.1; epochTime=0.0439679s
01/17/2018 06:13:20: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.26'

01/17/2018 06:13:20: Starting Epoch 27: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:20: Starting minibatch loop.
01/17/2018 06:13:20:  Epoch[27 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.16252776 * 1280; EvalClassificationError = 0.07734375 * 1280; time = 0.0056s; samplesPerSecond = 227758.0
01/17/2018 06:13:20:  Epoch[27 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.16304196 * 1280; EvalClassificationError = 0.09218750 * 1280; time = 0.0053s; samplesPerSecond = 242277.4
01/17/2018 06:13:20:  Epoch[27 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.14196091 * 1280; EvalClassificationError = 0.06718750 * 1280; time = 0.0052s; samplesPerSecond = 244265.5
01/17/2018 06:13:20:  Epoch[27 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.15195932 * 1280; EvalClassificationError = 0.06328125 * 1280; time = 0.0072s; samplesPerSecond = 177243.6
01/17/2018 06:13:20:  Epoch[27 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.15752149 * 1280; EvalClassificationError = 0.08046875 * 1280; time = 0.0053s; samplesPerSecond = 242598.8
01/17/2018 06:13:20:  Epoch[27 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.15266395 * 1280; EvalClassificationError = 0.07187500 * 1280; time = 0.0056s; samplesPerSecond = 230104.1
01/17/2018 06:13:20:  Epoch[27 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.16626177 * 1280; EvalClassificationError = 0.08437500 * 1280; time = 0.0052s; samplesPerSecond = 243981.5
01/17/2018 06:13:20: Finished Epoch[27 of 50]: [Training] CrossEntropyWithSoftmax = 0.15593679 * 10000; EvalClassificationError = 0.07620000 * 10000; totalSamplesSeen = 270000; learningRatePerSample = 0.1; epochTime=0.0456036s
01/17/2018 06:13:20: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.27'

01/17/2018 06:13:20: Starting Epoch 28: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:20: Starting minibatch loop.
01/17/2018 06:13:20:  Epoch[28 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.15549572 * 1280; EvalClassificationError = 0.07656250 * 1280; time = 0.0058s; samplesPerSecond = 221097.5
01/17/2018 06:13:20:  Epoch[28 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.16504967 * 1280; EvalClassificationError = 0.07968750 * 1280; time = 0.0053s; samplesPerSecond = 242557.5
01/17/2018 06:13:20:  Epoch[28 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.16705046 * 1280; EvalClassificationError = 0.08750000 * 1280; time = 0.0053s; samplesPerSecond = 243545.1
01/17/2018 06:13:20:  Epoch[28 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.14224138 * 1280; EvalClassificationError = 0.07421875 * 1280; time = 0.0061s; samplesPerSecond = 211346.7
01/17/2018 06:13:20:  Epoch[28 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.14113736 * 1280; EvalClassificationError = 0.07031250 * 1280; time = 0.0054s; samplesPerSecond = 236839.7
01/17/2018 06:13:20:  Epoch[28 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.15290103 * 1280; EvalClassificationError = 0.07265625 * 1280; time = 0.0059s; samplesPerSecond = 217159.0
01/17/2018 06:13:20:  Epoch[28 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.16408501 * 1280; EvalClassificationError = 0.07968750 * 1280; time = 0.0052s; samplesPerSecond = 245771.0
01/17/2018 06:13:20: Finished Epoch[28 of 50]: [Training] CrossEntropyWithSoftmax = 0.15697496 * 10000; EvalClassificationError = 0.07790000 * 10000; totalSamplesSeen = 280000; learningRatePerSample = 0.1; epochTime=0.0451096s
01/17/2018 06:13:20: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.28'

01/17/2018 06:13:20: Starting Epoch 29: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:20: Starting minibatch loop.
01/17/2018 06:13:20:  Epoch[29 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.15316752 * 1280; EvalClassificationError = 0.07578125 * 1280; time = 0.0057s; samplesPerSecond = 224896.8
01/17/2018 06:13:20:  Epoch[29 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.13558725 * 1280; EvalClassificationError = 0.05859375 * 1280; time = 0.0052s; samplesPerSecond = 246866.0
01/17/2018 06:13:20:  Epoch[29 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.16534040 * 1280; EvalClassificationError = 0.08750000 * 1280; time = 0.0051s; samplesPerSecond = 249298.8
01/17/2018 06:13:20:  Epoch[29 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.14491835 * 1280; EvalClassificationError = 0.07187500 * 1280; time = 0.0068s; samplesPerSecond = 189002.4
01/17/2018 06:13:20:  Epoch[29 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.16231375 * 1280; EvalClassificationError = 0.07968750 * 1280; time = 0.0052s; samplesPerSecond = 247630.1
01/17/2018 06:13:20:  Epoch[29 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.14071465 * 1280; EvalClassificationError = 0.06953125 * 1280; time = 0.0056s; samplesPerSecond = 230062.7
01/17/2018 06:13:20:  Epoch[29 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.16054001 * 1280; EvalClassificationError = 0.07578125 * 1280; time = 0.0052s; samplesPerSecond = 246400.2
01/17/2018 06:13:20: Finished Epoch[29 of 50]: [Training] CrossEntropyWithSoftmax = 0.15658409 * 10000; EvalClassificationError = 0.07700000 * 10000; totalSamplesSeen = 290000; learningRatePerSample = 0.1; epochTime=0.0451446s
01/17/2018 06:13:20: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.29'

01/17/2018 06:13:20: Starting Epoch 30: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:20: Starting minibatch loop.
01/17/2018 06:13:20:  Epoch[30 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.15555742 * 1280; EvalClassificationError = 0.08359375 * 1280; time = 0.0057s; samplesPerSecond = 222996.5
01/17/2018 06:13:20:  Epoch[30 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.14723208 * 1280; EvalClassificationError = 0.08046875 * 1280; time = 0.0054s; samplesPerSecond = 237481.2
01/17/2018 06:13:20:  Epoch[30 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.14683332 * 1280; EvalClassificationError = 0.07812500 * 1280; time = 0.0058s; samplesPerSecond = 222512.0
01/17/2018 06:13:20:  Epoch[30 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.16358609 * 1280; EvalClassificationError = 0.08125000 * 1280; time = 0.0057s; samplesPerSecond = 225904.9
01/17/2018 06:13:20:  Epoch[30 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.15779209 * 1280; EvalClassificationError = 0.07187500 * 1280; time = 0.0051s; samplesPerSecond = 248785.2
01/17/2018 06:13:20:  Epoch[30 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.15689759 * 1280; EvalClassificationError = 0.07031250 * 1280; time = 0.0053s; samplesPerSecond = 243096.4
01/17/2018 06:13:20:  Epoch[30 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.15914001 * 1280; EvalClassificationError = 0.07812500 * 1280; time = 0.0055s; samplesPerSecond = 231766.5
01/17/2018 06:13:20: Finished Epoch[30 of 50]: [Training] CrossEntropyWithSoftmax = 0.15398387 * 10000; EvalClassificationError = 0.07660000 * 10000; totalSamplesSeen = 300000; learningRatePerSample = 0.1; epochTime=0.0447575s
01/17/2018 06:13:20: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.30'

01/17/2018 06:13:20: Starting Epoch 31: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:20: Starting minibatch loop.
01/17/2018 06:13:20:  Epoch[31 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.15854740 * 1280; EvalClassificationError = 0.07343750 * 1280; time = 0.0129s; samplesPerSecond = 99540.4
01/17/2018 06:13:20:  Epoch[31 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.15418732 * 1280; EvalClassificationError = 0.08046875 * 1280; time = 0.0115s; samplesPerSecond = 111151.6
01/17/2018 06:13:20:  Epoch[31 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.15642931 * 1280; EvalClassificationError = 0.07421875 * 1280; time = 0.0076s; samplesPerSecond = 167915.1
01/17/2018 06:13:20:  Epoch[31 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.14554515 * 1280; EvalClassificationError = 0.07734375 * 1280; time = 0.0160s; samplesPerSecond = 79984.5
01/17/2018 06:13:20:  Epoch[31 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.17074995 * 1280; EvalClassificationError = 0.08359375 * 1280; time = 0.0054s; samplesPerSecond = 237856.3
01/17/2018 06:13:20:  Epoch[31 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.13945341 * 1280; EvalClassificationError = 0.07109375 * 1280; time = 0.0055s; samplesPerSecond = 232389.3
01/17/2018 06:13:20:  Epoch[31 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.14870253 * 1280; EvalClassificationError = 0.07421875 * 1280; time = 0.0052s; samplesPerSecond = 246966.0
01/17/2018 06:13:20: Finished Epoch[31 of 50]: [Training] CrossEntropyWithSoftmax = 0.15280536 * 10000; EvalClassificationError = 0.07570000 * 10000; totalSamplesSeen = 310000; learningRatePerSample = 0.1; epochTime=0.070549s
01/17/2018 06:13:20: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.31'

01/17/2018 06:13:20: Starting Epoch 32: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:20: Starting minibatch loop.
01/17/2018 06:13:20:  Epoch[32 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.15338740 * 1280; EvalClassificationError = 0.07734375 * 1280; time = 0.0054s; samplesPerSecond = 237111.7
01/17/2018 06:13:20:  Epoch[32 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.14309559 * 1280; EvalClassificationError = 0.06562500 * 1280; time = 0.0056s; samplesPerSecond = 229872.7
01/17/2018 06:13:20:  Epoch[32 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.14224100 * 1280; EvalClassificationError = 0.07031250 * 1280; time = 0.0051s; samplesPerSecond = 251093.6
01/17/2018 06:13:20:  Epoch[32 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.15188251 * 1280; EvalClassificationError = 0.07812500 * 1280; time = 0.0053s; samplesPerSecond = 243503.4
01/17/2018 06:13:20:  Epoch[32 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.15083723 * 1280; EvalClassificationError = 0.07500000 * 1280; time = 0.0054s; samplesPerSecond = 235553.9
01/17/2018 06:13:20:  Epoch[32 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.16061201 * 1280; EvalClassificationError = 0.07890625 * 1280; time = 0.0053s; samplesPerSecond = 239449.3
01/17/2018 06:13:20:  Epoch[32 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.15455847 * 1280; EvalClassificationError = 0.08203125 * 1280; time = 0.0051s; samplesPerSecond = 248785.2
01/17/2018 06:13:20: Finished Epoch[32 of 50]: [Training] CrossEntropyWithSoftmax = 0.15326575 * 10000; EvalClassificationError = 0.07650000 * 10000; totalSamplesSeen = 320000; learningRatePerSample = 0.1; epochTime=0.0439062s
01/17/2018 06:13:20: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.32'

01/17/2018 06:13:20: Starting Epoch 33: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:20: Starting minibatch loop.
01/17/2018 06:13:20:  Epoch[33 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.15867492 * 1280; EvalClassificationError = 0.07968750 * 1280; time = 0.0056s; samplesPerSecond = 228820.7
01/17/2018 06:13:20:  Epoch[33 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.17103983 * 1280; EvalClassificationError = 0.08750000 * 1280; time = 0.0055s; samplesPerSecond = 231984.9
01/17/2018 06:13:20:  Epoch[33 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.14906037 * 1280; EvalClassificationError = 0.07265625 * 1280; time = 0.0057s; samplesPerSecond = 222849.0
01/17/2018 06:13:20:  Epoch[33 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.15649753 * 1280; EvalClassificationError = 0.07421875 * 1280; time = 0.0051s; samplesPerSecond = 248616.1
01/17/2018 06:13:20:  Epoch[33 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.13757324 * 1280; EvalClassificationError = 0.07031250 * 1280; time = 0.0051s; samplesPerSecond = 251167.5
01/17/2018 06:13:20:  Epoch[33 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.14185958 * 1280; EvalClassificationError = 0.06953125 * 1280; time = 0.0057s; samplesPerSecond = 223983.8
01/17/2018 06:13:20:  Epoch[33 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.16102781 * 1280; EvalClassificationError = 0.08984375 * 1280; time = 0.0054s; samplesPerSecond = 238930.8
01/17/2018 06:13:20: Finished Epoch[33 of 50]: [Training] CrossEntropyWithSoftmax = 0.15331584 * 10000; EvalClassificationError = 0.07750000 * 10000; totalSamplesSeen = 330000; learningRatePerSample = 0.1; epochTime=0.0442492s
01/17/2018 06:13:20: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.33'

01/17/2018 06:13:20: Starting Epoch 34: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:20: Starting minibatch loop.
01/17/2018 06:13:20:  Epoch[34 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.14242564 * 1280; EvalClassificationError = 0.07421875 * 1280; time = 0.0066s; samplesPerSecond = 193672.4
01/17/2018 06:13:20:  Epoch[34 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.15581409 * 1280; EvalClassificationError = 0.07109375 * 1280; time = 0.0054s; samplesPerSecond = 237195.2
01/17/2018 06:13:20:  Epoch[34 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.15887907 * 1280; EvalClassificationError = 0.07968750 * 1280; time = 0.0057s; samplesPerSecond = 224778.3
01/17/2018 06:13:20:  Epoch[34 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.14696274 * 1280; EvalClassificationError = 0.08125000 * 1280; time = 0.0058s; samplesPerSecond = 222353.5
01/17/2018 06:13:20:  Epoch[34 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.15781937 * 1280; EvalClassificationError = 0.07656250 * 1280; time = 0.0085s; samplesPerSecond = 150088.5
01/17/2018 06:13:20:  Epoch[34 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.15656815 * 1280; EvalClassificationError = 0.07890625 * 1280; time = 0.0091s; samplesPerSecond = 140594.5
01/17/2018 06:13:20:  Epoch[34 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.15388136 * 1280; EvalClassificationError = 0.07656250 * 1280; time = 0.0062s; samplesPerSecond = 207176.7
01/17/2018 06:13:20: Finished Epoch[34 of 50]: [Training] CrossEntropyWithSoftmax = 0.15178218 * 10000; EvalClassificationError = 0.07640000 * 10000; totalSamplesSeen = 340000; learningRatePerSample = 0.1; epochTime=0.0538137s
01/17/2018 06:13:20: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.34'

01/17/2018 06:13:20: Starting Epoch 35: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:20: Starting minibatch loop.
01/17/2018 06:13:20:  Epoch[35 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.15963540 * 1280; EvalClassificationError = 0.08437500 * 1280; time = 0.0082s; samplesPerSecond = 156718.7
01/17/2018 06:13:20:  Epoch[35 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.14121661 * 1280; EvalClassificationError = 0.06875000 * 1280; time = 0.0055s; samplesPerSecond = 233632.1
01/17/2018 06:13:20:  Epoch[35 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.14684043 * 1280; EvalClassificationError = 0.07265625 * 1280; time = 0.0055s; samplesPerSecond = 233504.2
01/17/2018 06:13:20:  Epoch[35 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.14852238 * 1280; EvalClassificationError = 0.07421875 * 1280; time = 0.0061s; samplesPerSecond = 210994.8
01/17/2018 06:13:20:  Epoch[35 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.14079437 * 1280; EvalClassificationError = 0.06406250 * 1280; time = 0.0055s; samplesPerSecond = 234458.0
01/17/2018 06:13:20:  Epoch[35 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.14343047 * 1280; EvalClassificationError = 0.07031250 * 1280; time = 0.0051s; samplesPerSecond = 249488.4
01/17/2018 06:13:20:  Epoch[35 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.15912304 * 1280; EvalClassificationError = 0.08359375 * 1280; time = 0.0062s; samplesPerSecond = 206016.3
01/17/2018 06:13:20: Finished Epoch[35 of 50]: [Training] CrossEntropyWithSoftmax = 0.14920563 * 10000; EvalClassificationError = 0.07370000 * 10000; totalSamplesSeen = 350000; learningRatePerSample = 0.1; epochTime=0.0486591s
01/17/2018 06:13:20: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.35'

01/17/2018 06:13:20: Starting Epoch 36: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:20: Starting minibatch loop.
01/17/2018 06:13:20:  Epoch[36 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.14718659 * 1280; EvalClassificationError = 0.07656250 * 1280; time = 0.0058s; samplesPerSecond = 218889.5
01/17/2018 06:13:20:  Epoch[36 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.17286856 * 1280; EvalClassificationError = 0.08359375 * 1280; time = 0.0053s; samplesPerSecond = 243540.5
01/17/2018 06:13:20:  Epoch[36 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.15483851 * 1280; EvalClassificationError = 0.08125000 * 1280; time = 0.0055s; samplesPerSecond = 233117.2
01/17/2018 06:13:20:  Epoch[36 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.15607777 * 1280; EvalClassificationError = 0.08437500 * 1280; time = 0.0071s; samplesPerSecond = 180274.1
01/17/2018 06:13:20:  Epoch[36 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.15352387 * 1280; EvalClassificationError = 0.08125000 * 1280; time = 0.0069s; samplesPerSecond = 184741.5
01/17/2018 06:13:20:  Epoch[36 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.13617873 * 1280; EvalClassificationError = 0.06640625 * 1280; time = 0.0104s; samplesPerSecond = 122726.4
01/17/2018 06:13:20:  Epoch[36 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.14273376 * 1280; EvalClassificationError = 0.07343750 * 1280; time = 0.0058s; samplesPerSecond = 220644.0
01/17/2018 06:13:20: Finished Epoch[36 of 50]: [Training] CrossEntropyWithSoftmax = 0.15350920 * 10000; EvalClassificationError = 0.07830000 * 10000; totalSamplesSeen = 360000; learningRatePerSample = 0.1; epochTime=0.053722s
01/17/2018 06:13:20: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.36'

01/17/2018 06:13:20: Starting Epoch 37: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:20: Starting minibatch loop.
01/17/2018 06:13:20:  Epoch[37 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.16884112 * 1280; EvalClassificationError = 0.08515625 * 1280; time = 0.0062s; samplesPerSecond = 206959.0
01/17/2018 06:13:20:  Epoch[37 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.14396670 * 1280; EvalClassificationError = 0.07500000 * 1280; time = 0.0077s; samplesPerSecond = 165839.6
01/17/2018 06:13:20:  Epoch[37 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.14920857 * 1280; EvalClassificationError = 0.07343750 * 1280; time = 0.0056s; samplesPerSecond = 228873.9
01/17/2018 06:13:20:  Epoch[37 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.12401195 * 1280; EvalClassificationError = 0.05781250 * 1280; time = 0.0061s; samplesPerSecond = 210942.6
01/17/2018 06:13:20:  Epoch[37 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.15502052 * 1280; EvalClassificationError = 0.08593750 * 1280; time = 0.0058s; samplesPerSecond = 220022.0
01/17/2018 06:13:20:  Epoch[37 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.14399314 * 1280; EvalClassificationError = 0.06875000 * 1280; time = 0.0057s; samplesPerSecond = 222953.8
01/17/2018 06:13:20:  Epoch[37 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.16337833 * 1280; EvalClassificationError = 0.07734375 * 1280; time = 0.0059s; samplesPerSecond = 215491.8
01/17/2018 06:13:20: Finished Epoch[37 of 50]: [Training] CrossEntropyWithSoftmax = 0.14998220 * 10000; EvalClassificationError = 0.07480000 * 10000; totalSamplesSeen = 370000; learningRatePerSample = 0.1; epochTime=0.0499222s
01/17/2018 06:13:20: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.37'

01/17/2018 06:13:20: Starting Epoch 38: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:20: Starting minibatch loop.
01/17/2018 06:13:20:  Epoch[38 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.16702759 * 1280; EvalClassificationError = 0.08671875 * 1280; time = 0.0066s; samplesPerSecond = 194997.1
01/17/2018 06:13:20:  Epoch[38 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.14558318 * 1280; EvalClassificationError = 0.07968750 * 1280; time = 0.0064s; samplesPerSecond = 199694.2
01/17/2018 06:13:20:  Epoch[38 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.13988147 * 1280; EvalClassificationError = 0.06953125 * 1280; time = 0.0057s; samplesPerSecond = 223576.9
01/17/2018 06:13:20:  Epoch[38 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.15891962 * 1280; EvalClassificationError = 0.08750000 * 1280; time = 0.0061s; samplesPerSecond = 208720.6
01/17/2018 06:13:20:  Epoch[38 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.13741221 * 1280; EvalClassificationError = 0.06562500 * 1280; time = 0.0053s; samplesPerSecond = 243013.4
01/17/2018 06:13:20:  Epoch[38 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.15441995 * 1280; EvalClassificationError = 0.08906250 * 1280; time = 0.0054s; samplesPerSecond = 236432.8
01/17/2018 06:13:20:  Epoch[38 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.13708830 * 1280; EvalClassificationError = 0.06875000 * 1280; time = 0.0059s; samplesPerSecond = 215648.0
01/17/2018 06:13:20: Finished Epoch[38 of 50]: [Training] CrossEntropyWithSoftmax = 0.14877498 * 10000; EvalClassificationError = 0.07770000 * 10000; totalSamplesSeen = 380000; learningRatePerSample = 0.1; epochTime=0.0479252s
01/17/2018 06:13:20: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.38'

01/17/2018 06:13:20: Starting Epoch 39: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:20: Starting minibatch loop.
01/17/2018 06:13:20:  Epoch[39 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.14556807 * 1280; EvalClassificationError = 0.07734375 * 1280; time = 0.0054s; samplesPerSecond = 237529.7
01/17/2018 06:13:20:  Epoch[39 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.13892974 * 1280; EvalClassificationError = 0.06953125 * 1280; time = 0.0051s; samplesPerSecond = 253214.6
01/17/2018 06:13:20:  Epoch[39 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.14860389 * 1280; EvalClassificationError = 0.07734375 * 1280; time = 0.0052s; samplesPerSecond = 245544.7
01/17/2018 06:13:20:  Epoch[39 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.14665070 * 1280; EvalClassificationError = 0.07500000 * 1280; time = 0.0060s; samplesPerSecond = 212022.3
01/17/2018 06:13:20:  Epoch[39 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.15661945 * 1280; EvalClassificationError = 0.08750000 * 1280; time = 0.0055s; samplesPerSecond = 233623.5
01/17/2018 06:13:20:  Epoch[39 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.17463212 * 1280; EvalClassificationError = 0.09062500 * 1280; time = 0.0052s; samplesPerSecond = 247922.7
01/17/2018 06:13:20:  Epoch[39 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.18018923 * 1280; EvalClassificationError = 0.09609375 * 1280; time = 0.0054s; samplesPerSecond = 235627.6
01/17/2018 06:13:20: Finished Epoch[39 of 50]: [Training] CrossEntropyWithSoftmax = 0.15433470 * 10000; EvalClassificationError = 0.08030000 * 10000; totalSamplesSeen = 390000; learningRatePerSample = 0.1; epochTime=0.0446369s
01/17/2018 06:13:20: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.39'

01/17/2018 06:13:20: Starting Epoch 40: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:20: Starting minibatch loop.
01/17/2018 06:13:20:  Epoch[40 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.18012108 * 1280; EvalClassificationError = 0.09140625 * 1280; time = 0.0065s; samplesPerSecond = 197698.7
01/17/2018 06:13:20:  Epoch[40 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.14649881 * 1280; EvalClassificationError = 0.07187500 * 1280; time = 0.0060s; samplesPerSecond = 212741.2
01/17/2018 06:13:20:  Epoch[40 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.15320475 * 1280; EvalClassificationError = 0.07421875 * 1280; time = 0.0054s; samplesPerSecond = 235164.4
01/17/2018 06:13:20:  Epoch[40 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.15182514 * 1280; EvalClassificationError = 0.07578125 * 1280; time = 0.0056s; samplesPerSecond = 228759.3
01/17/2018 06:13:20:  Epoch[40 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.15425653 * 1280; EvalClassificationError = 0.08046875 * 1280; time = 0.0054s; samplesPerSecond = 235796.9
01/17/2018 06:13:20:  Epoch[40 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.15766792 * 1280; EvalClassificationError = 0.08437500 * 1280; time = 0.0051s; samplesPerSecond = 248969.1
01/17/2018 06:13:20:  Epoch[40 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.14060678 * 1280; EvalClassificationError = 0.07578125 * 1280; time = 0.0057s; samplesPerSecond = 222996.5
01/17/2018 06:13:20: Finished Epoch[40 of 50]: [Training] CrossEntropyWithSoftmax = 0.15285542 * 10000; EvalClassificationError = 0.07730000 * 10000; totalSamplesSeen = 400000; learningRatePerSample = 0.1; epochTime=0.0465558s
01/17/2018 06:13:20: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.40'

01/17/2018 06:13:20: Starting Epoch 41: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:20: Starting minibatch loop.
01/17/2018 06:13:20:  Epoch[41 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.16204326 * 1280; EvalClassificationError = 0.08046875 * 1280; time = 0.0065s; samplesPerSecond = 196397.3
01/17/2018 06:13:20:  Epoch[41 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.13063951 * 1280; EvalClassificationError = 0.06406250 * 1280; time = 0.0053s; samplesPerSecond = 242635.6
01/17/2018 06:13:20:  Epoch[41 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.14536316 * 1280; EvalClassificationError = 0.07265625 * 1280; time = 0.0064s; samplesPerSecond = 199215.6
01/17/2018 06:13:20:  Epoch[41 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.14750862 * 1280; EvalClassificationError = 0.07187500 * 1280; time = 0.0063s; samplesPerSecond = 201673.3
01/17/2018 06:13:20:  Epoch[41 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.14682159 * 1280; EvalClassificationError = 0.07265625 * 1280; time = 0.0057s; samplesPerSecond = 224813.8
01/17/2018 06:13:20:  Epoch[41 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.14233489 * 1280; EvalClassificationError = 0.07031250 * 1280; time = 0.0060s; samplesPerSecond = 212226.2
01/17/2018 06:13:20:  Epoch[41 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.15025969 * 1280; EvalClassificationError = 0.07656250 * 1280; time = 0.0081s; samplesPerSecond = 158541.4
01/17/2018 06:13:20: Finished Epoch[41 of 50]: [Training] CrossEntropyWithSoftmax = 0.14567303 * 10000; EvalClassificationError = 0.07300000 * 10000; totalSamplesSeen = 410000; learningRatePerSample = 0.1; epochTime=0.0524497s
01/17/2018 06:13:20: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.41'

01/17/2018 06:13:20: Starting Epoch 42: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:20: Starting minibatch loop.
01/17/2018 06:13:20:  Epoch[42 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.14131975 * 1280; EvalClassificationError = 0.07187500 * 1280; time = 0.0058s; samplesPerSecond = 220781.0
01/17/2018 06:13:20:  Epoch[42 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.14017251 * 1280; EvalClassificationError = 0.07187500 * 1280; time = 0.0054s; samplesPerSecond = 237878.4
01/17/2018 06:13:20:  Epoch[42 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.15156863 * 1280; EvalClassificationError = 0.07734375 * 1280; time = 0.0062s; samplesPerSecond = 206182.2
01/17/2018 06:13:20:  Epoch[42 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.14153347 * 1280; EvalClassificationError = 0.07265625 * 1280; time = 0.0058s; samplesPerSecond = 219287.0
01/17/2018 06:13:20:  Epoch[42 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.14911618 * 1280; EvalClassificationError = 0.06796875 * 1280; time = 0.0059s; samplesPerSecond = 218240.1
01/17/2018 06:13:20:  Epoch[42 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.15729055 * 1280; EvalClassificationError = 0.08203125 * 1280; time = 0.0065s; samplesPerSecond = 197436.4
01/17/2018 06:13:20:  Epoch[42 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.17820501 * 1280; EvalClassificationError = 0.09062500 * 1280; time = 0.0064s; samplesPerSecond = 198788.6
01/17/2018 06:13:20: Finished Epoch[42 of 50]: [Training] CrossEntropyWithSoftmax = 0.15131997 * 10000; EvalClassificationError = 0.07720000 * 10000; totalSamplesSeen = 420000; learningRatePerSample = 0.1; epochTime=0.0517574s
01/17/2018 06:13:20: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.42'

01/17/2018 06:13:20: Starting Epoch 43: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:20: Starting minibatch loop.
01/17/2018 06:13:21:  Epoch[43 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.16356061 * 1280; EvalClassificationError = 0.08593750 * 1280; time = 0.0056s; samplesPerSecond = 229456.5
01/17/2018 06:13:21:  Epoch[43 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.12194111 * 1280; EvalClassificationError = 0.06171875 * 1280; time = 0.0058s; samplesPerSecond = 222314.9
01/17/2018 06:13:21:  Epoch[43 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.14642954 * 1280; EvalClassificationError = 0.07031250 * 1280; time = 0.0052s; samplesPerSecond = 245337.6
01/17/2018 06:13:21:  Epoch[43 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.15378962 * 1280; EvalClassificationError = 0.08203125 * 1280; time = 0.0053s; samplesPerSecond = 243091.8
01/17/2018 06:13:21:  Epoch[43 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.15869656 * 1280; EvalClassificationError = 0.08125000 * 1280; time = 0.0059s; samplesPerSecond = 217288.0
01/17/2018 06:13:21:  Epoch[43 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.14140344 * 1280; EvalClassificationError = 0.07890625 * 1280; time = 0.0053s; samplesPerSecond = 242277.4
01/17/2018 06:13:21:  Epoch[43 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.16585588 * 1280; EvalClassificationError = 0.08515625 * 1280; time = 0.0056s; samplesPerSecond = 228522.5
01/17/2018 06:13:21: Finished Epoch[43 of 50]: [Training] CrossEntropyWithSoftmax = 0.14952623 * 10000; EvalClassificationError = 0.07730000 * 10000; totalSamplesSeen = 430000; learningRatePerSample = 0.1; epochTime=0.0452791s
01/17/2018 06:13:21: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.43'

01/17/2018 06:13:21: Starting Epoch 44: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:21: Starting minibatch loop.
01/17/2018 06:13:21:  Epoch[44 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.16138092 * 1280; EvalClassificationError = 0.08125000 * 1280; time = 0.0055s; samplesPerSecond = 231218.0
01/17/2018 06:13:21:  Epoch[44 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.15588461 * 1280; EvalClassificationError = 0.07578125 * 1280; time = 0.0055s; samplesPerSecond = 233414.8
01/17/2018 06:13:21:  Epoch[44 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.12850945 * 1280; EvalClassificationError = 0.06328125 * 1280; time = 0.0057s; samplesPerSecond = 224062.2
01/17/2018 06:13:21:  Epoch[44 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.14631443 * 1280; EvalClassificationError = 0.07421875 * 1280; time = 0.0060s; samplesPerSecond = 213614.6
01/17/2018 06:13:21:  Epoch[44 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.15218453 * 1280; EvalClassificationError = 0.07656250 * 1280; time = 0.0059s; samplesPerSecond = 217476.3
01/17/2018 06:13:21:  Epoch[44 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.13426723 * 1280; EvalClassificationError = 0.06640625 * 1280; time = 0.0057s; samplesPerSecond = 224411.8
01/17/2018 06:13:21:  Epoch[44 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.15782404 * 1280; EvalClassificationError = 0.08125000 * 1280; time = 0.0057s; samplesPerSecond = 223245.4
01/17/2018 06:13:21: Finished Epoch[44 of 50]: [Training] CrossEntropyWithSoftmax = 0.14878915 * 10000; EvalClassificationError = 0.07400000 * 10000; totalSamplesSeen = 440000; learningRatePerSample = 0.1; epochTime=0.0470097s
01/17/2018 06:13:21: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.44'

01/17/2018 06:13:21: Starting Epoch 45: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:21: Starting minibatch loop.
01/17/2018 06:13:21:  Epoch[45 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.16446162 * 1280; EvalClassificationError = 0.08906250 * 1280; time = 0.0055s; samplesPerSecond = 233423.3
01/17/2018 06:13:21:  Epoch[45 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.13921901 * 1280; EvalClassificationError = 0.07265625 * 1280; time = 0.0056s; samplesPerSecond = 230050.3
01/17/2018 06:13:21:  Epoch[45 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.15290625 * 1280; EvalClassificationError = 0.07343750 * 1280; time = 0.0052s; samplesPerSecond = 244854.2
01/17/2018 06:13:21:  Epoch[45 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.15271430 * 1280; EvalClassificationError = 0.07890625 * 1280; time = 0.0057s; samplesPerSecond = 223748.8
01/17/2018 06:13:21:  Epoch[45 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.14689426 * 1280; EvalClassificationError = 0.07968750 * 1280; time = 0.0056s; samplesPerSecond = 227640.5
01/17/2018 06:13:21:  Epoch[45 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.13420410 * 1280; EvalClassificationError = 0.06406250 * 1280; time = 0.0057s; samplesPerSecond = 225841.2
01/17/2018 06:13:21:  Epoch[45 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.14893570 * 1280; EvalClassificationError = 0.07968750 * 1280; time = 0.0053s; samplesPerSecond = 240800.7
01/17/2018 06:13:21: Finished Epoch[45 of 50]: [Training] CrossEntropyWithSoftmax = 0.14720029 * 10000; EvalClassificationError = 0.07620000 * 10000; totalSamplesSeen = 450000; learningRatePerSample = 0.1; epochTime=0.0475341s
01/17/2018 06:13:21: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.45'

01/17/2018 06:13:21: Starting Epoch 46: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:21: Starting minibatch loop.
01/17/2018 06:13:21:  Epoch[46 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.12671082 * 1280; EvalClassificationError = 0.05703125 * 1280; time = 0.0056s; samplesPerSecond = 228575.5
01/17/2018 06:13:21:  Epoch[46 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.16278114 * 1280; EvalClassificationError = 0.08125000 * 1280; time = 0.0055s; samplesPerSecond = 232216.4
01/17/2018 06:13:21:  Epoch[46 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.14169729 * 1280; EvalClassificationError = 0.07734375 * 1280; time = 0.0053s; samplesPerSecond = 241984.3
01/17/2018 06:13:21:  Epoch[46 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.13762536 * 1280; EvalClassificationError = 0.06796875 * 1280; time = 0.0055s; samplesPerSecond = 231779.1
01/17/2018 06:13:21:  Epoch[46 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.14371409 * 1280; EvalClassificationError = 0.06875000 * 1280; time = 0.0057s; samplesPerSecond = 225272.8
01/17/2018 06:13:21:  Epoch[46 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.14948044 * 1280; EvalClassificationError = 0.07578125 * 1280; time = 0.0055s; samplesPerSecond = 230784.5
01/17/2018 06:13:21:  Epoch[46 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.14947481 * 1280; EvalClassificationError = 0.07500000 * 1280; time = 0.0059s; samplesPerSecond = 216567.4
01/17/2018 06:13:21: Finished Epoch[46 of 50]: [Training] CrossEntropyWithSoftmax = 0.14464042 * 10000; EvalClassificationError = 0.07170000 * 10000; totalSamplesSeen = 460000; learningRatePerSample = 0.1; epochTime=0.04652s
01/17/2018 06:13:21: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.46'

01/17/2018 06:13:21: Starting Epoch 47: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:21: Starting minibatch loop.
01/17/2018 06:13:21:  Epoch[47 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.14711735 * 1280; EvalClassificationError = 0.07890625 * 1280; time = 0.0056s; samplesPerSecond = 229612.9
01/17/2018 06:13:21:  Epoch[47 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.14898081 * 1280; EvalClassificationError = 0.07656250 * 1280; time = 0.0060s; samplesPerSecond = 211587.7
01/17/2018 06:13:21:  Epoch[47 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.15361421 * 1280; EvalClassificationError = 0.08671875 * 1280; time = 0.0059s; samplesPerSecond = 216252.7
01/17/2018 06:13:21:  Epoch[47 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.11621261 * 1280; EvalClassificationError = 0.06171875 * 1280; time = 0.0069s; samplesPerSecond = 184650.9
01/17/2018 06:13:21:  Epoch[47 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.13497615 * 1280; EvalClassificationError = 0.06953125 * 1280; time = 0.0055s; samplesPerSecond = 231322.5
01/17/2018 06:13:21:  Epoch[47 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.14900246 * 1280; EvalClassificationError = 0.07890625 * 1280; time = 0.0055s; samplesPerSecond = 233555.3
01/17/2018 06:13:21:  Epoch[47 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.13549509 * 1280; EvalClassificationError = 0.06796875 * 1280; time = 0.0057s; samplesPerSecond = 224529.9
01/17/2018 06:13:21: Finished Epoch[47 of 50]: [Training] CrossEntropyWithSoftmax = 0.14077632 * 10000; EvalClassificationError = 0.07380000 * 10000; totalSamplesSeen = 470000; learningRatePerSample = 0.1; epochTime=0.0480979s
01/17/2018 06:13:21: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.47'

01/17/2018 06:13:21: Starting Epoch 48: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:21: Starting minibatch loop.
01/17/2018 06:13:21:  Epoch[48 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.14503465 * 1280; EvalClassificationError = 0.07578125 * 1280; time = 0.0057s; samplesPerSecond = 225375.9
01/17/2018 06:13:21:  Epoch[48 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.14941282 * 1280; EvalClassificationError = 0.08203125 * 1280; time = 0.0054s; samplesPerSecond = 237958.0
01/17/2018 06:13:21:  Epoch[48 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.14857764 * 1280; EvalClassificationError = 0.07734375 * 1280; time = 0.0053s; samplesPerSecond = 242884.3
01/17/2018 06:13:21:  Epoch[48 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.13969493 * 1280; EvalClassificationError = 0.07187500 * 1280; time = 0.0072s; samplesPerSecond = 177619.9
01/17/2018 06:13:21:  Epoch[48 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.14006305 * 1280; EvalClassificationError = 0.07187500 * 1280; time = 0.0061s; samplesPerSecond = 210394.8
01/17/2018 06:13:21:  Epoch[48 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.13965473 * 1280; EvalClassificationError = 0.06875000 * 1280; time = 0.0053s; samplesPerSecond = 239965.5
01/17/2018 06:13:21:  Epoch[48 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.12673740 * 1280; EvalClassificationError = 0.06796875 * 1280; time = 0.0058s; samplesPerSecond = 219433.6
01/17/2018 06:13:21: Finished Epoch[48 of 50]: [Training] CrossEntropyWithSoftmax = 0.14152292 * 10000; EvalClassificationError = 0.07300000 * 10000; totalSamplesSeen = 480000; learningRatePerSample = 0.1; epochTime=0.047683s
01/17/2018 06:13:21: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.48'

01/17/2018 06:13:21: Starting Epoch 49: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:21: Starting minibatch loop.
01/17/2018 06:13:21:  Epoch[49 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.15557060 * 1280; EvalClassificationError = 0.08437500 * 1280; time = 0.0056s; samplesPerSecond = 230112.4
01/17/2018 06:13:21:  Epoch[49 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.13022280 * 1280; EvalClassificationError = 0.07656250 * 1280; time = 0.0054s; samplesPerSecond = 238592.3
01/17/2018 06:13:21:  Epoch[49 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.14758434 * 1280; EvalClassificationError = 0.07890625 * 1280; time = 0.0055s; samplesPerSecond = 234759.0
01/17/2018 06:13:21:  Epoch[49 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.14358916 * 1280; EvalClassificationError = 0.07343750 * 1280; time = 0.0055s; samplesPerSecond = 232418.8
01/17/2018 06:13:21:  Epoch[49 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.14146266 * 1280; EvalClassificationError = 0.06406250 * 1280; time = 0.0064s; samplesPerSecond = 200554.7
01/17/2018 06:13:21:  Epoch[49 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.13917103 * 1280; EvalClassificationError = 0.06015625 * 1280; time = 0.0062s; samplesPerSecond = 206611.6
01/17/2018 06:13:21:  Epoch[49 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.17072220 * 1280; EvalClassificationError = 0.08750000 * 1280; time = 0.0054s; samplesPerSecond = 235302.8
01/17/2018 06:13:21: Finished Epoch[49 of 50]: [Training] CrossEntropyWithSoftmax = 0.14609703 * 10000; EvalClassificationError = 0.07400000 * 10000; totalSamplesSeen = 490000; learningRatePerSample = 0.1; epochTime=0.0466362s
01/17/2018 06:13:21: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.49'

01/17/2018 06:13:21: Starting Epoch 50: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:21: Starting minibatch loop.
01/17/2018 06:13:21:  Epoch[50 of 50]-Minibatch[   1-  10, 14.29%]: CrossEntropyWithSoftmax = 0.14391767 * 1280; EvalClassificationError = 0.07656250 * 1280; time = 0.0059s; samplesPerSecond = 216747.1
01/17/2018 06:13:21:  Epoch[50 of 50]-Minibatch[  11-  20, 28.57%]: CrossEntropyWithSoftmax = 0.13435644 * 1280; EvalClassificationError = 0.07578125 * 1280; time = 0.0053s; samplesPerSecond = 242704.6
01/17/2018 06:13:21:  Epoch[50 of 50]-Minibatch[  21-  30, 42.86%]: CrossEntropyWithSoftmax = 0.13637500 * 1280; EvalClassificationError = 0.07500000 * 1280; time = 0.0057s; samplesPerSecond = 224687.5
01/17/2018 06:13:21:  Epoch[50 of 50]-Minibatch[  31-  40, 57.14%]: CrossEntropyWithSoftmax = 0.14619503 * 1280; EvalClassificationError = 0.07656250 * 1280; time = 0.0057s; samplesPerSecond = 224058.3
01/17/2018 06:13:21:  Epoch[50 of 50]-Minibatch[  41-  50, 71.43%]: CrossEntropyWithSoftmax = 0.13345990 * 1280; EvalClassificationError = 0.06640625 * 1280; time = 0.0056s; samplesPerSecond = 230381.6
01/17/2018 06:13:21:  Epoch[50 of 50]-Minibatch[  51-  60, 85.71%]: CrossEntropyWithSoftmax = 0.14153485 * 1280; EvalClassificationError = 0.07343750 * 1280; time = 0.0056s; samplesPerSecond = 227055.0
01/17/2018 06:13:21:  Epoch[50 of 50]-Minibatch[  61-  70, 100.00%]: CrossEntropyWithSoftmax = 0.14213562 * 1280; EvalClassificationError = 0.06875000 * 1280; time = 0.0054s; samplesPerSecond = 235091.0
01/17/2018 06:13:21: Finished Epoch[50 of 50]: [Training] CrossEntropyWithSoftmax = 0.13897399 * 10000; EvalClassificationError = 0.07310000 * 10000; totalSamplesSeen = 500000; learningRatePerSample = 0.1; epochTime=0.0455668s
01/17/2018 06:13:21: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn'

01/17/2018 06:13:21: Action "train" complete.


01/17/2018 06:13:21: ##############################################################################
01/17/2018 06:13:21: #                                                                            #
01/17/2018 06:13:21: # Simple_Demo_Output command (write action)                                  #
01/17/2018 06:13:21: #                                                                            #
01/17/2018 06:13:21: ##############################################################################


Post-processing network...

7 roots:
	CrossEntropyWithSoftmax = CrossEntropyWithSoftmax()
	EvalClassificationError = ClassificationError()
	InvStdOfFeatures = InvStdDev()
	MeanOfFeatures = Mean()
	PosteriorProb = Softmax()
	Prior = Mean()
	ScaledLogLikelihood = Minus()

Validating network. 25 nodes to process in pass 1.

Validating --> labels = InputValue() :  -> [2 x *1]
Validating --> W2 = LearnableParameter() :  -> [2 x 50]
Validating --> W1 = LearnableParameter() :  -> [50 x 50]
Validating --> W0 = LearnableParameter() :  -> [50 x 2]
Validating --> features = InputValue() :  -> [2 x *1]
Validating --> MeanOfFeatures = Mean (features) : [2 x *1] -> [2]
Validating --> InvStdOfFeatures = InvStdDev (features) : [2 x *1] -> [2]
Validating --> MVNormalizedFeatures = PerDimMeanVarNormalization (features, MeanOfFeatures, InvStdOfFeatures) : [2 x *1], [2], [2] -> [2 x *1]
Validating --> W0*features = Times (W0, MVNormalizedFeatures) : [50 x 2], [2 x *1] -> [50 x *1]
Validating --> B0 = LearnableParameter() :  -> [50 x 1]
Validating --> W0*features+B0 = Plus (W0*features, B0) : [50 x *1], [50 x 1] -> [50 x 1 x *1]
Validating --> H1 = Sigmoid (W0*features+B0) : [50 x 1 x *1] -> [50 x 1 x *1]
Validating --> W1*H1 = Times (W1, H1) : [50 x 50], [50 x 1 x *1] -> [50 x 1 x *1]
Validating --> B1 = LearnableParameter() :  -> [50 x 1]
Validating --> W1*H1+B1 = Plus (W1*H1, B1) : [50 x 1 x *1], [50 x 1] -> [50 x 1 x *1]
Validating --> H2 = Sigmoid (W1*H1+B1) : [50 x 1 x *1] -> [50 x 1 x *1]
Validating --> W2*H1 = Times (W2, H2) : [2 x 50], [50 x 1 x *1] -> [2 x 1 x *1]
Validating --> B2 = LearnableParameter() :  -> [2 x 1]
Validating --> HLast = Plus (W2*H1, B2) : [2 x 1 x *1], [2 x 1] -> [2 x 1 x *1]
Validating --> CrossEntropyWithSoftmax = CrossEntropyWithSoftmax (labels, HLast) : [2 x *1], [2 x 1 x *1] -> [1]
Validating --> EvalClassificationError = ClassificationError (labels, HLast) : [2 x *1], [2 x 1 x *1] -> [1]
Validating --> PosteriorProb = Softmax (HLast) : [2 x 1 x *1] -> [2 x 1 x *1]
Validating --> Prior = Mean (labels) : [2 x *1] -> [2]
Validating --> LogOfPrior = Log (Prior) : [2] -> [2]
Validating --> ScaledLogLikelihood = Minus (HLast, LogOfPrior) : [2 x 1 x *1], [2] -> [2 x 1 x *1]

Validating network. 17 nodes to process in pass 2.


Validating network, final pass.




Post-processing network complete.



Allocating matrices for forward and/or backward propagation.

Memory Sharing: Out of 25 matrices, 12 are shared as 3, and 13 are not shared.

Here are the ones that share memory:
	{ CrossEntropyWithSoftmax : [1]
	  EvalClassificationError : [1]
	  PosteriorProb : [2 x 1 x *1] }
	{ H1 : [50 x 1 x *1]
	  W0*features : [50 x *1]
	  W1*H1+B1 : [50 x 1 x *1]
	  W2*H1 : [2 x 1 x *1] }
	{ H2 : [50 x 1 x *1]
	  HLast : [2 x 1 x *1]
	  MVNormalizedFeatures : [2 x *1]
	  W0*features+B0 : [50 x 1 x *1]
	  W1*H1 : [50 x 1 x *1] }

Here are the ones that don't share memory:
	{ScaledLogLikelihood : [2 x 1 x *1]}
	{B0 : [50 x 1]}
	{B1 : [50 x 1]}
	{B2 : [2 x 1]}
	{features : [2 x *1]}
	{InvStdOfFeatures : [2]}
	{labels : [2 x *1]}
	{MeanOfFeatures : [2]}
	{Prior : [2]}
	{W0 : [50 x 2]}
	{W1 : [50 x 50]}
	{W2 : [2 x 50]}
	{LogOfPrior : [2]}

Minibatch[0]: ActualMBSize = 603
Written to /tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/SimpleOutput*
Total Samples Evaluated = 603

01/17/2018 06:13:21: Action "write" complete.

01/17/2018 06:13:21: __COMPLETED__
=== Deleting last epoch data
==== Re-running from checkpoint
=== Running /home/ubuntu/workspace/build/gpu/release/bin/cntk configFile=/home/ubuntu/workspace/Tests/EndToEndTests/Speech/Simple/cntk.cntk currentDirectory=/home/ubuntu/workspace/Tests/EndToEndTests/Speech/Data RunDir=/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu DataDir=/home/ubuntu/workspace/Tests/EndToEndTests/Speech/Data ConfigDir=/home/ubuntu/workspace/Tests/EndToEndTests/Speech/Simple OutputDir=/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu DeviceId=0 timestamping=true forceDeterministicAlgorithms=true makeMode=true
CNTK 2.3.1+ (HEAD b7b3e4, Jan 17 2018 02:42:45) at 2018/01/17 06:13:21

/home/ubuntu/workspace/build/gpu/release/bin/cntk  configFile=/home/ubuntu/workspace/Tests/EndToEndTests/Speech/Simple/cntk.cntk  currentDirectory=/home/ubuntu/workspace/Tests/EndToEndTests/Speech/Data  RunDir=/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu  DataDir=/home/ubuntu/workspace/Tests/EndToEndTests/Speech/Data  ConfigDir=/home/ubuntu/workspace/Tests/EndToEndTests/Speech/Simple  OutputDir=/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu  DeviceId=0  timestamping=true  forceDeterministicAlgorithms=true  makeMode=true
Changed current directory to /home/ubuntu/workspace/Tests/EndToEndTests/Speech/Data
01/17/2018 06:13:21: -------------------------------------------------------------------
01/17/2018 06:13:21: Build info: 

01/17/2018 06:13:21: 		Built time: Jan 17 2018 02:36:21
01/17/2018 06:13:21: 		Last modified date: Wed Jan 17 02:34:37 2018
01/17/2018 06:13:21: 		Build type: release
01/17/2018 06:13:21: 		Build target: GPU
01/17/2018 06:13:21: 		With ASGD: yes
01/17/2018 06:13:21: 		Math lib: mkl
01/17/2018 06:13:21: 		CUDA version: 9.0.0
01/17/2018 06:13:21: 		CUDNN version: 7.0.4
01/17/2018 06:13:21: 		Build Branch: HEAD
01/17/2018 06:13:21: 		Build SHA1: b7b3e4fb3ff0f69024ce19a19b8f2780fb63078b
01/17/2018 06:13:21: 		MPI distribution: Open MPI
01/17/2018 06:13:21: 		MPI version: 1.10.7
01/17/2018 06:13:21: -------------------------------------------------------------------
01/17/2018 06:13:21: -------------------------------------------------------------------
01/17/2018 06:13:21: GPU info:

01/17/2018 06:13:21: 		Device[0]: cores = 3072; computeCapability = 5.2; type = "Tesla M60"; total memory = 8123 MB; free memory = 8112 MB
01/17/2018 06:13:21: -------------------------------------------------------------------

Configuration After Processing and Variable Resolution:

configparameters: cntk.cntk:command=Simple_Demo:Simple_Demo_Output
configparameters: cntk.cntk:ConfigDir=/home/ubuntu/workspace/Tests/EndToEndTests/Speech/Simple
configparameters: cntk.cntk:currentDirectory=/home/ubuntu/workspace/Tests/EndToEndTests/Speech/Data
configparameters: cntk.cntk:DataDir=/home/ubuntu/workspace/Tests/EndToEndTests/Speech/Data
configparameters: cntk.cntk:deviceId=0
configparameters: cntk.cntk:DeviceNumber=-1
configparameters: cntk.cntk:forceDeterministicAlgorithms=true
configparameters: cntk.cntk:makeMode=true
configparameters: cntk.cntk:modelPath=/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn
configparameters: cntk.cntk:OutputDir=/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu
configparameters: cntk.cntk:outputNodeNames=ScaledLogLikelihood
configparameters: cntk.cntk:precision=float
configparameters: cntk.cntk:RunDir=/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu
configparameters: cntk.cntk:Simple_Demo=[
    action=train
    SimpleNetworkBuilder=[
        layerSizes=2:50*2:2
        trainingCriterion=CrossEntropyWithSoftmax
        evalCriterion=ClassificationError
        layerTypes=Sigmoid
        initValueScale=1.0
        applyMeanVarNorm=true
        uniformInit=true
        needPrior=true
    ]
    SGD=[
        epochSize=0 
        minibatchSize=128
        learningRatesPerSample=0.1
        momentumAsTimeConstant=2500
        dropoutRate=0.0
        maxEpochs=50
        keepCheckPointFiles = true
    ]
    reader=[
        readerType=CNTKTextFormatReader
        file=/home/ubuntu/workspace/Tests/EndToEndTests/Speech/Data/SimpleDataTrain_cntk_text.txt
        input = [
            features=[
dim = 2      
                format = "dense"
            ]
            labels=[
dim = 2 
                format = "dense"
            ]
        ]
    ]
]

configparameters: cntk.cntk:Simple_Demo_Output=[
    action=write
    reader=[
        readerType=CNTKTextFormatReader
        file=/home/ubuntu/workspace/Tests/EndToEndTests/Speech/Data/SimpleDataTest_cntk_text.txt
        input = [
            features=[
dim = 2 
                format = "dense" 
            ]
            labels=[
dim = 2 
                format = "dense"
            ]
        ]
    ]
outputPath=/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/SimpleOutput    
]

configparameters: cntk.cntk:timestamping=true
configparameters: cntk.cntk:traceLevel=1
01/17/2018 06:13:21: Commands: Simple_Demo Simple_Demo_Output
01/17/2018 06:13:21: precision = "float"
01/17/2018 06:13:21: WARNING: forceDeterministicAlgorithms flag is specified. Using 1 CPU thread for processing.

01/17/2018 06:13:21: ##############################################################################
01/17/2018 06:13:21: #                                                                            #
01/17/2018 06:13:21: # Simple_Demo command (train action)                                         #
01/17/2018 06:13:21: #                                                                            #
01/17/2018 06:13:21: ##############################################################################

01/17/2018 06:13:21: 
Starting from checkpoint. Loading network from '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn.49'.
SimpleNetworkBuilder Using GPU 0
01/17/2018 06:13:21: 
Model has 25 nodes. Using GPU 0.

01/17/2018 06:13:21: Training criterion:   CrossEntropyWithSoftmax = CrossEntropyWithSoftmax
01/17/2018 06:13:21: Evaluation criterion: EvalClassificationError = ClassificationError

01/17/2018 06:13:21: Training 2802 parameters in 6 out of 6 parameter tensors and 15 nodes with gradient:

01/17/2018 06:13:21: 	Node 'B0' (LearnableParameter operation) : [50 x 1]
01/17/2018 06:13:21: 	Node 'B1' (LearnableParameter operation) : [50 x 1]
01/17/2018 06:13:21: 	Node 'B2' (LearnableParameter operation) : [2 x 1]
01/17/2018 06:13:21: 	Node 'W0' (LearnableParameter operation) : [50 x 2]
01/17/2018 06:13:21: 	Node 'W1' (LearnableParameter operation) : [50 x 50]
01/17/2018 06:13:21: 	Node 'W2' (LearnableParameter operation) : [2 x 50]

01/17/2018 06:13:21: No PreCompute nodes found, or all already computed. Skipping pre-computation step.

01/17/2018 06:13:21: Starting Epoch 50: learning rate per sample = 0.100000  effective momentum = 0.950085  momentum as time constant = 2499.8 samples

01/17/2018 06:13:21: Starting minibatch loop.
01/17/2018 06:13:22:  Epoch[50 of 50]-Minibatch[   1-  10]: CrossEntropyWithSoftmax = 0.14391767 * 1280; EvalClassificationError = 0.07656250 * 1280; time = 0.2976s; samplesPerSecond = 4300.7
01/17/2018 06:13:22:  Epoch[50 of 50]-Minibatch[  11-  20]: CrossEntropyWithSoftmax = 0.13435644 * 1280; EvalClassificationError = 0.07578125 * 1280; time = 0.0053s; samplesPerSecond = 240778.0
01/17/2018 06:13:22:  Epoch[50 of 50]-Minibatch[  21-  30]: CrossEntropyWithSoftmax = 0.13637500 * 1280; EvalClassificationError = 0.07500000 * 1280; time = 0.0052s; samplesPerSecond = 247999.5
01/17/2018 06:13:22:  Epoch[50 of 50]-Minibatch[  31-  40]: CrossEntropyWithSoftmax = 0.14619503 * 1280; EvalClassificationError = 0.07656250 * 1280; time = 0.0056s; samplesPerSecond = 229757.1
01/17/2018 06:13:22:  Epoch[50 of 50]-Minibatch[  41-  50]: CrossEntropyWithSoftmax = 0.13345990 * 1280; EvalClassificationError = 0.06640625 * 1280; time = 0.0052s; samplesPerSecond = 247047.0
01/17/2018 06:13:22:  Epoch[50 of 50]-Minibatch[  51-  60]: CrossEntropyWithSoftmax = 0.14153485 * 1280; EvalClassificationError = 0.07343750 * 1280; time = 0.0052s; samplesPerSecond = 244788.7
01/17/2018 06:13:22:  Epoch[50 of 50]-Minibatch[  61-  70]: CrossEntropyWithSoftmax = 0.14213562 * 1280; EvalClassificationError = 0.06875000 * 1280; time = 0.0056s; samplesPerSecond = 229386.6
01/17/2018 06:13:22: Finished Epoch[50 of 50]: [Training] CrossEntropyWithSoftmax = 0.13897399 * 10000; EvalClassificationError = 0.07310000 * 10000; totalSamplesSeen = 500000; learningRatePerSample = 0.1; epochTime=0.336333s
01/17/2018 06:13:22: SGD: Saving checkpoint model '/tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/models/simple.dnn'

01/17/2018 06:13:22: Action "train" complete.


01/17/2018 06:13:22: ##############################################################################
01/17/2018 06:13:22: #                                                                            #
01/17/2018 06:13:22: # Simple_Demo_Output command (write action)                                  #
01/17/2018 06:13:22: #                                                                            #
01/17/2018 06:13:22: ##############################################################################


Post-processing network...

7 roots:
	CrossEntropyWithSoftmax = CrossEntropyWithSoftmax()
	EvalClassificationError = ClassificationError()
	InvStdOfFeatures = InvStdDev()
	MeanOfFeatures = Mean()
	PosteriorProb = Softmax()
	Prior = Mean()
	ScaledLogLikelihood = Minus()

Validating network. 25 nodes to process in pass 1.

Validating --> labels = InputValue() :  -> [2 x *2]
Validating --> W2 = LearnableParameter() :  -> [2 x 50]
Validating --> W1 = LearnableParameter() :  -> [50 x 50]
Validating --> W0 = LearnableParameter() :  -> [50 x 2]
Validating --> features = InputValue() :  -> [2 x *2]
Validating --> MeanOfFeatures = Mean (features) : [2 x *2] -> [2]
Validating --> InvStdOfFeatures = InvStdDev (features) : [2 x *2] -> [2]
Validating --> MVNormalizedFeatures = PerDimMeanVarNormalization (features, MeanOfFeatures, InvStdOfFeatures) : [2 x *2], [2], [2] -> [2 x *2]
Validating --> W0*features = Times (W0, MVNormalizedFeatures) : [50 x 2], [2 x *2] -> [50 x *2]
Validating --> B0 = LearnableParameter() :  -> [50 x 1]
Validating --> W0*features+B0 = Plus (W0*features, B0) : [50 x *2], [50 x 1] -> [50 x 1 x *2]
Validating --> H1 = Sigmoid (W0*features+B0) : [50 x 1 x *2] -> [50 x 1 x *2]
Validating --> W1*H1 = Times (W1, H1) : [50 x 50], [50 x 1 x *2] -> [50 x 1 x *2]
Validating --> B1 = LearnableParameter() :  -> [50 x 1]
Validating --> W1*H1+B1 = Plus (W1*H1, B1) : [50 x 1 x *2], [50 x 1] -> [50 x 1 x *2]
Validating --> H2 = Sigmoid (W1*H1+B1) : [50 x 1 x *2] -> [50 x 1 x *2]
Validating --> W2*H1 = Times (W2, H2) : [2 x 50], [50 x 1 x *2] -> [2 x 1 x *2]
Validating --> B2 = LearnableParameter() :  -> [2 x 1]
Validating --> HLast = Plus (W2*H1, B2) : [2 x 1 x *2], [2 x 1] -> [2 x 1 x *2]
Validating --> CrossEntropyWithSoftmax = CrossEntropyWithSoftmax (labels, HLast) : [2 x *2], [2 x 1 x *2] -> [1]
Validating --> EvalClassificationError = ClassificationError (labels, HLast) : [2 x *2], [2 x 1 x *2] -> [1]
Validating --> PosteriorProb = Softmax (HLast) : [2 x 1 x *2] -> [2 x 1 x *2]
Validating --> Prior = Mean (labels) : [2 x *2] -> [2]
Validating --> LogOfPrior = Log (Prior) : [2] -> [2]
Validating --> ScaledLogLikelihood = Minus (HLast, LogOfPrior) : [2 x 1 x *2], [2] -> [2 x 1 x *2]

Validating network. 17 nodes to process in pass 2.


Validating network, final pass.




Post-processing network complete.



Allocating matrices for forward and/or backward propagation.

Memory Sharing: Out of 25 matrices, 12 are shared as 3, and 13 are not shared.

Here are the ones that share memory:
	{ CrossEntropyWithSoftmax : [1]
	  EvalClassificationError : [1]
	  PosteriorProb : [2 x 1 x *2] }
	{ H1 : [50 x 1 x *2]
	  W0*features : [50 x *2]
	  W1*H1+B1 : [50 x 1 x *2]
	  W2*H1 : [2 x 1 x *2] }
	{ H2 : [50 x 1 x *2]
	  HLast : [2 x 1 x *2]
	  MVNormalizedFeatures : [2 x *2]
	  W0*features+B0 : [50 x 1 x *2]
	  W1*H1 : [50 x 1 x *2] }

Here are the ones that don't share memory:
	{B2 : [2 x 1]}
	{features : [2 x *2]}
	{InvStdOfFeatures : [2]}
	{labels : [2 x *2]}
	{MeanOfFeatures : [2]}
	{Prior : [2]}
	{W0 : [50 x 2]}
	{W1 : [50 x 50]}
	{W2 : [2 x 50]}
	{LogOfPrior : [2]}
	{B0 : [50 x 1]}
	{B1 : [50 x 1]}
	{ScaledLogLikelihood : [2 x 1 x *2]}

Minibatch[0]: ActualMBSize = 603
Written to /tmp/cntk-test-20180117061317.742222/Speech_Simple@release_gpu/SimpleOutput*
Total Samples Evaluated = 603

01/17/2018 06:13:22: Action "write" complete.

01/17/2018 06:13:22: __COMPLETED__