CPU info:
    CPU Model Name: Intel(R) Xeon(R) CPU E5-2630 v2 @ 2.60GHz
    Hardware threads: 24
    Total Memory: 264172964 kB
-------------------------------------------------------------------
=== Running /home/philly/jenkins/workspace/CNTK-Test-Linux-W1/build/gpu/release/bin/cntk configFile=/home/philly/jenkins/workspace/CNTK-Test-Linux-W1/Tests/EndToEndTests/Speech/HTKDeserializers/TIMIT/TrainWithPreTrain/../../../../../../Examples/Speech/Miscellaneous/TIMIT/config/TIMIT_TrainWithPreTrain_ndl_deprecated.cntk currentDirectory=/home/philly/data/CNTKTestData/Speech/ASR RunDir=/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu DataDir=/home/philly/data/CNTKTestData/Speech/ASR ConfigDir=/home/philly/jenkins/workspace/CNTK-Test-Linux-W1/Tests/EndToEndTests/Speech/HTKDeserializers/TIMIT/TrainWithPreTrain/../../../../../../Examples/Speech/Miscellaneous/TIMIT/config OutputDir=/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu DeviceId=0 timestamping=true reader=[readerType=HTKDeserializers] LibDir=/home/philly/jenkins/workspace/CNTK-Test-Linux-W1/Tests/EndToEndTests/Speech/HTKDeserializers/TIMIT/TrainWithPreTrain/../../../../../../Examples/Speech/Miscellaneous/TIMIT/config/../lib ScpDir=/home/philly/data/CNTKTestData/Speech/ASR MlfDir=/home/philly/data/CNTKTestData/Speech/ASR NdlDir=/home/philly/jenkins/workspace/CNTK-Test-Linux-W1/Tests/EndToEndTests/Speech/HTKDeserializers/TIMIT/TrainWithPreTrain/../../../../../../Examples/Speech/Miscellaneous/TIMIT/config MelDir=/home/philly/jenkins/workspace/CNTK-Test-Linux-W1/Tests/EndToEndTests/Speech/HTKDeserializers/TIMIT/TrainWithPreTrain/../../../../../../Examples/Speech/Miscellaneous/TIMIT/config ExpDir=/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp DeviceNumber=0
CNTK 2.0.beta6.0+ (HEAD bf0ca9, Dec 20 2016 11:40:12) on localhost at 2016/12/20 15:26:37

/home/philly/jenkins/workspace/CNTK-Test-Linux-W1/build/gpu/release/bin/cntk  configFile=/home/philly/jenkins/workspace/CNTK-Test-Linux-W1/Tests/EndToEndTests/Speech/HTKDeserializers/TIMIT/TrainWithPreTrain/../../../../../../Examples/Speech/Miscellaneous/TIMIT/config/TIMIT_TrainWithPreTrain_ndl_deprecated.cntk  currentDirectory=/home/philly/data/CNTKTestData/Speech/ASR  RunDir=/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu  DataDir=/home/philly/data/CNTKTestData/Speech/ASR  ConfigDir=/home/philly/jenkins/workspace/CNTK-Test-Linux-W1/Tests/EndToEndTests/Speech/HTKDeserializers/TIMIT/TrainWithPreTrain/../../../../../../Examples/Speech/Miscellaneous/TIMIT/config  OutputDir=/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu  DeviceId=0  timestamping=true  reader=[readerType=HTKDeserializers]  LibDir=/home/philly/jenkins/workspace/CNTK-Test-Linux-W1/Tests/EndToEndTests/Speech/HTKDeserializers/TIMIT/TrainWithPreTrain/../../../../../../Examples/Speech/Miscellaneous/TIMIT/config/../lib  ScpDir=/home/philly/data/CNTKTestData/Speech/ASR  MlfDir=/home/philly/data/CNTKTestData/Speech/ASR  NdlDir=/home/philly/jenkins/workspace/CNTK-Test-Linux-W1/Tests/EndToEndTests/Speech/HTKDeserializers/TIMIT/TrainWithPreTrain/../../../../../../Examples/Speech/Miscellaneous/TIMIT/config  MelDir=/home/philly/jenkins/workspace/CNTK-Test-Linux-W1/Tests/EndToEndTests/Speech/HTKDeserializers/TIMIT/TrainWithPreTrain/../../../../../../Examples/Speech/Miscellaneous/TIMIT/config  ExpDir=/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp  DeviceNumber=0
Changed current directory to /home/philly/data/CNTKTestData/Speech/ASR
12/20/2016 15:26:38: -------------------------------------------------------------------
12/20/2016 15:26:38: Build info: 

12/20/2016 15:26:38: 		Built time: Dec 20 2016 11:40:12
12/20/2016 15:26:38: 		Last modified date: Tue Dec 20 11:38:27 2016
12/20/2016 15:26:38: 		Build type: release
12/20/2016 15:26:38: 		Build target: GPU
12/20/2016 15:26:38: 		With ASGD: yes
12/20/2016 15:26:38: 		Math lib: mkl
12/20/2016 15:26:38: 		CUDA_PATH: /usr/local/cuda-8.0
12/20/2016 15:26:38: 		CUB_PATH: /usr/local/cub-1.4.1
12/20/2016 15:26:38: 		CUDNN_PATH: /usr/local
12/20/2016 15:26:38: 		Build Branch: HEAD
12/20/2016 15:26:38: 		Build SHA1: bf0ca998cd077aa28c04371fd2093770e819ffd0
12/20/2016 15:26:38: 		Built by Source/CNTK/buildinfo.h$$0 on b4b39bc07965
12/20/2016 15:26:38: 		Build Path: /home/philly/jenkins/workspace/CNTK-Build-Linux
12/20/2016 15:26:38: -------------------------------------------------------------------
12/20/2016 15:26:39: -------------------------------------------------------------------
12/20/2016 15:26:39: GPU info:

12/20/2016 15:26:39: 		Device[0]: cores = 2880; computeCapability = 3.5; type = "GeForce GTX 780 Ti"; memory = 3020 MB
12/20/2016 15:26:39: 		Device[1]: cores = 2880; computeCapability = 3.5; type = "GeForce GTX 780 Ti"; memory = 3020 MB
12/20/2016 15:26:39: 		Device[2]: cores = 2880; computeCapability = 3.5; type = "GeForce GTX 780 Ti"; memory = 3020 MB
12/20/2016 15:26:39: 		Device[3]: cores = 2880; computeCapability = 3.5; type = "GeForce GTX 780 Ti"; memory = 3020 MB
12/20/2016 15:26:39: -------------------------------------------------------------------

Configuration After Processing and Variable Resolution:

configparameters: TIMIT_TrainWithPreTrain_ndl_deprecated.cntk:command=TIMIT_DiscrimPreTrain1:TIMIT_AddLayer2:TIMIT_DiscrimPreTrain2:TIMIT_AddLayer3:TIMIT_Train3
configparameters: TIMIT_TrainWithPreTrain_ndl_deprecated.cntk:ConfigDir=/home/philly/jenkins/workspace/CNTK-Test-Linux-W1/Tests/EndToEndTests/Speech/HTKDeserializers/TIMIT/TrainWithPreTrain/../../../../../../Examples/Speech/Miscellaneous/TIMIT/config
configparameters: TIMIT_TrainWithPreTrain_ndl_deprecated.cntk:currentDirectory=/home/philly/data/CNTKTestData/Speech/ASR
configparameters: TIMIT_TrainWithPreTrain_ndl_deprecated.cntk:DataDir=/home/philly/data/CNTKTestData/Speech/ASR
configparameters: TIMIT_TrainWithPreTrain_ndl_deprecated.cntk:deviceId=0
configparameters: TIMIT_TrainWithPreTrain_ndl_deprecated.cntk:DeviceNumber=0
configparameters: TIMIT_TrainWithPreTrain_ndl_deprecated.cntk:ExpDir=/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp
configparameters: TIMIT_TrainWithPreTrain_ndl_deprecated.cntk:initOnCPUOnly=true
configparameters: TIMIT_TrainWithPreTrain_ndl_deprecated.cntk:LibDir=/home/philly/jenkins/workspace/CNTK-Test-Linux-W1/Tests/EndToEndTests/Speech/HTKDeserializers/TIMIT/TrainWithPreTrain/../../../../../../Examples/Speech/Miscellaneous/TIMIT/config/../lib
configparameters: TIMIT_TrainWithPreTrain_ndl_deprecated.cntk:MelDir=/home/philly/jenkins/workspace/CNTK-Test-Linux-W1/Tests/EndToEndTests/Speech/HTKDeserializers/TIMIT/TrainWithPreTrain/../../../../../../Examples/Speech/Miscellaneous/TIMIT/config
configparameters: TIMIT_TrainWithPreTrain_ndl_deprecated.cntk:MlfDir=/home/philly/data/CNTKTestData/Speech/ASR
configparameters: TIMIT_TrainWithPreTrain_ndl_deprecated.cntk:NdlDir=/home/philly/jenkins/workspace/CNTK-Test-Linux-W1/Tests/EndToEndTests/Speech/HTKDeserializers/TIMIT/TrainWithPreTrain/../../../../../../Examples/Speech/Miscellaneous/TIMIT/config
configparameters: TIMIT_TrainWithPreTrain_ndl_deprecated.cntk:ndlMacros=/home/philly/jenkins/workspace/CNTK-Test-Linux-W1/Tests/EndToEndTests/Speech/HTKDeserializers/TIMIT/TrainWithPreTrain/../../../../../../Examples/Speech/Miscellaneous/TIMIT/config/default_macros.ndl
configparameters: TIMIT_TrainWithPreTrain_ndl_deprecated.cntk:OutputDir=/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu
configparameters: TIMIT_TrainWithPreTrain_ndl_deprecated.cntk:precision=float
configparameters: TIMIT_TrainWithPreTrain_ndl_deprecated.cntk:reader=[
    readerType=HTKMLFReader
    readMethod=blockRandomize
    miniBatchMode=Partial
    randomize=Auto
    verbosity=0
    features=[
        dim=792
        scpFile=/home/philly/data/CNTKTestData/Speech/ASR/TIMIT.train.scp.fbank.fullpath.rnn
    ]
    labels=[
        mlfFile=/home/philly/data/CNTKTestData/Speech/ASR/TIMIT.train.align_cistate.mlf.cntk
        labelDim=183
        labelMappingFile=/home/philly/data/CNTKTestData/Speech/ASR/TIMIT.statelist
    ]
] [readerType=HTKDeserializers]

configparameters: TIMIT_TrainWithPreTrain_ndl_deprecated.cntk:RunDir=/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu
configparameters: TIMIT_TrainWithPreTrain_ndl_deprecated.cntk:ScpDir=/home/philly/data/CNTKTestData/Speech/ASR
configparameters: TIMIT_TrainWithPreTrain_ndl_deprecated.cntk:SGD=[
    epochSize=0 
    minibatchSize=256
    learningRatesPerMB=0.1
    momentumPerMB=0.9
    dropoutRate=0.0
    maxEpochs=2
]

configparameters: TIMIT_TrainWithPreTrain_ndl_deprecated.cntk:timestamping=true
configparameters: TIMIT_TrainWithPreTrain_ndl_deprecated.cntk:TIMIT_AddLayer2=[    
    action=edit
    CurrLayer=1
    NewLayer=2
    CurrModel=/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/dptmodel1/cntkSpeech.dnn
    NewModel=/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/dptmodel2/cntkSpeech.dnn.0
    editPath=/home/philly/jenkins/workspace/CNTK-Test-Linux-W1/Tests/EndToEndTests/Speech/HTKDeserializers/TIMIT/TrainWithPreTrain/../../../../../../Examples/Speech/Miscellaneous/TIMIT/config/add_layer.mel
]

configparameters: TIMIT_TrainWithPreTrain_ndl_deprecated.cntk:TIMIT_AddLayer3=[
    action=edit
    CurrLayer=2
    NewLayer=3
    CurrModel=/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/dptmodel2/cntkSpeech.dnn
    NewModel=/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn.0
    editPath=/home/philly/jenkins/workspace/CNTK-Test-Linux-W1/Tests/EndToEndTests/Speech/HTKDeserializers/TIMIT/TrainWithPreTrain/../../../../../../Examples/Speech/Miscellaneous/TIMIT/config/add_layer.mel
]

configparameters: TIMIT_TrainWithPreTrain_ndl_deprecated.cntk:TIMIT_DiscrimPreTrain1=[
    action=train    
    modelPath=/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/dptmodel1/cntkSpeech.dnn
    NDLNetworkBuilder=[
        NetworkDescription=/home/philly/jenkins/workspace/CNTK-Test-Linux-W1/Tests/EndToEndTests/Speech/HTKDeserializers/TIMIT/TrainWithPreTrain/../../../../../../Examples/Speech/Miscellaneous/TIMIT/config/create_1layer.ndl
    ]
]

configparameters: TIMIT_TrainWithPreTrain_ndl_deprecated.cntk:TIMIT_DiscrimPreTrain2=[
    action=train
    modelPath=/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/dptmodel2/cntkSpeech.dnn
    NDLNetworkBuilder=[
        NetworkDescription=/home/philly/jenkins/workspace/CNTK-Test-Linux-W1/Tests/EndToEndTests/Speech/HTKDeserializers/TIMIT/TrainWithPreTrain/../../../../../../Examples/Speech/Miscellaneous/TIMIT/config/create_1layer.ndl
    ]
]

configparameters: TIMIT_TrainWithPreTrain_ndl_deprecated.cntk:TIMIT_Train3=[
    action=train
    modelPath=/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn
    NDLNetworkBuilder=[
        NetworkDescription=/home/philly/jenkins/workspace/CNTK-Test-Linux-W1/Tests/EndToEndTests/Speech/HTKDeserializers/TIMIT/TrainWithPreTrain/../../../../../../Examples/Speech/Miscellaneous/TIMIT/config/create_1layer.ndl
    ]
    SGD=[
        epochSize=0 
        minibatchSize=256:1024
        learningRatesPerMB=0.8:3.2*14:0.08
        momentumPerMB=0.9
        dropoutRate=0.0
        maxEpochs=25
    ]  
]

configparameters: TIMIT_TrainWithPreTrain_ndl_deprecated.cntk:traceLevel=1
12/20/2016 15:26:39: Commands: TIMIT_DiscrimPreTrain1 TIMIT_AddLayer2 TIMIT_DiscrimPreTrain2 TIMIT_AddLayer3 TIMIT_Train3
12/20/2016 15:26:39: precision = "float"

12/20/2016 15:26:39: ##############################################################################
12/20/2016 15:26:39: #                                                                            #
12/20/2016 15:26:39: # TIMIT_DiscrimPreTrain1 command (train action)                              #
12/20/2016 15:26:39: #                                                                            #
12/20/2016 15:26:39: ##############################################################################

12/20/2016 15:26:39: 
Creating virgin network.
NDLBuilder Using GPU 0
SetUniformRandomValue (GPU): creating curand object with seed 1, sizeof(ElemType)==4
Reading script file /home/philly/data/CNTKTestData/Speech/ASR/TIMIT.train.scp.fbank.fullpath.rnn ... 3696 entries
HTKDataDeserializer::HTKDataDeserializer: selected 3696 utterances grouped into 13 chunks, average chunk size: 284.3 utterances, 86524.8 frames (for I/O: 284.3 utterances, 86524.8 frames)
HTKDataDeserializer::HTKDataDeserializer: determined feature kind as 72-dimensional 'FBANK_D_A_Z' with frame shift 10.0 ms
total 183 state names in state list /home/philly/data/CNTKTestData/Speech/ASR/TIMIT.statelist
htkmlfreader: reading MLF file /home/philly/data/CNTKTestData/Speech/ASR/TIMIT.train.align_cistate.mlf.cntk ... total 3696 entries
MLFDataDeserializer::MLFDataDeserializer: 3696 utterances with 1124823 frames in 183 classes
12/20/2016 15:26:40: 
Model has 19 nodes. Using GPU 0.

12/20/2016 15:26:40: Training criterion:   CE.SM = CrossEntropyWithSoftmax
12/20/2016 15:26:40: Evaluation criterion: Err = ClassificationError


Allocating matrices for forward and/or backward propagation.

Memory Sharing: Out of 29 matrices, 11 are shared as 5, and 18 are not shared.

	{ L1.BFF.FF.P : [512 x *]
	  L1.BFF.W : [512 x 792] (gradient) }
	{ L1.BFF.FF.T : [512 x *] (gradient)
	  L1.S : [512 x *] }
	{ CE.BFF.FF.T : [183 x *]
	  L1.BFF.FF.P : [512 x *] (gradient) }
	{ CE.BFF.FF.P : [183 x *] (gradient)
	  L1.BFF.B : [512] (gradient)
	  L1.S : [512 x *] (gradient) }
	{ CE.BFF.FF.P : [183 x *]
	  CE.BFF.W : [183 x 512] (gradient) }


12/20/2016 15:26:40: Training 499895 parameters in 4 out of 4 parameter tensors and 10 nodes with gradient:

12/20/2016 15:26:40: 	Node 'CE.BFF.B' (LearnableParameter operation) : [183]
12/20/2016 15:26:40: 	Node 'CE.BFF.W' (LearnableParameter operation) : [183 x 512]
12/20/2016 15:26:40: 	Node 'L1.BFF.B' (LearnableParameter operation) : [512]
12/20/2016 15:26:40: 	Node 'L1.BFF.W' (LearnableParameter operation) : [512 x 792]


12/20/2016 15:26:40: Precomputing --> 3 PreCompute nodes found.

12/20/2016 15:26:40: 	featNorm.xMean = Mean()
12/20/2016 15:26:40: 	featNorm.xStdDev = InvStdDev()
12/20/2016 15:26:40: 	logPrior.Prior = Mean()

12/20/2016 15:26:45: Precomputing --> Completed.


12/20/2016 15:26:45: Starting Epoch 1: learning rate per sample = 0.000391  effective momentum = 0.900000  momentum as time constant = 2429.8 samples

12/20/2016 15:26:45: Starting minibatch loop.
12/20/2016 15:26:45:  Epoch[ 1 of 2]-Minibatch[   1-  10]: CE.SM = 5.16985092 * 2560; Err = 0.97539062 * 2560; time = 0.0146s; samplesPerSecond = 175679.4
12/20/2016 15:26:45:  Epoch[ 1 of 2]-Minibatch[  11-  20]: CE.SM = 4.88109818 * 2560; Err = 0.94882813 * 2560; time = 0.0114s; samplesPerSecond = 224956.1
12/20/2016 15:26:45:  Epoch[ 1 of 2]-Minibatch[  21-  30]: CE.SM = 4.67409973 * 2560; Err = 0.91250000 * 2560; time = 0.0111s; samplesPerSecond = 230651.4
12/20/2016 15:26:45:  Epoch[ 1 of 2]-Minibatch[  31-  40]: CE.SM = 4.59894714 * 2560; Err = 0.91171875 * 2560; time = 0.0111s; samplesPerSecond = 229781.9
12/20/2016 15:26:45:  Epoch[ 1 of 2]-Minibatch[  41-  50]: CE.SM = 4.46573792 * 2560; Err = 0.90390625 * 2560; time = 0.0113s; samplesPerSecond = 227192.0
12/20/2016 15:26:45:  Epoch[ 1 of 2]-Minibatch[  51-  60]: CE.SM = 4.40809021 * 2560; Err = 0.91640625 * 2560; time = 0.0111s; samplesPerSecond = 230526.8
12/20/2016 15:26:45:  Epoch[ 1 of 2]-Minibatch[  61-  70]: CE.SM = 4.31769409 * 2560; Err = 0.88984375 * 2560; time = 0.0112s; samplesPerSecond = 228612.3
12/20/2016 15:26:45:  Epoch[ 1 of 2]-Minibatch[  71-  80]: CE.SM = 4.26254272 * 2560; Err = 0.87187500 * 2560; time = 0.0106s; samplesPerSecond = 240511.1
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[  81-  90]: CE.SM = 4.12560120 * 2560; Err = 0.84765625 * 2560; time = 0.0146s; samplesPerSecond = 174947.0
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[  91- 100]: CE.SM = 4.13090210 * 2560; Err = 0.85976562 * 2560; time = 0.0276s; samplesPerSecond = 92679.7
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 101- 110]: CE.SM = 4.02817688 * 2560; Err = 0.84531250 * 2560; time = 0.0128s; samplesPerSecond = 199392.5
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 111- 120]: CE.SM = 4.02988586 * 2560; Err = 0.84101563 * 2560; time = 0.0131s; samplesPerSecond = 194765.7
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 121- 130]: CE.SM = 3.88218384 * 2560; Err = 0.81562500 * 2560; time = 0.0129s; samplesPerSecond = 198265.2
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 131- 140]: CE.SM = 3.83023071 * 2560; Err = 0.80273438 * 2560; time = 0.0132s; samplesPerSecond = 193470.4
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 141- 150]: CE.SM = 3.81896362 * 2560; Err = 0.80742187 * 2560; time = 0.0123s; samplesPerSecond = 207893.5
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 151- 160]: CE.SM = 3.79920044 * 2560; Err = 0.80156250 * 2560; time = 0.0127s; samplesPerSecond = 201686.0
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 161- 170]: CE.SM = 3.69235840 * 2560; Err = 0.78359375 * 2560; time = 0.0126s; samplesPerSecond = 203610.9
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 171- 180]: CE.SM = 3.67468262 * 2560; Err = 0.79531250 * 2560; time = 0.0123s; samplesPerSecond = 208503.0
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 181- 190]: CE.SM = 3.59113770 * 2560; Err = 0.77734375 * 2560; time = 0.0106s; samplesPerSecond = 241897.4
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 191- 200]: CE.SM = 3.61049805 * 2560; Err = 0.78789062 * 2560; time = 0.0117s; samplesPerSecond = 218990.6
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 201- 210]: CE.SM = 3.60463867 * 2560; Err = 0.79375000 * 2560; time = 0.0109s; samplesPerSecond = 233896.8
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 211- 220]: CE.SM = 3.52816162 * 2560; Err = 0.77734375 * 2560; time = 0.0107s; samplesPerSecond = 239880.1
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 221- 230]: CE.SM = 3.51124878 * 2560; Err = 0.77382812 * 2560; time = 0.0106s; samplesPerSecond = 240533.7
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 231- 240]: CE.SM = 3.49269409 * 2560; Err = 0.77265625 * 2560; time = 0.0103s; samplesPerSecond = 249731.7
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 241- 250]: CE.SM = 3.42261353 * 2560; Err = 0.75937500 * 2560; time = 0.0108s; samplesPerSecond = 237785.6
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 251- 260]: CE.SM = 3.47793579 * 2560; Err = 0.77265625 * 2560; time = 0.0107s; samplesPerSecond = 238516.7
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 261- 270]: CE.SM = 3.40418701 * 2560; Err = 0.76132813 * 2560; time = 0.0108s; samplesPerSecond = 237256.7
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 271- 280]: CE.SM = 3.44434814 * 2560; Err = 0.78164062 * 2560; time = 0.0112s; samplesPerSecond = 228042.0
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 281- 290]: CE.SM = 3.38089600 * 2560; Err = 0.76171875 * 2560; time = 0.0112s; samplesPerSecond = 227555.6
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 291- 300]: CE.SM = 3.35212402 * 2560; Err = 0.75742188 * 2560; time = 0.0111s; samplesPerSecond = 231464.7
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 301- 310]: CE.SM = 3.33083496 * 2560; Err = 0.74765625 * 2560; time = 0.0117s; samplesPerSecond = 217909.4
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 311- 320]: CE.SM = 3.25802002 * 2560; Err = 0.75273437 * 2560; time = 0.0112s; samplesPerSecond = 228286.1
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 321- 330]: CE.SM = 3.31188965 * 2560; Err = 0.75117188 * 2560; time = 0.0112s; samplesPerSecond = 228469.4
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 331- 340]: CE.SM = 3.31600342 * 2560; Err = 0.76835937 * 2560; time = 0.0112s; samplesPerSecond = 227920.2
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 341- 350]: CE.SM = 3.20303955 * 2560; Err = 0.75000000 * 2560; time = 0.0113s; samplesPerSecond = 226869.9
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 351- 360]: CE.SM = 3.25258789 * 2560; Err = 0.74570313 * 2560; time = 0.0111s; samplesPerSecond = 230485.3
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 361- 370]: CE.SM = 3.15861816 * 2560; Err = 0.73359375 * 2560; time = 0.0121s; samplesPerSecond = 210855.8
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 371- 380]: CE.SM = 3.17800293 * 2560; Err = 0.74726563 * 2560; time = 0.0114s; samplesPerSecond = 225134.1
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 381- 390]: CE.SM = 3.16248779 * 2560; Err = 0.73828125 * 2560; time = 0.0123s; samplesPerSecond = 208469.1
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 391- 400]: CE.SM = 3.18978271 * 2560; Err = 0.73164063 * 2560; time = 0.0118s; samplesPerSecond = 217724.1
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 401- 410]: CE.SM = 3.17496338 * 2560; Err = 0.73632812 * 2560; time = 0.0117s; samplesPerSecond = 219704.8
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 411- 420]: CE.SM = 3.10482178 * 2560; Err = 0.73554688 * 2560; time = 0.0117s; samplesPerSecond = 218747.3
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 421- 430]: CE.SM = 3.10295410 * 2560; Err = 0.72500000 * 2560; time = 0.0115s; samplesPerSecond = 223054.8
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 431- 440]: CE.SM = 3.09871826 * 2560; Err = 0.72578125 * 2560; time = 0.0119s; samplesPerSecond = 214279.7
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 441- 450]: CE.SM = 3.11589355 * 2560; Err = 0.74296875 * 2560; time = 0.0117s; samplesPerSecond = 218747.3
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 451- 460]: CE.SM = 3.05733643 * 2560; Err = 0.72421875 * 2560; time = 0.0115s; samplesPerSecond = 222280.1
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 461- 470]: CE.SM = 3.06462402 * 2560; Err = 0.71679688 * 2560; time = 0.0119s; samplesPerSecond = 215742.5
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 471- 480]: CE.SM = 3.06251221 * 2560; Err = 0.72695312 * 2560; time = 0.0118s; samplesPerSecond = 216949.2
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 481- 490]: CE.SM = 3.00922852 * 2560; Err = 0.70273438 * 2560; time = 0.0116s; samplesPerSecond = 221510.8
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 491- 500]: CE.SM = 2.99824219 * 2560; Err = 0.71289062 * 2560; time = 0.0115s; samplesPerSecond = 222996.5
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 501- 510]: CE.SM = 2.98570557 * 2560; Err = 0.72187500 * 2560; time = 0.0125s; samplesPerSecond = 205144.6
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 511- 520]: CE.SM = 2.98309326 * 2560; Err = 0.72109375 * 2560; time = 0.0119s; samplesPerSecond = 214351.5
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 521- 530]: CE.SM = 2.98503418 * 2560; Err = 0.71093750 * 2560; time = 0.0116s; samplesPerSecond = 220290.9
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 531- 540]: CE.SM = 2.96007080 * 2560; Err = 0.70312500 * 2560; time = 0.0118s; samplesPerSecond = 217539.1
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 541- 550]: CE.SM = 3.00593262 * 2560; Err = 0.72812500 * 2560; time = 0.0116s; samplesPerSecond = 219855.7
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 551- 560]: CE.SM = 2.98140869 * 2560; Err = 0.70585937 * 2560; time = 0.0114s; samplesPerSecond = 223737.1
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 561- 570]: CE.SM = 2.95456543 * 2560; Err = 0.70390625 * 2560; time = 0.0116s; samplesPerSecond = 221089.9
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 571- 580]: CE.SM = 2.95152588 * 2560; Err = 0.71796875 * 2560; time = 0.0116s; samplesPerSecond = 221606.6
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 581- 590]: CE.SM = 2.91608887 * 2560; Err = 0.69531250 * 2560; time = 0.0119s; samplesPerSecond = 215669.8
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 591- 600]: CE.SM = 2.86706543 * 2560; Err = 0.70117188 * 2560; time = 0.0115s; samplesPerSecond = 223385.7
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 601- 610]: CE.SM = 2.94248047 * 2560; Err = 0.71210938 * 2560; time = 0.0117s; samplesPerSecond = 219384.7
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 611- 620]: CE.SM = 2.90012207 * 2560; Err = 0.70000000 * 2560; time = 0.0117s; samplesPerSecond = 219347.1
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 621- 630]: CE.SM = 2.88273926 * 2560; Err = 0.70117188 * 2560; time = 0.0115s; samplesPerSecond = 222666.8
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 631- 640]: CE.SM = 2.82316895 * 2560; Err = 0.69765625 * 2560; time = 0.0116s; samplesPerSecond = 219818.0
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 641- 650]: CE.SM = 2.88974609 * 2560; Err = 0.70859375 * 2560; time = 0.0116s; samplesPerSecond = 220044.7
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 651- 660]: CE.SM = 2.84401855 * 2560; Err = 0.69023437 * 2560; time = 0.0115s; samplesPerSecond = 223249.3
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 661- 670]: CE.SM = 2.86259766 * 2560; Err = 0.71367187 * 2560; time = 0.0118s; samplesPerSecond = 217816.7
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 671- 680]: CE.SM = 2.86748047 * 2560; Err = 0.70429688 * 2560; time = 0.0117s; samplesPerSecond = 219271.9
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 681- 690]: CE.SM = 2.82419434 * 2560; Err = 0.69726562 * 2560; time = 0.0117s; samplesPerSecond = 219347.1
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 691- 700]: CE.SM = 2.79973145 * 2560; Err = 0.67968750 * 2560; time = 0.0118s; samplesPerSecond = 216216.2
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 701- 710]: CE.SM = 2.86638184 * 2560; Err = 0.71992188 * 2560; time = 0.0119s; samplesPerSecond = 214279.7
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 711- 720]: CE.SM = 2.83571777 * 2560; Err = 0.70273438 * 2560; time = 0.0116s; samplesPerSecond = 220480.6
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 721- 730]: CE.SM = 2.79506836 * 2560; Err = 0.68476563 * 2560; time = 0.0117s; samplesPerSecond = 218822.1
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 731- 740]: CE.SM = 2.82990723 * 2560; Err = 0.68945312 * 2560; time = 0.0116s; samplesPerSecond = 220803.9
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 741- 750]: CE.SM = 2.78061523 * 2560; Err = 0.68945312 * 2560; time = 0.0115s; samplesPerSecond = 222106.5
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 751- 760]: CE.SM = 2.74165039 * 2560; Err = 0.67851562 * 2560; time = 0.0117s; samplesPerSecond = 219328.3
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 761- 770]: CE.SM = 2.72746582 * 2560; Err = 0.67187500 * 2560; time = 0.0116s; samplesPerSecond = 220822.9
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 771- 780]: CE.SM = 2.75603027 * 2560; Err = 0.69726562 * 2560; time = 0.0116s; samplesPerSecond = 220082.5
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 781- 790]: CE.SM = 2.75217285 * 2560; Err = 0.67929688 * 2560; time = 0.0116s; samplesPerSecond = 220556.6
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 791- 800]: CE.SM = 2.74196777 * 2560; Err = 0.69140625 * 2560; time = 0.0117s; samplesPerSecond = 218953.1
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 801- 810]: CE.SM = 2.74924316 * 2560; Err = 0.67890625 * 2560; time = 0.0116s; samplesPerSecond = 220803.9
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 811- 820]: CE.SM = 2.73066406 * 2560; Err = 0.67968750 * 2560; time = 0.0117s; samplesPerSecond = 219215.6
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 821- 830]: CE.SM = 2.74926758 * 2560; Err = 0.68007812 * 2560; time = 0.0115s; samplesPerSecond = 221856.3
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 831- 840]: CE.SM = 2.71374512 * 2560; Err = 0.67929688 * 2560; time = 0.0115s; samplesPerSecond = 221741.0
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 841- 850]: CE.SM = 2.70029297 * 2560; Err = 0.66835937 * 2560; time = 0.0117s; samplesPerSecond = 219403.5
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 851- 860]: CE.SM = 2.70891113 * 2560; Err = 0.67500000 * 2560; time = 0.0115s; samplesPerSecond = 223483.2
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 861- 870]: CE.SM = 2.72526855 * 2560; Err = 0.67304688 * 2560; time = 0.0116s; samplesPerSecond = 220765.8
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 871- 880]: CE.SM = 2.71545410 * 2560; Err = 0.67304688 * 2560; time = 0.0115s; samplesPerSecond = 222705.5
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 881- 890]: CE.SM = 2.68872070 * 2560; Err = 0.66132813 * 2560; time = 0.0115s; samplesPerSecond = 222241.5
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 891- 900]: CE.SM = 2.67675781 * 2560; Err = 0.67148438 * 2560; time = 0.0115s; samplesPerSecond = 223307.7
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 901- 910]: CE.SM = 2.77314453 * 2560; Err = 0.69140625 * 2560; time = 0.0115s; samplesPerSecond = 222183.6
12/20/2016 15:26:46:  Epoch[ 1 of 2]-Minibatch[ 911- 920]: CE.SM = 2.67338867 * 2560; Err = 0.67460937 * 2560; time = 0.0106s; samplesPerSecond = 241623.4
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[ 921- 930]: CE.SM = 2.73891602 * 2560; Err = 0.69335938 * 2560; time = 0.0115s; samplesPerSecond = 222783.0
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[ 931- 940]: CE.SM = 2.64516602 * 2560; Err = 0.66171875 * 2560; time = 0.0114s; samplesPerSecond = 224266.3
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[ 941- 950]: CE.SM = 2.63547363 * 2560; Err = 0.65312500 * 2560; time = 0.0115s; samplesPerSecond = 222029.5
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[ 951- 960]: CE.SM = 2.64746094 * 2560; Err = 0.66679687 * 2560; time = 0.0115s; samplesPerSecond = 222686.2
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[ 961- 970]: CE.SM = 2.67902832 * 2560; Err = 0.67929688 * 2560; time = 0.0116s; samplesPerSecond = 221395.8
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[ 971- 980]: CE.SM = 2.62915039 * 2560; Err = 0.66796875 * 2560; time = 0.0117s; samplesPerSecond = 218971.9
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[ 981- 990]: CE.SM = 2.62128906 * 2560; Err = 0.64570313 * 2560; time = 0.0114s; samplesPerSecond = 224148.5
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[ 991-1000]: CE.SM = 2.63820801 * 2560; Err = 0.66445312 * 2560; time = 0.0111s; samplesPerSecond = 231172.1
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1001-1010]: CE.SM = 2.60112305 * 2560; Err = 0.66523438 * 2560; time = 0.0111s; samplesPerSecond = 229967.7
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1011-1020]: CE.SM = 2.64233398 * 2560; Err = 0.66367188 * 2560; time = 0.0108s; samplesPerSecond = 237454.8
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1021-1030]: CE.SM = 2.59216309 * 2560; Err = 0.66210938 * 2560; time = 0.0111s; samplesPerSecond = 230589.1
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1031-1040]: CE.SM = 2.57377930 * 2560; Err = 0.66250000 * 2560; time = 0.0116s; samplesPerSecond = 220727.7
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1041-1050]: CE.SM = 2.58576660 * 2560; Err = 0.64648438 * 2560; time = 0.0116s; samplesPerSecond = 221395.8
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1051-1060]: CE.SM = 2.61821289 * 2560; Err = 0.65390625 * 2560; time = 0.0114s; samplesPerSecond = 224168.1
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1061-1070]: CE.SM = 2.58085938 * 2560; Err = 0.65468750 * 2560; time = 0.0116s; samplesPerSecond = 220328.8
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1071-1080]: CE.SM = 2.60266113 * 2560; Err = 0.66406250 * 2560; time = 0.0115s; samplesPerSecond = 222396.0
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1081-1090]: CE.SM = 2.58383789 * 2560; Err = 0.65429688 * 2560; time = 0.0115s; samplesPerSecond = 221933.2
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1091-1100]: CE.SM = 2.62839355 * 2560; Err = 0.67187500 * 2560; time = 0.0115s; samplesPerSecond = 222125.8
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1101-1110]: CE.SM = 2.56481934 * 2560; Err = 0.64375000 * 2560; time = 0.0115s; samplesPerSecond = 222453.9
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1111-1120]: CE.SM = 2.57800293 * 2560; Err = 0.64453125 * 2560; time = 0.0115s; samplesPerSecond = 222512.0
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1121-1130]: CE.SM = 2.55087891 * 2560; Err = 0.63984375 * 2560; time = 0.0114s; samplesPerSecond = 223698.0
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1131-1140]: CE.SM = 2.55930176 * 2560; Err = 0.64882812 * 2560; time = 0.0114s; samplesPerSecond = 224541.7
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1141-1150]: CE.SM = 2.49941406 * 2560; Err = 0.62812500 * 2560; time = 0.0110s; samplesPerSecond = 232981.4
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1151-1160]: CE.SM = 2.50043945 * 2560; Err = 0.64843750 * 2560; time = 0.0116s; samplesPerSecond = 221376.7
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1161-1170]: CE.SM = 2.52336426 * 2560; Err = 0.64570313 * 2560; time = 0.0115s; samplesPerSecond = 223054.8
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1171-1180]: CE.SM = 2.55410156 * 2560; Err = 0.65039062 * 2560; time = 0.0115s; samplesPerSecond = 223444.2
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1181-1190]: CE.SM = 2.54072266 * 2560; Err = 0.63906250 * 2560; time = 0.0116s; samplesPerSecond = 220177.2
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1191-1200]: CE.SM = 2.52595215 * 2560; Err = 0.64804688 * 2560; time = 0.0116s; samplesPerSecond = 221338.4
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1201-1210]: CE.SM = 2.52065430 * 2560; Err = 0.65351563 * 2560; time = 0.0115s; samplesPerSecond = 221702.6
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1211-1220]: CE.SM = 2.53261719 * 2560; Err = 0.63359375 * 2560; time = 0.0117s; samplesPerSecond = 218430.0
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1221-1230]: CE.SM = 2.52692871 * 2560; Err = 0.66015625 * 2560; time = 0.0112s; samplesPerSecond = 228225.0
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1231-1240]: CE.SM = 2.49379883 * 2560; Err = 0.64257812 * 2560; time = 0.0114s; samplesPerSecond = 224482.6
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1241-1250]: CE.SM = 2.52558594 * 2560; Err = 0.63945312 * 2560; time = 0.0115s; samplesPerSecond = 223229.9
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1251-1260]: CE.SM = 2.47526855 * 2560; Err = 0.63828125 * 2560; time = 0.0113s; samplesPerSecond = 225650.1
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1261-1270]: CE.SM = 2.48032227 * 2560; Err = 0.63281250 * 2560; time = 0.0115s; samplesPerSecond = 223249.3
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1271-1280]: CE.SM = 2.54611816 * 2560; Err = 0.64843750 * 2560; time = 0.0114s; samplesPerSecond = 224030.8
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1281-1290]: CE.SM = 2.49899902 * 2560; Err = 0.64648438 * 2560; time = 0.0113s; samplesPerSecond = 225610.3
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1291-1300]: CE.SM = 2.51059570 * 2560; Err = 0.64062500 * 2560; time = 0.0112s; samplesPerSecond = 229452.4
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1301-1310]: CE.SM = 2.47578125 * 2560; Err = 0.63320312 * 2560; time = 0.0112s; samplesPerSecond = 227960.8
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1311-1320]: CE.SM = 2.45412598 * 2560; Err = 0.64101562 * 2560; time = 0.0113s; samplesPerSecond = 225650.1
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1321-1330]: CE.SM = 2.46501465 * 2560; Err = 0.63007813 * 2560; time = 0.0115s; samplesPerSecond = 222628.1
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1331-1340]: CE.SM = 2.48574219 * 2560; Err = 0.63828125 * 2560; time = 0.0113s; samplesPerSecond = 226288.3
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1341-1350]: CE.SM = 2.45244141 * 2560; Err = 0.63476562 * 2560; time = 0.0115s; samplesPerSecond = 223152.0
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1351-1360]: CE.SM = 2.42282715 * 2560; Err = 0.62031250 * 2560; time = 0.0113s; samplesPerSecond = 225769.5
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1361-1370]: CE.SM = 2.46435547 * 2560; Err = 0.63281250 * 2560; time = 0.0113s; samplesPerSecond = 226869.9
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1371-1380]: CE.SM = 2.44033203 * 2560; Err = 0.62109375 * 2560; time = 0.0116s; samplesPerSecond = 220252.9
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1381-1390]: CE.SM = 2.48945312 * 2560; Err = 0.65078125 * 2560; time = 0.0112s; samplesPerSecond = 228001.4
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1391-1400]: CE.SM = 2.38950195 * 2560; Err = 0.62031250 * 2560; time = 0.0111s; samplesPerSecond = 230963.6
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1401-1410]: CE.SM = 2.45888672 * 2560; Err = 0.62773437 * 2560; time = 0.0113s; samplesPerSecond = 227292.9
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1411-1420]: CE.SM = 2.47119141 * 2560; Err = 0.64687500 * 2560; time = 0.0114s; samplesPerSecond = 224266.3
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1421-1430]: CE.SM = 2.44711914 * 2560; Err = 0.64218750 * 2560; time = 0.0113s; samplesPerSecond = 226890.0
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1431-1440]: CE.SM = 2.44638672 * 2560; Err = 0.63085938 * 2560; time = 0.0111s; samplesPerSecond = 230071.0
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1441-1450]: CE.SM = 2.38784180 * 2560; Err = 0.61757812 * 2560; time = 0.0115s; samplesPerSecond = 223366.2
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1451-1460]: CE.SM = 2.44887695 * 2560; Err = 0.64726562 * 2560; time = 0.0110s; samplesPerSecond = 232579.3
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1461-1470]: CE.SM = 2.42319336 * 2560; Err = 0.61718750 * 2560; time = 0.0110s; samplesPerSecond = 231842.1
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1471-1480]: CE.SM = 2.40356445 * 2560; Err = 0.61484375 * 2560; time = 0.0111s; samplesPerSecond = 230568.3
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1481-1490]: CE.SM = 2.44067383 * 2560; Err = 0.62421875 * 2560; time = 0.0113s; samplesPerSecond = 225729.7
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1491-1500]: CE.SM = 2.42299805 * 2560; Err = 0.62070313 * 2560; time = 0.0110s; samplesPerSecond = 232685.0
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1501-1510]: CE.SM = 2.36215820 * 2560; Err = 0.62226563 * 2560; time = 0.0110s; samplesPerSecond = 232073.2
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1511-1520]: CE.SM = 2.39360352 * 2560; Err = 0.60976562 * 2560; time = 0.0111s; samplesPerSecond = 230215.8
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1521-1530]: CE.SM = 2.42197266 * 2560; Err = 0.62265625 * 2560; time = 0.0110s; samplesPerSecond = 232094.3
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1531-1540]: CE.SM = 2.44008789 * 2560; Err = 0.62890625 * 2560; time = 0.0110s; samplesPerSecond = 231695.2
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1541-1550]: CE.SM = 2.46518555 * 2560; Err = 0.63828125 * 2560; time = 0.0110s; samplesPerSecond = 232917.8
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1551-1560]: CE.SM = 2.38061523 * 2560; Err = 0.63203125 * 2560; time = 0.0110s; samplesPerSecond = 233236.2
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1561-1570]: CE.SM = 2.36083984 * 2560; Err = 0.62460938 * 2560; time = 0.0121s; samplesPerSecond = 210960.0
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1571-1580]: CE.SM = 2.37851562 * 2560; Err = 0.62031250 * 2560; time = 0.0111s; samplesPerSecond = 231485.7
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1581-1590]: CE.SM = 2.41503906 * 2560; Err = 0.61445313 * 2560; time = 0.0115s; samplesPerSecond = 223015.9
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1591-1600]: CE.SM = 2.37304688 * 2560; Err = 0.60195312 * 2560; time = 0.0118s; samplesPerSecond = 217372.8
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1601-1610]: CE.SM = 2.39086914 * 2560; Err = 0.63906250 * 2560; time = 0.0114s; samplesPerSecond = 225530.8
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1611-1620]: CE.SM = 2.39985352 * 2560; Err = 0.63164062 * 2560; time = 0.0117s; samplesPerSecond = 219065.5
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1621-1630]: CE.SM = 2.37680664 * 2560; Err = 0.62031250 * 2560; time = 0.0115s; samplesPerSecond = 222280.1
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1631-1640]: CE.SM = 2.35844727 * 2560; Err = 0.61171875 * 2560; time = 0.0114s; samplesPerSecond = 225253.0
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1641-1650]: CE.SM = 2.33872070 * 2560; Err = 0.61445313 * 2560; time = 0.0114s; samplesPerSecond = 225015.4
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1651-1660]: CE.SM = 2.34477539 * 2560; Err = 0.61250000 * 2560; time = 0.0114s; samplesPerSecond = 225391.8
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1661-1670]: CE.SM = 2.37587891 * 2560; Err = 0.62265625 * 2560; time = 0.0112s; samplesPerSecond = 228959.8
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1671-1680]: CE.SM = 2.37065430 * 2560; Err = 0.61757812 * 2560; time = 0.0112s; samplesPerSecond = 229082.8
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1681-1690]: CE.SM = 2.37475586 * 2560; Err = 0.61914062 * 2560; time = 0.0114s; samplesPerSecond = 225471.2
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1691-1700]: CE.SM = 2.33686523 * 2560; Err = 0.61367187 * 2560; time = 0.0112s; samplesPerSecond = 229472.9
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1701-1710]: CE.SM = 2.28505859 * 2560; Err = 0.58671875 * 2560; time = 0.0113s; samplesPerSecond = 227535.3
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1711-1720]: CE.SM = 2.37153320 * 2560; Err = 0.61640625 * 2560; time = 0.0113s; samplesPerSecond = 226729.3
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1721-1730]: CE.SM = 2.38281250 * 2560; Err = 0.62304688 * 2560; time = 0.0114s; samplesPerSecond = 225451.3
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1731-1740]: CE.SM = 2.35830078 * 2560; Err = 0.62695312 * 2560; time = 0.0116s; samplesPerSecond = 221261.9
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1741-1750]: CE.SM = 2.35527344 * 2560; Err = 0.61718750 * 2560; time = 0.0114s; samplesPerSecond = 224659.9
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1751-1760]: CE.SM = 2.34887695 * 2560; Err = 0.59960938 * 2560; time = 0.0114s; samplesPerSecond = 224266.3
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1761-1770]: CE.SM = 2.29965820 * 2560; Err = 0.59023437 * 2560; time = 0.0114s; samplesPerSecond = 224128.9
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1771-1780]: CE.SM = 2.37202148 * 2560; Err = 0.62343750 * 2560; time = 0.0117s; samplesPerSecond = 218020.8
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1781-1790]: CE.SM = 2.30996094 * 2560; Err = 0.59414062 * 2560; time = 0.0119s; samplesPerSecond = 215488.2
12/20/2016 15:26:47:  Epoch[ 1 of 2]-Minibatch[1791-1800]: CE.SM = 2.37407227 * 2560; Err = 0.61406250 * 2560; time = 0.0108s; samplesPerSecond = 237520.9
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[1801-1810]: CE.SM = 2.32983398 * 2560; Err = 0.61171875 * 2560; time = 0.0120s; samplesPerSecond = 213689.5
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[1811-1820]: CE.SM = 2.28408203 * 2560; Err = 0.61601562 * 2560; time = 0.0117s; samplesPerSecond = 219196.8
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[1821-1830]: CE.SM = 2.27836914 * 2560; Err = 0.60117188 * 2560; time = 0.0117s; samplesPerSecond = 219667.1
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[1831-1840]: CE.SM = 2.32036133 * 2560; Err = 0.61445313 * 2560; time = 0.0118s; samplesPerSecond = 216912.4
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[1841-1850]: CE.SM = 2.33315430 * 2560; Err = 0.61679688 * 2560; time = 0.0116s; samplesPerSecond = 220290.9
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[1851-1860]: CE.SM = 2.25014648 * 2560; Err = 0.59921875 * 2560; time = 0.0115s; samplesPerSecond = 222164.4
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[1861-1870]: CE.SM = 2.30361328 * 2560; Err = 0.60625000 * 2560; time = 0.0117s; samplesPerSecond = 217983.7
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[1871-1880]: CE.SM = 2.32231445 * 2560; Err = 0.59726563 * 2560; time = 0.0117s; samplesPerSecond = 218318.3
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[1881-1890]: CE.SM = 2.34755859 * 2560; Err = 0.61562500 * 2560; time = 0.0116s; samplesPerSecond = 221128.1
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[1891-1900]: CE.SM = 2.29873047 * 2560; Err = 0.60351562 * 2560; time = 0.0117s; samplesPerSecond = 218057.9
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[1901-1910]: CE.SM = 2.25966797 * 2560; Err = 0.60000000 * 2560; time = 0.0118s; samplesPerSecond = 216710.4
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[1911-1920]: CE.SM = 2.29873047 * 2560; Err = 0.59960938 * 2560; time = 0.0119s; samplesPerSecond = 214459.2
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[1921-1930]: CE.SM = 2.29389648 * 2560; Err = 0.61015625 * 2560; time = 0.0117s; samplesPerSecond = 219234.4
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[1931-1940]: CE.SM = 2.29184570 * 2560; Err = 0.60234375 * 2560; time = 0.0116s; samplesPerSecond = 221089.9
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[1941-1950]: CE.SM = 2.27509766 * 2560; Err = 0.60664063 * 2560; time = 0.0118s; samplesPerSecond = 216216.2
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[1951-1960]: CE.SM = 2.26972656 * 2560; Err = 0.59843750 * 2560; time = 0.0118s; samplesPerSecond = 216508.8
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[1961-1970]: CE.SM = 2.31513672 * 2560; Err = 0.61289063 * 2560; time = 0.0117s; samplesPerSecond = 218953.1
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[1971-1980]: CE.SM = 2.26860352 * 2560; Err = 0.58750000 * 2560; time = 0.0119s; samplesPerSecond = 215288.9
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[1981-1990]: CE.SM = 2.25581055 * 2560; Err = 0.58867187 * 2560; time = 0.0118s; samplesPerSecond = 217594.6
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[1991-2000]: CE.SM = 2.28090820 * 2560; Err = 0.60195312 * 2560; time = 0.0118s; samplesPerSecond = 216234.5
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2001-2010]: CE.SM = 2.25107422 * 2560; Err = 0.58789062 * 2560; time = 0.0118s; samplesPerSecond = 216307.6
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2011-2020]: CE.SM = 2.30180664 * 2560; Err = 0.60234375 * 2560; time = 0.0119s; samplesPerSecond = 215815.2
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2021-2030]: CE.SM = 2.26474609 * 2560; Err = 0.59218750 * 2560; time = 0.0118s; samplesPerSecond = 216673.7
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2031-2040]: CE.SM = 2.28779297 * 2560; Err = 0.61250000 * 2560; time = 0.0121s; samplesPerSecond = 211832.9
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2041-2050]: CE.SM = 2.30791016 * 2560; Err = 0.61210937 * 2560; time = 0.0117s; samplesPerSecond = 219309.5
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2051-2060]: CE.SM = 2.26098633 * 2560; Err = 0.60468750 * 2560; time = 0.0119s; samplesPerSecond = 215451.9
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2061-2070]: CE.SM = 2.27451172 * 2560; Err = 0.60117188 * 2560; time = 0.0118s; samplesPerSecond = 216875.6
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2071-2080]: CE.SM = 2.26469727 * 2560; Err = 0.60781250 * 2560; time = 0.0119s; samplesPerSecond = 215506.4
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2081-2090]: CE.SM = 2.28339844 * 2560; Err = 0.60625000 * 2560; time = 0.0120s; samplesPerSecond = 214118.4
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2091-2100]: CE.SM = 2.24809570 * 2560; Err = 0.58632812 * 2560; time = 0.0118s; samplesPerSecond = 216033.8
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2101-2110]: CE.SM = 2.22905273 * 2560; Err = 0.59570312 * 2560; time = 0.0120s; samplesPerSecond = 214028.9
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2111-2120]: CE.SM = 2.29106445 * 2560; Err = 0.61250000 * 2560; time = 0.0120s; samplesPerSecond = 213138.0
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2121-2130]: CE.SM = 2.22280273 * 2560; Err = 0.59648437 * 2560; time = 0.0120s; samplesPerSecond = 212624.6
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2131-2140]: CE.SM = 2.26313477 * 2560; Err = 0.61093750 * 2560; time = 0.0121s; samplesPerSecond = 211325.7
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2141-2150]: CE.SM = 2.22944336 * 2560; Err = 0.61132812 * 2560; time = 0.0120s; samplesPerSecond = 212659.9
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2151-2160]: CE.SM = 2.20976563 * 2560; Err = 0.59296875 * 2560; time = 0.0121s; samplesPerSecond = 210855.8
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2161-2170]: CE.SM = 2.23154297 * 2560; Err = 0.58476562 * 2560; time = 0.0122s; samplesPerSecond = 210059.9
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2171-2180]: CE.SM = 2.26240234 * 2560; Err = 0.60468750 * 2560; time = 0.0115s; samplesPerSecond = 222145.1
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2181-2190]: CE.SM = 2.20190430 * 2560; Err = 0.59179688 * 2560; time = 0.0122s; samplesPerSecond = 209475.5
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2191-2200]: CE.SM = 2.19291992 * 2560; Err = 0.58398438 * 2560; time = 0.0112s; samplesPerSecond = 229123.8
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2201-2210]: CE.SM = 2.21699219 * 2560; Err = 0.58320313 * 2560; time = 0.0113s; samplesPerSecond = 227252.6
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2211-2220]: CE.SM = 2.18994141 * 2560; Err = 0.57382813 * 2560; time = 0.0122s; samplesPerSecond = 210439.8
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2221-2230]: CE.SM = 2.20864258 * 2560; Err = 0.57695312 * 2560; time = 0.0122s; samplesPerSecond = 209235.8
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2231-2240]: CE.SM = 2.18833008 * 2560; Err = 0.56953125 * 2560; time = 0.0118s; samplesPerSecond = 217170.0
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2241-2250]: CE.SM = 2.22241211 * 2560; Err = 0.58046875 * 2560; time = 0.0116s; samplesPerSecond = 221051.7
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2251-2260]: CE.SM = 2.28413086 * 2560; Err = 0.60546875 * 2560; time = 0.0117s; samplesPerSecond = 218188.0
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2261-2270]: CE.SM = 2.16279297 * 2560; Err = 0.58046875 * 2560; time = 0.0116s; samplesPerSecond = 221606.6
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2271-2280]: CE.SM = 2.18374023 * 2560; Err = 0.58593750 * 2560; time = 0.0117s; samplesPerSecond = 219065.5
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2281-2290]: CE.SM = 2.20312500 * 2560; Err = 0.56914062 * 2560; time = 0.0117s; samplesPerSecond = 218336.9
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2291-2300]: CE.SM = 2.17187500 * 2560; Err = 0.59960938 * 2560; time = 0.0118s; samplesPerSecond = 217225.3
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2301-2310]: CE.SM = 2.20664062 * 2560; Err = 0.57773438 * 2560; time = 0.0118s; samplesPerSecond = 217687.1
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2311-2320]: CE.SM = 2.25551758 * 2560; Err = 0.60156250 * 2560; time = 0.0118s; samplesPerSecond = 216967.5
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2321-2330]: CE.SM = 2.20771484 * 2560; Err = 0.60000000 * 2560; time = 0.0117s; samplesPerSecond = 218822.1
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2331-2340]: CE.SM = 2.18901367 * 2560; Err = 0.59414062 * 2560; time = 0.0116s; samplesPerSecond = 220271.9
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2341-2350]: CE.SM = 2.18730469 * 2560; Err = 0.58593750 * 2560; time = 0.0115s; samplesPerSecond = 221894.8
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2351-2360]: CE.SM = 2.20336914 * 2560; Err = 0.58906250 * 2560; time = 0.0119s; samplesPerSecond = 215833.4
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2361-2370]: CE.SM = 2.24916992 * 2560; Err = 0.59726563 * 2560; time = 0.0115s; samplesPerSecond = 221741.0
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2371-2380]: CE.SM = 2.18256836 * 2560; Err = 0.57382813 * 2560; time = 0.0116s; samplesPerSecond = 220215.1
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2381-2390]: CE.SM = 2.20000000 * 2560; Err = 0.59023437 * 2560; time = 0.0116s; samplesPerSecond = 221625.8
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2391-2400]: CE.SM = 2.23159180 * 2560; Err = 0.59648437 * 2560; time = 0.0115s; samplesPerSecond = 222434.6
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2401-2410]: CE.SM = 2.18208008 * 2560; Err = 0.58007812 * 2560; time = 0.0118s; samplesPerSecond = 217391.3
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2411-2420]: CE.SM = 2.19296875 * 2560; Err = 0.59492188 * 2560; time = 0.0114s; samplesPerSecond = 225173.7
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2421-2430]: CE.SM = 2.20561523 * 2560; Err = 0.58945313 * 2560; time = 0.0114s; samplesPerSecond = 225391.8
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2431-2440]: CE.SM = 2.13676758 * 2560; Err = 0.57656250 * 2560; time = 0.0122s; samplesPerSecond = 209047.9
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2441-2450]: CE.SM = 2.18452148 * 2560; Err = 0.59023437 * 2560; time = 0.0119s; samplesPerSecond = 214423.3
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2451-2460]: CE.SM = 2.23198242 * 2560; Err = 0.59023437 * 2560; time = 0.0122s; samplesPerSecond = 209441.2
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2461-2470]: CE.SM = 2.20742188 * 2560; Err = 0.58281250 * 2560; time = 0.0122s; samplesPerSecond = 209767.3
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2471-2480]: CE.SM = 2.19599609 * 2560; Err = 0.57578125 * 2560; time = 0.0121s; samplesPerSecond = 211640.2
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2481-2490]: CE.SM = 2.15083008 * 2560; Err = 0.57656250 * 2560; time = 0.0125s; samplesPerSecond = 204211.9
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2491-2500]: CE.SM = 2.17768555 * 2560; Err = 0.58593750 * 2560; time = 0.0123s; samplesPerSecond = 208894.3
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2501-2510]: CE.SM = 2.16074219 * 2560; Err = 0.58593750 * 2560; time = 0.0121s; samplesPerSecond = 212148.8
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2511-2520]: CE.SM = 2.15156250 * 2560; Err = 0.58085937 * 2560; time = 0.0121s; samplesPerSecond = 210960.0
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2521-2530]: CE.SM = 2.19785156 * 2560; Err = 0.58515625 * 2560; time = 0.0121s; samplesPerSecond = 211850.4
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2531-2540]: CE.SM = 2.15917969 * 2560; Err = 0.57031250 * 2560; time = 0.0122s; samplesPerSecond = 210491.7
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2541-2550]: CE.SM = 2.19301758 * 2560; Err = 0.58632812 * 2560; time = 0.0121s; samplesPerSecond = 212324.8
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2551-2560]: CE.SM = 2.20317383 * 2560; Err = 0.59453125 * 2560; time = 0.0120s; samplesPerSecond = 212677.6
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2561-2570]: CE.SM = 2.19335938 * 2560; Err = 0.58710938 * 2560; time = 0.0122s; samplesPerSecond = 210284.2
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2571-2580]: CE.SM = 2.17050781 * 2560; Err = 0.58437500 * 2560; time = 0.0124s; samplesPerSecond = 205837.4
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2581-2590]: CE.SM = 2.15166016 * 2560; Err = 0.56328125 * 2560; time = 0.0120s; samplesPerSecond = 212748.3
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2591-2600]: CE.SM = 2.19687500 * 2560; Err = 0.59218750 * 2560; time = 0.0123s; samplesPerSecond = 208724.0
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2601-2610]: CE.SM = 2.21098633 * 2560; Err = 0.58945313 * 2560; time = 0.0121s; samplesPerSecond = 212219.2
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2611-2620]: CE.SM = 2.17670898 * 2560; Err = 0.58671875 * 2560; time = 0.0121s; samplesPerSecond = 211570.2
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2621-2630]: CE.SM = 2.14199219 * 2560; Err = 0.58125000 * 2560; time = 0.0119s; samplesPerSecond = 215089.9
12/20/2016 15:26:48:  Epoch[ 1 of 2]-Minibatch[2631-2640]: CE.SM = 2.19086914 * 2560; Err = 0.58046875 * 2560; time = 0.0111s; samplesPerSecond = 231193.0
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2641-2650]: CE.SM = 2.18139648 * 2560; Err = 0.57226562 * 2560; time = 0.0121s; samplesPerSecond = 212289.6
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2651-2660]: CE.SM = 2.17573242 * 2560; Err = 0.59062500 * 2560; time = 0.0117s; samplesPerSecond = 218915.7
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2661-2670]: CE.SM = 2.15190430 * 2560; Err = 0.57343750 * 2560; time = 0.0122s; samplesPerSecond = 209235.8
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2671-2680]: CE.SM = 2.12607422 * 2560; Err = 0.58593750 * 2560; time = 0.0124s; samplesPerSecond = 207019.2
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2681-2690]: CE.SM = 2.07285156 * 2560; Err = 0.57148438 * 2560; time = 0.0122s; samplesPerSecond = 209099.1
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2691-2700]: CE.SM = 2.15703125 * 2560; Err = 0.59257812 * 2560; time = 0.0125s; samplesPerSecond = 204980.4
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2701-2710]: CE.SM = 2.18784180 * 2560; Err = 0.59296875 * 2560; time = 0.0128s; samplesPerSecond = 200125.1
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2711-2720]: CE.SM = 2.11215820 * 2560; Err = 0.56328125 * 2560; time = 0.0124s; samplesPerSecond = 205936.8
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2721-2730]: CE.SM = 2.14897461 * 2560; Err = 0.58320313 * 2560; time = 0.0121s; samplesPerSecond = 210994.8
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2731-2740]: CE.SM = 2.20468750 * 2560; Err = 0.58750000 * 2560; time = 0.0125s; samplesPerSecond = 204898.4
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2741-2750]: CE.SM = 2.13842773 * 2560; Err = 0.58476562 * 2560; time = 0.0122s; samplesPerSecond = 210025.4
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2751-2760]: CE.SM = 2.19257812 * 2560; Err = 0.59062500 * 2560; time = 0.0122s; samplesPerSecond = 209235.8
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2761-2770]: CE.SM = 2.12314453 * 2560; Err = 0.57187500 * 2560; time = 0.0121s; samplesPerSecond = 210873.1
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2771-2780]: CE.SM = 2.11518555 * 2560; Err = 0.57460937 * 2560; time = 0.0124s; samplesPerSecond = 206218.8
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2781-2790]: CE.SM = 2.16567383 * 2560; Err = 0.57851562 * 2560; time = 0.0123s; samplesPerSecond = 207775.3
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2791-2800]: CE.SM = 2.12832031 * 2560; Err = 0.57890625 * 2560; time = 0.0124s; samplesPerSecond = 207186.8
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2801-2810]: CE.SM = 2.16328125 * 2560; Err = 0.57812500 * 2560; time = 0.0120s; samplesPerSecond = 212659.9
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2811-2820]: CE.SM = 2.10776367 * 2560; Err = 0.56406250 * 2560; time = 0.0120s; samplesPerSecond = 213939.5
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2821-2830]: CE.SM = 2.17670898 * 2560; Err = 0.58906250 * 2560; time = 0.0120s; samplesPerSecond = 213725.2
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2831-2840]: CE.SM = 2.15180664 * 2560; Err = 0.57929688 * 2560; time = 0.0123s; samplesPerSecond = 207657.4
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2841-2850]: CE.SM = 2.15478516 * 2560; Err = 0.58320313 * 2560; time = 0.0120s; samplesPerSecond = 212801.3
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2851-2860]: CE.SM = 2.17939453 * 2560; Err = 0.58281250 * 2560; time = 0.0121s; samplesPerSecond = 211116.6
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2861-2870]: CE.SM = 2.17099609 * 2560; Err = 0.57890625 * 2560; time = 0.0121s; samplesPerSecond = 210925.3
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2871-2880]: CE.SM = 2.20527344 * 2560; Err = 0.58593750 * 2560; time = 0.0122s; samplesPerSecond = 209509.8
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2881-2890]: CE.SM = 2.10473633 * 2560; Err = 0.58125000 * 2560; time = 0.0124s; samplesPerSecond = 206634.9
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2891-2900]: CE.SM = 2.07514648 * 2560; Err = 0.56210938 * 2560; time = 0.0122s; samplesPerSecond = 210146.1
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2901-2910]: CE.SM = 2.10834961 * 2560; Err = 0.57148438 * 2560; time = 0.0122s; samplesPerSecond = 210301.5
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2911-2920]: CE.SM = 2.12221680 * 2560; Err = 0.56640625 * 2560; time = 0.0121s; samplesPerSecond = 210907.9
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2921-2930]: CE.SM = 2.09565430 * 2560; Err = 0.58125000 * 2560; time = 0.0121s; samplesPerSecond = 211657.7
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2931-2940]: CE.SM = 2.14960937 * 2560; Err = 0.57304687 * 2560; time = 0.0115s; samplesPerSecond = 221779.4
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2941-2950]: CE.SM = 2.10634766 * 2560; Err = 0.56093750 * 2560; time = 0.0120s; samplesPerSecond = 212801.3
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2951-2960]: CE.SM = 2.12333984 * 2560; Err = 0.57343750 * 2560; time = 0.0122s; samplesPerSecond = 209698.6
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2961-2970]: CE.SM = 2.12602539 * 2560; Err = 0.56640625 * 2560; time = 0.0127s; samplesPerSecond = 201972.4
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2971-2980]: CE.SM = 2.11376953 * 2560; Err = 0.56757813 * 2560; time = 0.0121s; samplesPerSecond = 211903.0
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2981-2990]: CE.SM = 2.13481445 * 2560; Err = 0.57343750 * 2560; time = 0.0120s; samplesPerSecond = 213209.0
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[2991-3000]: CE.SM = 2.11572266 * 2560; Err = 0.57539063 * 2560; time = 0.0121s; samplesPerSecond = 210890.5
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3001-3010]: CE.SM = 2.15283203 * 2560; Err = 0.58828125 * 2560; time = 0.0120s; samplesPerSecond = 213351.1
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3011-3020]: CE.SM = 2.13154297 * 2560; Err = 0.56289062 * 2560; time = 0.0122s; samplesPerSecond = 210560.9
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3021-3030]: CE.SM = 2.09907227 * 2560; Err = 0.56718750 * 2560; time = 0.0122s; samplesPerSecond = 209013.7
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3031-3040]: CE.SM = 2.09106445 * 2560; Err = 0.56328125 * 2560; time = 0.0119s; samplesPerSecond = 215669.8
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3041-3050]: CE.SM = 2.13281250 * 2560; Err = 0.57421875 * 2560; time = 0.0121s; samplesPerSecond = 211920.5
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3051-3060]: CE.SM = 2.09887695 * 2560; Err = 0.56679687 * 2560; time = 0.0123s; samplesPerSecond = 208248.6
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3061-3070]: CE.SM = 2.14501953 * 2560; Err = 0.57382813 * 2560; time = 0.0123s; samplesPerSecond = 208231.7
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3071-3080]: CE.SM = 2.11035156 * 2560; Err = 0.56953125 * 2560; time = 0.0123s; samplesPerSecond = 207573.2
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3081-3090]: CE.SM = 2.04194336 * 2560; Err = 0.54492188 * 2560; time = 0.0122s; samplesPerSecond = 209561.2
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3091-3100]: CE.SM = 2.10239258 * 2560; Err = 0.57968750 * 2560; time = 0.0121s; samplesPerSecond = 211570.2
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3101-3110]: CE.SM = 2.12138672 * 2560; Err = 0.58398438 * 2560; time = 0.0121s; samplesPerSecond = 211482.9
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3111-3120]: CE.SM = 2.10771484 * 2560; Err = 0.57773438 * 2560; time = 0.0124s; samplesPerSecond = 206818.5
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3121-3130]: CE.SM = 2.12001953 * 2560; Err = 0.56562500 * 2560; time = 0.0122s; samplesPerSecond = 210405.2
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3131-3140]: CE.SM = 2.10805664 * 2560; Err = 0.56562500 * 2560; time = 0.0125s; samplesPerSecond = 204260.8
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3141-3150]: CE.SM = 2.12368164 * 2560; Err = 0.57109375 * 2560; time = 0.0122s; samplesPerSecond = 209526.9
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3151-3160]: CE.SM = 2.12758789 * 2560; Err = 0.56835938 * 2560; time = 0.0123s; samplesPerSecond = 207287.4
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3161-3170]: CE.SM = 2.07055664 * 2560; Err = 0.55390625 * 2560; time = 0.0125s; samplesPerSecond = 204032.8
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3171-3180]: CE.SM = 2.00722656 * 2560; Err = 0.54843750 * 2560; time = 0.0125s; samplesPerSecond = 204865.6
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3181-3190]: CE.SM = 2.09648438 * 2560; Err = 0.56718750 * 2560; time = 0.0127s; samplesPerSecond = 201829.1
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3191-3200]: CE.SM = 2.06552734 * 2560; Err = 0.56015625 * 2560; time = 0.0125s; samplesPerSecond = 205490.4
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3201-3210]: CE.SM = 2.04140625 * 2560; Err = 0.54960937 * 2560; time = 0.0125s; samplesPerSecond = 205276.2
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3211-3220]: CE.SM = 2.07949219 * 2560; Err = 0.55820313 * 2560; time = 0.0121s; samplesPerSecond = 211517.8
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3221-3230]: CE.SM = 2.08046875 * 2560; Err = 0.55546875 * 2560; time = 0.0124s; samplesPerSecond = 206885.4
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3231-3240]: CE.SM = 2.08535156 * 2560; Err = 0.56601563 * 2560; time = 0.0116s; samplesPerSecond = 221587.5
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3241-3250]: CE.SM = 2.10986328 * 2560; Err = 0.58007812 * 2560; time = 0.0459s; samplesPerSecond = 55780.7
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3251-3260]: CE.SM = 2.07041016 * 2560; Err = 0.57773438 * 2560; time = 0.0128s; samplesPerSecond = 199672.4
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3261-3270]: CE.SM = 2.11689453 * 2560; Err = 0.58437500 * 2560; time = 0.0114s; samplesPerSecond = 224857.3
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3271-3280]: CE.SM = 2.04746094 * 2560; Err = 0.55937500 * 2560; time = 0.0116s; samplesPerSecond = 220366.7
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3281-3290]: CE.SM = 2.09833984 * 2560; Err = 0.56054688 * 2560; time = 0.0136s; samplesPerSecond = 187903.7
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3291-3300]: CE.SM = 2.10693359 * 2560; Err = 0.56953125 * 2560; time = 0.0137s; samplesPerSecond = 187298.8
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3301-3310]: CE.SM = 2.11972656 * 2560; Err = 0.57031250 * 2560; time = 0.0147s; samplesPerSecond = 173677.1
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3311-3320]: CE.SM = 2.08750000 * 2560; Err = 0.56289062 * 2560; time = 0.0146s; samplesPerSecond = 175872.5
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3321-3330]: CE.SM = 2.09931641 * 2560; Err = 0.57343750 * 2560; time = 0.0142s; samplesPerSecond = 180650.6
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3331-3340]: CE.SM = 2.04619141 * 2560; Err = 0.56054688 * 2560; time = 0.0131s; samplesPerSecond = 195330.4
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3341-3350]: CE.SM = 2.05039062 * 2560; Err = 0.55234375 * 2560; time = 0.0135s; samplesPerSecond = 189181.2
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3351-3360]: CE.SM = 2.08603516 * 2560; Err = 0.56679687 * 2560; time = 0.0129s; samplesPerSecond = 198004.5
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3361-3370]: CE.SM = 2.08281250 * 2560; Err = 0.57031250 * 2560; time = 0.0130s; samplesPerSecond = 196605.5
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3371-3380]: CE.SM = 2.07226562 * 2560; Err = 0.55781250 * 2560; time = 0.0134s; samplesPerSecond = 191602.4
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3381-3390]: CE.SM = 2.10009766 * 2560; Err = 0.56679687 * 2560; time = 0.0128s; samplesPerSecond = 199532.3
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3391-3400]: CE.SM = 2.11347656 * 2560; Err = 0.58515625 * 2560; time = 0.0136s; samplesPerSecond = 188110.8
12/20/2016 15:26:49:  Epoch[ 1 of 2]-Minibatch[3401-3410]: CE.SM = 2.01777344 * 2560; Err = 0.55937500 * 2560; time = 0.0110s; samplesPerSecond = 233768.6
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3411-3420]: CE.SM = 2.08798828 * 2560; Err = 0.56523437 * 2560; time = 0.0129s; samplesPerSecond = 198142.4
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3421-3430]: CE.SM = 2.06347656 * 2560; Err = 0.55429688 * 2560; time = 0.0132s; samplesPerSecond = 194381.2
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3431-3440]: CE.SM = 2.07470703 * 2560; Err = 0.55117187 * 2560; time = 0.0126s; samplesPerSecond = 202435.6
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3441-3450]: CE.SM = 2.07568359 * 2560; Err = 0.57070312 * 2560; time = 0.0130s; samplesPerSecond = 196349.1
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3451-3460]: CE.SM = 2.06816406 * 2560; Err = 0.55468750 * 2560; time = 0.0130s; samplesPerSecond = 196469.7
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3461-3470]: CE.SM = 2.08457031 * 2560; Err = 0.56210938 * 2560; time = 0.0129s; samplesPerSecond = 198788.6
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3471-3480]: CE.SM = 2.04833984 * 2560; Err = 0.55859375 * 2560; time = 0.0125s; samplesPerSecond = 205210.4
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3481-3490]: CE.SM = 2.08906250 * 2560; Err = 0.58046875 * 2560; time = 0.0126s; samplesPerSecond = 203287.5
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3491-3500]: CE.SM = 2.04052734 * 2560; Err = 0.55351562 * 2560; time = 0.0124s; samplesPerSecond = 206835.3
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3501-3510]: CE.SM = 2.06376953 * 2560; Err = 0.54960937 * 2560; time = 0.0125s; samplesPerSecond = 205309.2
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3511-3520]: CE.SM = 2.05683594 * 2560; Err = 0.55390625 * 2560; time = 0.0132s; samplesPerSecond = 194130.6
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3521-3530]: CE.SM = 2.09013672 * 2560; Err = 0.56601563 * 2560; time = 0.0135s; samplesPerSecond = 190150.8
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3531-3540]: CE.SM = 2.02324219 * 2560; Err = 0.54531250 * 2560; time = 0.0127s; samplesPerSecond = 202211.7
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3541-3550]: CE.SM = 2.12216797 * 2560; Err = 0.58867187 * 2560; time = 0.0125s; samplesPerSecond = 204032.8
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3551-3560]: CE.SM = 2.06728516 * 2560; Err = 0.56640625 * 2560; time = 0.0126s; samplesPerSecond = 203449.1
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3561-3570]: CE.SM = 2.05234375 * 2560; Err = 0.55585938 * 2560; time = 0.0123s; samplesPerSecond = 207910.3
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3571-3580]: CE.SM = 2.01748047 * 2560; Err = 0.56210938 * 2560; time = 0.0127s; samplesPerSecond = 200847.3
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3581-3590]: CE.SM = 2.08105469 * 2560; Err = 0.56484375 * 2560; time = 0.0128s; samplesPerSecond = 199501.2
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3591-3600]: CE.SM = 2.02500000 * 2560; Err = 0.54179687 * 2560; time = 0.0107s; samplesPerSecond = 239655.5
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3601-3610]: CE.SM = 2.09843750 * 2560; Err = 0.57265625 * 2560; time = 0.0102s; samplesPerSecond = 250587.3
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3611-3620]: CE.SM = 2.02832031 * 2560; Err = 0.55390625 * 2560; time = 0.0116s; samplesPerSecond = 221338.4
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3621-3630]: CE.SM = 2.06582031 * 2560; Err = 0.56054688 * 2560; time = 0.0103s; samplesPerSecond = 248471.3
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3631-3640]: CE.SM = 2.08798828 * 2560; Err = 0.57343750 * 2560; time = 0.0107s; samplesPerSecond = 239812.6
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3641-3650]: CE.SM = 2.05722656 * 2560; Err = 0.56601563 * 2560; time = 0.0111s; samplesPerSecond = 229617.0
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3651-3660]: CE.SM = 1.98515625 * 2560; Err = 0.53984375 * 2560; time = 0.0100s; samplesPerSecond = 255872.1
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3661-3670]: CE.SM = 2.03339844 * 2560; Err = 0.56562500 * 2560; time = 0.0106s; samplesPerSecond = 240533.7
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3671-3680]: CE.SM = 2.05312500 * 2560; Err = 0.56523437 * 2560; time = 0.0107s; samplesPerSecond = 239745.3
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3681-3690]: CE.SM = 2.05224609 * 2560; Err = 0.56718750 * 2560; time = 0.0115s; samplesPerSecond = 222802.4
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3691-3700]: CE.SM = 2.00039063 * 2560; Err = 0.54453125 * 2560; time = 0.0124s; samplesPerSecond = 207119.7
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3701-3710]: CE.SM = 2.12207031 * 2560; Err = 0.58515625 * 2560; time = 0.0124s; samplesPerSecond = 206584.9
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3711-3720]: CE.SM = 2.06455078 * 2560; Err = 0.56523437 * 2560; time = 0.0125s; samplesPerSecond = 205506.9
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3721-3730]: CE.SM = 2.03144531 * 2560; Err = 0.55312500 * 2560; time = 0.0123s; samplesPerSecond = 208011.7
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3731-3740]: CE.SM = 2.02548828 * 2560; Err = 0.56601563 * 2560; time = 0.0126s; samplesPerSecond = 203562.3
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3741-3750]: CE.SM = 1.99169922 * 2560; Err = 0.54648438 * 2560; time = 0.0121s; samplesPerSecond = 211955.6
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3751-3760]: CE.SM = 2.05332031 * 2560; Err = 0.56289062 * 2560; time = 0.0123s; samplesPerSecond = 208282.5
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3761-3770]: CE.SM = 2.02763672 * 2560; Err = 0.55859375 * 2560; time = 0.0122s; samplesPerSecond = 209287.1
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3771-3780]: CE.SM = 2.04130859 * 2560; Err = 0.56210938 * 2560; time = 0.0128s; samplesPerSecond = 200328.7
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3781-3790]: CE.SM = 2.09101562 * 2560; Err = 0.56953125 * 2560; time = 0.0129s; samplesPerSecond = 197759.8
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3791-3800]: CE.SM = 2.08935547 * 2560; Err = 0.56601563 * 2560; time = 0.0122s; samplesPerSecond = 209116.2
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3801-3810]: CE.SM = 2.05117188 * 2560; Err = 0.56718750 * 2560; time = 0.0124s; samplesPerSecond = 206285.3
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3811-3820]: CE.SM = 2.06044922 * 2560; Err = 0.56992188 * 2560; time = 0.0125s; samplesPerSecond = 205177.5
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3821-3830]: CE.SM = 2.05097656 * 2560; Err = 0.56289062 * 2560; time = 0.0123s; samplesPerSecond = 208079.3
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3831-3840]: CE.SM = 2.02744141 * 2560; Err = 0.55390625 * 2560; time = 0.0122s; samplesPerSecond = 210059.9
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3841-3850]: CE.SM = 2.04482422 * 2560; Err = 0.55585938 * 2560; time = 0.0119s; samplesPerSecond = 215343.2
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3851-3860]: CE.SM = 1.99902344 * 2560; Err = 0.54648438 * 2560; time = 0.0122s; samplesPerSecond = 210630.2
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3861-3870]: CE.SM = 1.99208984 * 2560; Err = 0.54492188 * 2560; time = 0.0126s; samplesPerSecond = 202547.7
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3871-3880]: CE.SM = 2.02431641 * 2560; Err = 0.54921875 * 2560; time = 0.0124s; samplesPerSecond = 206385.0
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3881-3890]: CE.SM = 2.03388672 * 2560; Err = 0.54960937 * 2560; time = 0.0122s; samplesPerSecond = 210146.1
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3891-3900]: CE.SM = 2.04580078 * 2560; Err = 0.56093750 * 2560; time = 0.0128s; samplesPerSecond = 200768.6
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3901-3910]: CE.SM = 2.07919922 * 2560; Err = 0.57695312 * 2560; time = 0.0121s; samplesPerSecond = 212148.8
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3911-3920]: CE.SM = 2.05058594 * 2560; Err = 0.56562500 * 2560; time = 0.0126s; samplesPerSecond = 203821.7
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3921-3930]: CE.SM = 2.02421875 * 2560; Err = 0.55507812 * 2560; time = 0.0125s; samplesPerSecond = 204423.9
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3931-3940]: CE.SM = 2.01181641 * 2560; Err = 0.55078125 * 2560; time = 0.0120s; samplesPerSecond = 213155.7
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3941-3950]: CE.SM = 1.99560547 * 2560; Err = 0.55703125 * 2560; time = 0.0121s; samplesPerSecond = 211047.0
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3951-3960]: CE.SM = 1.99824219 * 2560; Err = 0.54648438 * 2560; time = 0.0122s; samplesPerSecond = 209304.2
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3961-3970]: CE.SM = 2.02773438 * 2560; Err = 0.56093750 * 2560; time = 0.0125s; samplesPerSecond = 205490.4
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3971-3980]: CE.SM = 2.06162109 * 2560; Err = 0.56250000 * 2560; time = 0.0120s; samplesPerSecond = 213102.5
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3981-3990]: CE.SM = 1.99873047 * 2560; Err = 0.54687500 * 2560; time = 0.0123s; samplesPerSecond = 208316.4
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[3991-4000]: CE.SM = 2.01523437 * 2560; Err = 0.54765625 * 2560; time = 0.0129s; samplesPerSecond = 197897.3
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[4001-4010]: CE.SM = 2.04921875 * 2560; Err = 0.56562500 * 2560; time = 0.0124s; samplesPerSecond = 205887.1
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[4011-4020]: CE.SM = 2.01894531 * 2560; Err = 0.54726562 * 2560; time = 0.0127s; samplesPerSecond = 201163.0
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[4021-4030]: CE.SM = 2.01220703 * 2560; Err = 0.54531250 * 2560; time = 0.0125s; samplesPerSecond = 204734.5
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[4031-4040]: CE.SM = 2.02353516 * 2560; Err = 0.54921875 * 2560; time = 0.0126s; samplesPerSecond = 203594.7
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[4041-4050]: CE.SM = 2.02392578 * 2560; Err = 0.54843750 * 2560; time = 0.0123s; samplesPerSecond = 208384.2
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[4051-4060]: CE.SM = 1.99833984 * 2560; Err = 0.54726562 * 2560; time = 0.0124s; samplesPerSecond = 207220.3
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[4061-4070]: CE.SM = 2.04257812 * 2560; Err = 0.56718750 * 2560; time = 0.0123s; samplesPerSecond = 208792.1
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[4071-4080]: CE.SM = 1.98417969 * 2560; Err = 0.55273438 * 2560; time = 0.0125s; samplesPerSecond = 204538.2
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[4081-4090]: CE.SM = 1.97685547 * 2560; Err = 0.53906250 * 2560; time = 0.0122s; samplesPerSecond = 209338.5
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[4091-4100]: CE.SM = 1.98564453 * 2560; Err = 0.54609375 * 2560; time = 0.0125s; samplesPerSecond = 205523.4
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[4101-4110]: CE.SM = 1.99140625 * 2560; Err = 0.54765625 * 2560; time = 0.0123s; samplesPerSecond = 208911.4
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[4111-4120]: CE.SM = 2.02128906 * 2560; Err = 0.55351562 * 2560; time = 0.0121s; samplesPerSecond = 211012.2
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[4121-4130]: CE.SM = 1.97724609 * 2560; Err = 0.55000000 * 2560; time = 0.0121s; samplesPerSecond = 211973.2
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[4131-4140]: CE.SM = 2.01357422 * 2560; Err = 0.54648438 * 2560; time = 0.0119s; samplesPerSecond = 215706.1
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[4141-4150]: CE.SM = 2.00585938 * 2560; Err = 0.54179687 * 2560; time = 0.0119s; samplesPerSecond = 215906.2
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[4151-4160]: CE.SM = 2.07685547 * 2560; Err = 0.56992188 * 2560; time = 0.0117s; samplesPerSecond = 217909.4
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[4161-4170]: CE.SM = 2.00312500 * 2560; Err = 0.55156250 * 2560; time = 0.0119s; samplesPerSecond = 214981.5
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[4171-4180]: CE.SM = 1.94843750 * 2560; Err = 0.53867188 * 2560; time = 0.0117s; samplesPerSecond = 218299.7
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[4181-4190]: CE.SM = 1.99707031 * 2560; Err = 0.53828125 * 2560; time = 0.0119s; samplesPerSecond = 215126.1
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[4191-4200]: CE.SM = 1.99082031 * 2560; Err = 0.55781250 * 2560; time = 0.0120s; samplesPerSecond = 213707.3
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[4201-4210]: CE.SM = 1.98212891 * 2560; Err = 0.54062500 * 2560; time = 0.0127s; samplesPerSecond = 202163.8
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[4211-4220]: CE.SM = 2.03349609 * 2560; Err = 0.57070312 * 2560; time = 0.0127s; samplesPerSecond = 201701.9
12/20/2016 15:26:50:  Epoch[ 1 of 2]-Minibatch[4221-4230]: CE.SM = 1.99619141 * 2560; Err = 0.55234375 * 2560; time = 0.0109s; samplesPerSecond = 234582.6
12/20/2016 15:26:51:  Epoch[ 1 of 2]-Minibatch[4231-4240]: CE.SM = 1.96523438 * 2560; Err = 0.53203125 * 2560; time = 0.0121s; samplesPerSecond = 210977.4
12/20/2016 15:26:51:  Epoch[ 1 of 2]-Minibatch[4241-4250]: CE.SM = 2.01435547 * 2560; Err = 0.53437500 * 2560; time = 0.0119s; samplesPerSecond = 214675.1
12/20/2016 15:26:51:  Epoch[ 1 of 2]-Minibatch[4251-4260]: CE.SM = 2.02294922 * 2560; Err = 0.55156250 * 2560; time = 0.0121s; samplesPerSecond = 210838.4
12/20/2016 15:26:51:  Epoch[ 1 of 2]-Minibatch[4261-4270]: CE.SM = 2.03261719 * 2560; Err = 0.54882812 * 2560; time = 0.0127s; samplesPerSecond = 201876.8
12/20/2016 15:26:51:  Epoch[ 1 of 2]-Minibatch[4271-4280]: CE.SM = 1.99765625 * 2560; Err = 0.56015625 * 2560; time = 0.0119s; samplesPerSecond = 214495.2
12/20/2016 15:26:51:  Epoch[ 1 of 2]-Minibatch[4281-4290]: CE.SM = 2.02763672 * 2560; Err = 0.54648438 * 2560; time = 0.0122s; samplesPerSecond = 210077.1
12/20/2016 15:26:51:  Epoch[ 1 of 2]-Minibatch[4291-4300]: CE.SM = 1.97783203 * 2560; Err = 0.55546875 * 2560; time = 0.0121s; samplesPerSecond = 211570.2
12/20/2016 15:26:51:  Epoch[ 1 of 2]-Minibatch[4301-4310]: CE.SM = 2.00810547 * 2560; Err = 0.55664062 * 2560; time = 0.0119s; samplesPerSecond = 215597.1
12/20/2016 15:26:51:  Epoch[ 1 of 2]-Minibatch[4311-4320]: CE.SM = 2.00810547 * 2560; Err = 0.54335937 * 2560; time = 0.0122s; samplesPerSecond = 209767.3
12/20/2016 15:26:51:  Epoch[ 1 of 2]-Minibatch[4321-4330]: CE.SM = 1.95957031 * 2560; Err = 0.53671875 * 2560; time = 0.0121s; samplesPerSecond = 211395.5
12/20/2016 15:26:51:  Epoch[ 1 of 2]-Minibatch[4331-4340]: CE.SM = 1.97861328 * 2560; Err = 0.55820313 * 2560; time = 0.0126s; samplesPerSecond = 203708.1
12/20/2016 15:26:51:  Epoch[ 1 of 2]-Minibatch[4341-4350]: CE.SM = 2.03222656 * 2560; Err = 0.57343750 * 2560; time = 0.0126s; samplesPerSecond = 202772.3
12/20/2016 15:26:51:  Epoch[ 1 of 2]-Minibatch[4351-4360]: CE.SM = 1.97734375 * 2560; Err = 0.55468750 * 2560; time = 0.0126s; samplesPerSecond = 202419.5
12/20/2016 15:26:51:  Epoch[ 1 of 2]-Minibatch[4361-4370]: CE.SM = 2.00136719 * 2560; Err = 0.55742187 * 2560; time = 0.0121s; samplesPerSecond = 211552.8
12/20/2016 15:26:51:  Epoch[ 1 of 2]-Minibatch[4371-4380]: CE.SM = 1.99404297 * 2560; Err = 0.54492188 * 2560; time = 0.0123s; samplesPerSecond = 208673.0
12/20/2016 15:26:51:  Epoch[ 1 of 2]-Minibatch[4381-4390]: CE.SM = 1.97587891 * 2560; Err = 0.53828125 * 2560; time = 0.0127s; samplesPerSecond = 200815.8
12/20/2016 15:26:51: Finished Epoch[ 1 of 2]: [Training] CE.SM = 2.42878280 * 1124823; Err = 0.62160980 * 1124823; totalSamplesSeen = 1124823; learningRatePerSample = 0.00039062501; epochTime=5.30028s
12/20/2016 15:26:51: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/dptmodel1/cntkSpeech.dnn.1'

12/20/2016 15:26:51: Starting Epoch 2: learning rate per sample = 0.000391  effective momentum = 0.900000  momentum as time constant = 2429.8 samples

12/20/2016 15:26:51: Starting minibatch loop.
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[   1-  10, 0.23%]: CE.SM = 2.00334606 * 2560; Err = 0.55546875 * 2560; time = 0.0118s; samplesPerSecond = 216857.3
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[  11-  20, 0.46%]: CE.SM = 1.95379009 * 2560; Err = 0.54101562 * 2560; time = 0.0126s; samplesPerSecond = 203967.8
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[  21-  30, 0.68%]: CE.SM = 2.01865158 * 2560; Err = 0.55234375 * 2560; time = 0.0122s; samplesPerSecond = 209853.3
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[  31-  40, 0.91%]: CE.SM = 1.99790688 * 2560; Err = 0.54921875 * 2560; time = 0.0119s; samplesPerSecond = 214387.4
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[  41-  50, 1.14%]: CE.SM = 1.94115295 * 2560; Err = 0.53242188 * 2560; time = 0.0124s; samplesPerSecond = 206119.2
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[  51-  60, 1.37%]: CE.SM = 1.97341919 * 2560; Err = 0.54453125 * 2560; time = 0.0128s; samplesPerSecond = 199812.7
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[  61-  70, 1.59%]: CE.SM = 2.02834549 * 2560; Err = 0.55390625 * 2560; time = 0.0126s; samplesPerSecond = 202692.0
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[  71-  80, 1.82%]: CE.SM = 1.93781738 * 2560; Err = 0.54765625 * 2560; time = 0.0130s; samplesPerSecond = 197150.6
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[  81-  90, 2.05%]: CE.SM = 1.92423096 * 2560; Err = 0.54101562 * 2560; time = 0.0125s; samplesPerSecond = 204163.0
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[  91- 100, 2.28%]: CE.SM = 1.96589355 * 2560; Err = 0.54804688 * 2560; time = 0.0124s; samplesPerSecond = 206285.3
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 101- 110, 2.51%]: CE.SM = 1.95301514 * 2560; Err = 0.53945312 * 2560; time = 0.0127s; samplesPerSecond = 202004.3
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 111- 120, 2.73%]: CE.SM = 1.95447388 * 2560; Err = 0.54687500 * 2560; time = 0.0124s; samplesPerSecond = 205953.3
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 121- 130, 2.96%]: CE.SM = 1.99099274 * 2560; Err = 0.54570312 * 2560; time = 0.0121s; samplesPerSecond = 211465.4
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 131- 140, 3.19%]: CE.SM = 1.99067993 * 2560; Err = 0.54960937 * 2560; time = 0.0126s; samplesPerSecond = 202981.3
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 141- 150, 3.42%]: CE.SM = 1.92315369 * 2560; Err = 0.52187500 * 2560; time = 0.0124s; samplesPerSecond = 205804.3
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 151- 160, 3.64%]: CE.SM = 1.97778015 * 2560; Err = 0.54414063 * 2560; time = 0.0129s; samplesPerSecond = 199144.3
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 161- 170, 3.87%]: CE.SM = 2.02292786 * 2560; Err = 0.55546875 * 2560; time = 0.0127s; samplesPerSecond = 201210.4
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 171- 180, 4.10%]: CE.SM = 2.01116028 * 2560; Err = 0.54101562 * 2560; time = 0.0125s; samplesPerSecond = 204832.8
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 181- 190, 4.33%]: CE.SM = 1.99359131 * 2560; Err = 0.56210938 * 2560; time = 0.0126s; samplesPerSecond = 203562.3
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 191- 200, 4.56%]: CE.SM = 1.96360779 * 2560; Err = 0.54531250 * 2560; time = 0.0125s; samplesPerSecond = 204652.7
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 201- 210, 4.78%]: CE.SM = 1.99527283 * 2560; Err = 0.54843750 * 2560; time = 0.0123s; samplesPerSecond = 207792.2
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 211- 220, 5.01%]: CE.SM = 1.95928955 * 2560; Err = 0.54140625 * 2560; time = 0.0124s; samplesPerSecond = 206285.3
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 221- 230, 5.24%]: CE.SM = 1.98595886 * 2560; Err = 0.54804688 * 2560; time = 0.0124s; samplesPerSecond = 206318.5
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 231- 240, 5.47%]: CE.SM = 1.99875183 * 2560; Err = 0.54921875 * 2560; time = 0.0124s; samplesPerSecond = 206534.9
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 241- 250, 5.69%]: CE.SM = 1.98141785 * 2560; Err = 0.54414063 * 2560; time = 0.0127s; samplesPerSecond = 201084.0
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 251- 260, 5.92%]: CE.SM = 1.97168274 * 2560; Err = 0.53906250 * 2560; time = 0.0124s; samplesPerSecond = 205771.2
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 261- 270, 6.15%]: CE.SM = 1.97962036 * 2560; Err = 0.55312500 * 2560; time = 0.0128s; samplesPerSecond = 199953.1
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 271- 280, 6.38%]: CE.SM = 1.89807739 * 2560; Err = 0.52226562 * 2560; time = 0.0125s; samplesPerSecond = 205539.9
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 281- 290, 6.61%]: CE.SM = 1.96758423 * 2560; Err = 0.54843750 * 2560; time = 0.0124s; samplesPerSecond = 206735.0
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 291- 300, 6.83%]: CE.SM = 2.01719360 * 2560; Err = 0.55507812 * 2560; time = 0.0122s; samplesPerSecond = 210042.7
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 301- 310, 7.06%]: CE.SM = 1.97598267 * 2560; Err = 0.54101562 * 2560; time = 0.0121s; samplesPerSecond = 211081.8
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 311- 320, 7.29%]: CE.SM = 1.97039185 * 2560; Err = 0.54335937 * 2560; time = 0.0120s; samplesPerSecond = 214154.3
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 321- 330, 7.52%]: CE.SM = 1.98658447 * 2560; Err = 0.54335937 * 2560; time = 0.0122s; samplesPerSecond = 210474.4
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 331- 340, 7.74%]: CE.SM = 1.95869141 * 2560; Err = 0.54804688 * 2560; time = 0.0121s; samplesPerSecond = 212412.9
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 341- 350, 7.97%]: CE.SM = 2.00740356 * 2560; Err = 0.55664062 * 2560; time = 0.0120s; samplesPerSecond = 212748.3
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 351- 360, 8.20%]: CE.SM = 1.97414551 * 2560; Err = 0.53867188 * 2560; time = 0.0119s; samplesPerSecond = 214639.1
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 361- 370, 8.43%]: CE.SM = 1.96647339 * 2560; Err = 0.53789062 * 2560; time = 0.0120s; samplesPerSecond = 213832.3
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 371- 380, 8.66%]: CE.SM = 1.99505615 * 2560; Err = 0.55039063 * 2560; time = 0.0121s; samplesPerSecond = 211134.0
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 381- 390, 8.88%]: CE.SM = 1.90538330 * 2560; Err = 0.53710938 * 2560; time = 0.0119s; samplesPerSecond = 215488.2
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 391- 400, 9.11%]: CE.SM = 1.96569214 * 2560; Err = 0.55078125 * 2560; time = 0.0120s; samplesPerSecond = 213760.9
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 401- 410, 9.34%]: CE.SM = 1.98982544 * 2560; Err = 0.54453125 * 2560; time = 0.0121s; samplesPerSecond = 212043.4
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 411- 420, 9.57%]: CE.SM = 2.03132935 * 2560; Err = 0.55976563 * 2560; time = 0.0120s; samplesPerSecond = 213903.7
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 421- 430, 9.79%]: CE.SM = 1.97291870 * 2560; Err = 0.54687500 * 2560; time = 0.0121s; samplesPerSecond = 211657.7
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 431- 440, 10.02%]: CE.SM = 1.97991943 * 2560; Err = 0.54453125 * 2560; time = 0.0120s; samplesPerSecond = 214011.0
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 441- 450, 10.25%]: CE.SM = 1.90657959 * 2560; Err = 0.52343750 * 2560; time = 0.0120s; samplesPerSecond = 213921.6
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 451- 460, 10.48%]: CE.SM = 1.93512573 * 2560; Err = 0.54882812 * 2560; time = 0.0113s; samplesPerSecond = 226749.3
12/20/2016 15:26:51:  Epoch[ 2 of 2]-Minibatch[ 461- 470, 10.71%]: CE.SM = 1.89949951 * 2560; Err = 0.53085938 * 2560; time = 0.0118s; samplesPerSecond = 217354.4
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 471- 480, 10.93%]: CE.SM = 1.92949829 * 2560; Err = 0.52265625 * 2560; time = 0.0122s; samplesPerSecond = 210111.6
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 481- 490, 11.16%]: CE.SM = 1.97517090 * 2560; Err = 0.54570312 * 2560; time = 0.0123s; samplesPerSecond = 207421.8
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 491- 500, 11.39%]: CE.SM = 1.94895020 * 2560; Err = 0.53359375 * 2560; time = 0.0123s; samplesPerSecond = 208843.2
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 501- 510, 11.62%]: CE.SM = 1.91326904 * 2560; Err = 0.52578125 * 2560; time = 0.0120s; samplesPerSecond = 213743.0
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 511- 520, 11.85%]: CE.SM = 1.97700195 * 2560; Err = 0.54687500 * 2560; time = 0.0124s; samplesPerSecond = 206119.2
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 521- 530, 12.07%]: CE.SM = 1.91873779 * 2560; Err = 0.53046875 * 2560; time = 0.0121s; samplesPerSecond = 211116.6
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 531- 540, 12.30%]: CE.SM = 1.92512207 * 2560; Err = 0.53085938 * 2560; time = 0.0115s; samplesPerSecond = 222938.3
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 541- 550, 12.53%]: CE.SM = 1.94018555 * 2560; Err = 0.53867188 * 2560; time = 0.0121s; samplesPerSecond = 211570.2
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 551- 560, 12.76%]: CE.SM = 1.95364990 * 2560; Err = 0.53320312 * 2560; time = 0.0120s; samplesPerSecond = 213957.4
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 561- 570, 12.98%]: CE.SM = 1.97687988 * 2560; Err = 0.56132812 * 2560; time = 0.0120s; samplesPerSecond = 213297.8
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 571- 580, 13.21%]: CE.SM = 1.96450195 * 2560; Err = 0.54257813 * 2560; time = 0.0120s; samplesPerSecond = 212554.0
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 581- 590, 13.44%]: CE.SM = 1.93133545 * 2560; Err = 0.53203125 * 2560; time = 0.0121s; samplesPerSecond = 212254.4
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 591- 600, 13.67%]: CE.SM = 1.93786621 * 2560; Err = 0.53984375 * 2560; time = 0.0121s; samplesPerSecond = 212148.8
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 601- 610, 13.90%]: CE.SM = 2.00069580 * 2560; Err = 0.54687500 * 2560; time = 0.0119s; samplesPerSecond = 214495.2
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 611- 620, 14.12%]: CE.SM = 1.94119873 * 2560; Err = 0.54140625 * 2560; time = 0.0119s; samplesPerSecond = 214711.1
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 621- 630, 14.35%]: CE.SM = 1.92653809 * 2560; Err = 0.52773437 * 2560; time = 0.0120s; samplesPerSecond = 213226.7
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 631- 640, 14.58%]: CE.SM = 1.97145996 * 2560; Err = 0.54375000 * 2560; time = 0.0121s; samplesPerSecond = 210925.3
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 641- 650, 14.81%]: CE.SM = 1.91876221 * 2560; Err = 0.53125000 * 2560; time = 0.0120s; samplesPerSecond = 213529.1
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 651- 660, 15.03%]: CE.SM = 1.96845703 * 2560; Err = 0.53554687 * 2560; time = 0.0120s; samplesPerSecond = 214064.7
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 661- 670, 15.26%]: CE.SM = 1.93986816 * 2560; Err = 0.52812500 * 2560; time = 0.0120s; samplesPerSecond = 213031.5
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 671- 680, 15.49%]: CE.SM = 1.96250000 * 2560; Err = 0.55468750 * 2560; time = 0.0122s; samplesPerSecond = 210578.3
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 681- 690, 15.72%]: CE.SM = 1.91948242 * 2560; Err = 0.53984375 * 2560; time = 0.0121s; samplesPerSecond = 211640.2
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 691- 700, 15.95%]: CE.SM = 1.90634766 * 2560; Err = 0.53320312 * 2560; time = 0.0122s; samplesPerSecond = 210353.3
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 701- 710, 16.17%]: CE.SM = 1.91823730 * 2560; Err = 0.55390625 * 2560; time = 0.0120s; samplesPerSecond = 212978.4
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 711- 720, 16.40%]: CE.SM = 2.01961670 * 2560; Err = 0.54335937 * 2560; time = 0.0121s; samplesPerSecond = 211308.3
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 721- 730, 16.63%]: CE.SM = 1.92088623 * 2560; Err = 0.52031250 * 2560; time = 0.0121s; samplesPerSecond = 211064.4
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 731- 740, 16.86%]: CE.SM = 1.91947021 * 2560; Err = 0.54062500 * 2560; time = 0.0121s; samplesPerSecond = 210734.3
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 741- 750, 17.08%]: CE.SM = 1.93861084 * 2560; Err = 0.54140625 * 2560; time = 0.0121s; samplesPerSecond = 211710.2
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 751- 760, 17.31%]: CE.SM = 1.92675781 * 2560; Err = 0.53437500 * 2560; time = 0.0120s; samplesPerSecond = 213475.7
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 761- 770, 17.54%]: CE.SM = 1.99114990 * 2560; Err = 0.55664062 * 2560; time = 0.0126s; samplesPerSecond = 203142.4
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 771- 780, 17.77%]: CE.SM = 1.91751709 * 2560; Err = 0.54140625 * 2560; time = 0.0131s; samplesPerSecond = 195748.6
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 781- 790, 18.00%]: CE.SM = 1.92917480 * 2560; Err = 0.53906250 * 2560; time = 0.0115s; samplesPerSecond = 223229.9
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 791- 800, 18.22%]: CE.SM = 1.92446289 * 2560; Err = 0.54453125 * 2560; time = 0.0124s; samplesPerSecond = 206102.6
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 801- 810, 18.45%]: CE.SM = 1.95371094 * 2560; Err = 0.54023438 * 2560; time = 0.0127s; samplesPerSecond = 201876.8
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 811- 820, 18.68%]: CE.SM = 1.92972412 * 2560; Err = 0.54882812 * 2560; time = 0.0128s; samplesPerSecond = 200705.6
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 821- 830, 18.91%]: CE.SM = 1.98690186 * 2560; Err = 0.54648438 * 2560; time = 0.0126s; samplesPerSecond = 203158.5
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 831- 840, 19.13%]: CE.SM = 1.93276367 * 2560; Err = 0.52148438 * 2560; time = 0.0125s; samplesPerSecond = 204685.4
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 841- 850, 19.36%]: CE.SM = 1.93708496 * 2560; Err = 0.54179687 * 2560; time = 0.0125s; samplesPerSecond = 204391.2
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 851- 860, 19.59%]: CE.SM = 1.91385498 * 2560; Err = 0.53437500 * 2560; time = 0.0125s; samplesPerSecond = 204538.2
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 861- 870, 19.82%]: CE.SM = 1.93806152 * 2560; Err = 0.52929688 * 2560; time = 0.0123s; samplesPerSecond = 207539.5
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 871- 880, 20.05%]: CE.SM = 1.92178955 * 2560; Err = 0.52109375 * 2560; time = 0.0129s; samplesPerSecond = 197744.5
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 881- 890, 20.27%]: CE.SM = 1.98505859 * 2560; Err = 0.54531250 * 2560; time = 0.0125s; samplesPerSecond = 204996.8
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 891- 900, 20.50%]: CE.SM = 1.94667969 * 2560; Err = 0.53671875 * 2560; time = 0.0130s; samplesPerSecond = 196983.7
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 901- 910, 20.73%]: CE.SM = 1.92806396 * 2560; Err = 0.54140625 * 2560; time = 0.0129s; samplesPerSecond = 197989.2
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 911- 920, 20.96%]: CE.SM = 1.91748047 * 2560; Err = 0.52578125 * 2560; time = 0.0124s; samplesPerSecond = 206885.4
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 921- 930, 21.18%]: CE.SM = 1.95489502 * 2560; Err = 0.54101562 * 2560; time = 0.0122s; samplesPerSecond = 209578.4
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 931- 940, 21.41%]: CE.SM = 1.92094727 * 2560; Err = 0.54687500 * 2560; time = 0.0127s; samplesPerSecond = 201321.2
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 941- 950, 21.64%]: CE.SM = 1.95830078 * 2560; Err = 0.54335937 * 2560; time = 0.0124s; samplesPerSecond = 205903.6
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 951- 960, 21.87%]: CE.SM = 1.90950928 * 2560; Err = 0.53085938 * 2560; time = 0.0129s; samplesPerSecond = 198188.4
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 961- 970, 22.10%]: CE.SM = 1.89929199 * 2560; Err = 0.52500000 * 2560; time = 0.0132s; samplesPerSecond = 194057.0
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 971- 980, 22.32%]: CE.SM = 1.92894287 * 2560; Err = 0.52617187 * 2560; time = 0.0120s; samplesPerSecond = 213868.0
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 981- 990, 22.55%]: CE.SM = 1.93760986 * 2560; Err = 0.53984375 * 2560; time = 0.0128s; samplesPerSecond = 200015.6
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[ 991-1000, 22.78%]: CE.SM = 1.92568359 * 2560; Err = 0.54375000 * 2560; time = 0.0130s; samplesPerSecond = 197059.5
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[1001-1010, 23.01%]: CE.SM = 1.92332764 * 2560; Err = 0.53789062 * 2560; time = 0.0128s; samplesPerSecond = 199330.4
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[1011-1020, 23.23%]: CE.SM = 1.91445313 * 2560; Err = 0.52539062 * 2560; time = 0.0124s; samplesPerSecond = 206835.3
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[1021-1030, 23.46%]: CE.SM = 1.90157471 * 2560; Err = 0.53671875 * 2560; time = 0.0127s; samplesPerSecond = 201432.1
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[1031-1040, 23.69%]: CE.SM = 1.88779297 * 2560; Err = 0.53671875 * 2560; time = 0.0127s; samplesPerSecond = 200957.7
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[1041-1050, 23.92%]: CE.SM = 1.95721436 * 2560; Err = 0.54570312 * 2560; time = 0.0129s; samplesPerSecond = 198958.6
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[1051-1060, 24.15%]: CE.SM = 1.92363281 * 2560; Err = 0.53593750 * 2560; time = 0.0128s; samplesPerSecond = 200469.9
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[1061-1070, 24.37%]: CE.SM = 1.94333496 * 2560; Err = 0.52539062 * 2560; time = 0.0130s; samplesPerSecond = 196273.9
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[1071-1080, 24.60%]: CE.SM = 1.96972656 * 2560; Err = 0.56054688 * 2560; time = 0.0129s; samplesPerSecond = 198711.5
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[1081-1090, 24.83%]: CE.SM = 1.91049805 * 2560; Err = 0.52773437 * 2560; time = 0.0127s; samplesPerSecond = 201654.2
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[1091-1100, 25.06%]: CE.SM = 1.95676270 * 2560; Err = 0.54570312 * 2560; time = 0.0126s; samplesPerSecond = 203870.4
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[1101-1110, 25.28%]: CE.SM = 1.92434082 * 2560; Err = 0.53242188 * 2560; time = 0.0120s; samplesPerSecond = 212606.9
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[1111-1120, 25.51%]: CE.SM = 1.91074219 * 2560; Err = 0.53671875 * 2560; time = 0.0122s; samplesPerSecond = 209064.9
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[1121-1130, 25.74%]: CE.SM = 1.92702637 * 2560; Err = 0.53906250 * 2560; time = 0.0123s; samplesPerSecond = 207539.5
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[1131-1140, 25.97%]: CE.SM = 1.92128906 * 2560; Err = 0.52812500 * 2560; time = 0.0122s; samplesPerSecond = 209544.1
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[1141-1150, 26.20%]: CE.SM = 1.92951660 * 2560; Err = 0.54257813 * 2560; time = 0.0126s; samplesPerSecond = 203643.3
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[1151-1160, 26.42%]: CE.SM = 1.94497070 * 2560; Err = 0.52812500 * 2560; time = 0.0122s; samplesPerSecond = 209767.3
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[1161-1170, 26.65%]: CE.SM = 1.82529297 * 2560; Err = 0.50703125 * 2560; time = 0.0126s; samplesPerSecond = 203578.5
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[1171-1180, 26.88%]: CE.SM = 1.90930176 * 2560; Err = 0.52734375 * 2560; time = 0.0126s; samplesPerSecond = 202435.6
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[1181-1190, 27.11%]: CE.SM = 1.96052246 * 2560; Err = 0.53671875 * 2560; time = 0.0123s; samplesPerSecond = 207674.2
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[1191-1200, 27.33%]: CE.SM = 1.90151367 * 2560; Err = 0.53359375 * 2560; time = 0.0124s; samplesPerSecond = 207069.5
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[1201-1210, 27.56%]: CE.SM = 1.98168945 * 2560; Err = 0.54882812 * 2560; time = 0.0125s; samplesPerSecond = 204244.5
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[1211-1220, 27.79%]: CE.SM = 1.93017578 * 2560; Err = 0.53515625 * 2560; time = 0.0123s; samplesPerSecond = 208554.0
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[1221-1230, 28.02%]: CE.SM = 1.88046875 * 2560; Err = 0.52070313 * 2560; time = 0.0128s; samplesPerSecond = 199376.9
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[1231-1240, 28.25%]: CE.SM = 1.95361328 * 2560; Err = 0.53867188 * 2560; time = 0.0129s; samplesPerSecond = 198280.5
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[1241-1250, 28.47%]: CE.SM = 1.90041504 * 2560; Err = 0.53320312 * 2560; time = 0.0125s; samplesPerSecond = 205474.0
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[1251-1260, 28.70%]: CE.SM = 1.90810547 * 2560; Err = 0.53554687 * 2560; time = 0.0125s; samplesPerSecond = 205457.5
12/20/2016 15:26:52:  Epoch[ 2 of 2]-Minibatch[1261-1270, 28.93%]: CE.SM = 1.90544434 * 2560; Err = 0.52343750 * 2560; time = 0.0118s; samplesPerSecond = 216582.1
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1271-1280, 29.16%]: CE.SM = 1.94497070 * 2560; Err = 0.54921875 * 2560; time = 0.0129s; samplesPerSecond = 198004.5
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1281-1290, 29.38%]: CE.SM = 1.94506836 * 2560; Err = 0.53671875 * 2560; time = 0.0128s; samplesPerSecond = 200438.5
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1291-1300, 29.61%]: CE.SM = 1.88215332 * 2560; Err = 0.52734375 * 2560; time = 0.0126s; samplesPerSecond = 202627.8
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1301-1310, 29.84%]: CE.SM = 1.96296387 * 2560; Err = 0.55390625 * 2560; time = 0.0126s; samplesPerSecond = 202788.3
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1311-1320, 30.07%]: CE.SM = 1.92441406 * 2560; Err = 0.53281250 * 2560; time = 0.0127s; samplesPerSecond = 200878.8
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1321-1330, 30.30%]: CE.SM = 1.86906738 * 2560; Err = 0.52578125 * 2560; time = 0.0126s; samplesPerSecond = 203708.1
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1331-1340, 30.52%]: CE.SM = 1.91938477 * 2560; Err = 0.51757812 * 2560; time = 0.0127s; samplesPerSecond = 201447.9
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1341-1350, 30.75%]: CE.SM = 1.93967285 * 2560; Err = 0.54101562 * 2560; time = 0.0125s; samplesPerSecond = 205259.8
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1351-1360, 30.98%]: CE.SM = 1.87001953 * 2560; Err = 0.52070313 * 2560; time = 0.0126s; samplesPerSecond = 203919.1
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1361-1370, 31.21%]: CE.SM = 1.89152832 * 2560; Err = 0.52578125 * 2560; time = 0.0127s; samplesPerSecond = 202163.8
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1371-1380, 31.44%]: CE.SM = 1.88312988 * 2560; Err = 0.52148438 * 2560; time = 0.0125s; samplesPerSecond = 204800.0
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1381-1390, 31.66%]: CE.SM = 1.90270996 * 2560; Err = 0.52851563 * 2560; time = 0.0125s; samplesPerSecond = 204931.2
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1391-1400, 31.89%]: CE.SM = 1.91232910 * 2560; Err = 0.51953125 * 2560; time = 0.0124s; samplesPerSecond = 205771.2
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1401-1410, 32.12%]: CE.SM = 1.90432129 * 2560; Err = 0.53554687 * 2560; time = 0.0125s; samplesPerSecond = 205144.6
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1411-1420, 32.35%]: CE.SM = 1.94909668 * 2560; Err = 0.54453125 * 2560; time = 0.0127s; samplesPerSecond = 201084.0
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1421-1430, 32.57%]: CE.SM = 1.94726562 * 2560; Err = 0.54804688 * 2560; time = 0.0124s; samplesPerSecond = 205903.6
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1431-1440, 32.80%]: CE.SM = 1.95043945 * 2560; Err = 0.54179687 * 2560; time = 0.0120s; samplesPerSecond = 213618.2
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1441-1450, 33.03%]: CE.SM = 1.87446289 * 2560; Err = 0.53398437 * 2560; time = 0.0123s; samplesPerSecond = 207505.9
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1451-1460, 33.26%]: CE.SM = 1.94697266 * 2560; Err = 0.54062500 * 2560; time = 0.0124s; samplesPerSecond = 207270.7
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1461-1470, 33.49%]: CE.SM = 1.92399902 * 2560; Err = 0.52695313 * 2560; time = 0.0120s; samplesPerSecond = 214011.0
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1471-1480, 33.71%]: CE.SM = 1.93967285 * 2560; Err = 0.52929688 * 2560; time = 0.0170s; samplesPerSecond = 150774.5
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1481-1490, 33.94%]: CE.SM = 1.91445313 * 2560; Err = 0.54179687 * 2560; time = 0.0129s; samplesPerSecond = 198388.1
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1491-1500, 34.17%]: CE.SM = 1.92492676 * 2560; Err = 0.53554687 * 2560; time = 0.0125s; samplesPerSecond = 204211.9
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1501-1510, 34.40%]: CE.SM = 1.94604492 * 2560; Err = 0.54765625 * 2560; time = 0.0120s; samplesPerSecond = 214172.2
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1511-1520, 34.62%]: CE.SM = 1.87524414 * 2560; Err = 0.52226562 * 2560; time = 0.0124s; samplesPerSecond = 206268.6
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1521-1530, 34.85%]: CE.SM = 1.87575684 * 2560; Err = 0.51757812 * 2560; time = 0.0125s; samplesPerSecond = 204277.1
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1531-1540, 35.08%]: CE.SM = 1.86340332 * 2560; Err = 0.52187500 * 2560; time = 0.0122s; samplesPerSecond = 209082.0
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1541-1550, 35.31%]: CE.SM = 1.88261719 * 2560; Err = 0.51875000 * 2560; time = 0.0127s; samplesPerSecond = 202084.0
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1551-1560, 35.54%]: CE.SM = 1.90952148 * 2560; Err = 0.52500000 * 2560; time = 0.0121s; samplesPerSecond = 211064.4
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1561-1570, 35.76%]: CE.SM = 1.85102539 * 2560; Err = 0.51679688 * 2560; time = 0.0122s; samplesPerSecond = 209956.5
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1571-1580, 35.99%]: CE.SM = 1.93891602 * 2560; Err = 0.54140625 * 2560; time = 0.0145s; samplesPerSecond = 176211.5
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1581-1590, 36.22%]: CE.SM = 1.93068848 * 2560; Err = 0.53085938 * 2560; time = 0.0120s; samplesPerSecond = 212854.4
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1591-1600, 36.45%]: CE.SM = 1.91975098 * 2560; Err = 0.53984375 * 2560; time = 0.0121s; samplesPerSecond = 210960.0
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1601-1610, 36.67%]: CE.SM = 1.86154785 * 2560; Err = 0.52929688 * 2560; time = 0.0120s; samplesPerSecond = 213262.2
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1611-1620, 36.90%]: CE.SM = 1.85810547 * 2560; Err = 0.50820312 * 2560; time = 0.0126s; samplesPerSecond = 203578.5
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1621-1630, 37.13%]: CE.SM = 1.90646973 * 2560; Err = 0.52304688 * 2560; time = 0.0124s; samplesPerSecond = 206634.9
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1631-1640, 37.36%]: CE.SM = 1.96494141 * 2560; Err = 0.54804688 * 2560; time = 0.0126s; samplesPerSecond = 203627.1
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1641-1650, 37.59%]: CE.SM = 1.90776367 * 2560; Err = 0.52031250 * 2560; time = 0.0121s; samplesPerSecond = 211500.3
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1651-1660, 37.81%]: CE.SM = 1.88945312 * 2560; Err = 0.52812500 * 2560; time = 0.0122s; samplesPerSecond = 209372.7
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1661-1670, 38.04%]: CE.SM = 1.88034668 * 2560; Err = 0.52460938 * 2560; time = 0.0126s; samplesPerSecond = 203465.3
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1671-1680, 38.27%]: CE.SM = 1.84389648 * 2560; Err = 0.50390625 * 2560; time = 0.0123s; samplesPerSecond = 207321.0
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1681-1690, 38.50%]: CE.SM = 1.90368652 * 2560; Err = 0.52578125 * 2560; time = 0.0121s; samplesPerSecond = 211482.9
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1691-1700, 38.72%]: CE.SM = 1.92539062 * 2560; Err = 0.54296875 * 2560; time = 0.0123s; samplesPerSecond = 207927.2
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1701-1710, 38.95%]: CE.SM = 1.84721680 * 2560; Err = 0.51367188 * 2560; time = 0.0123s; samplesPerSecond = 207893.5
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1711-1720, 39.18%]: CE.SM = 1.88515625 * 2560; Err = 0.53320312 * 2560; time = 0.0123s; samplesPerSecond = 208605.0
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1721-1730, 39.41%]: CE.SM = 1.91279297 * 2560; Err = 0.52382812 * 2560; time = 0.0122s; samplesPerSecond = 210111.6
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1731-1740, 39.64%]: CE.SM = 1.89794922 * 2560; Err = 0.54257813 * 2560; time = 0.0120s; samplesPerSecond = 213689.5
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1741-1750, 39.86%]: CE.SM = 1.90334473 * 2560; Err = 0.54101562 * 2560; time = 0.0121s; samplesPerSecond = 211151.4
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1751-1760, 40.09%]: CE.SM = 1.93254395 * 2560; Err = 0.54023438 * 2560; time = 0.0124s; samplesPerSecond = 207069.5
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1761-1770, 40.32%]: CE.SM = 1.88393555 * 2560; Err = 0.52343750 * 2560; time = 0.0127s; samplesPerSecond = 202259.6
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1771-1780, 40.55%]: CE.SM = 1.87016602 * 2560; Err = 0.51796875 * 2560; time = 0.0125s; samplesPerSecond = 204000.3
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1781-1790, 40.77%]: CE.SM = 1.91818848 * 2560; Err = 0.52929688 * 2560; time = 0.0121s; samplesPerSecond = 210751.6
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1791-1800, 41.00%]: CE.SM = 1.88271484 * 2560; Err = 0.52304688 * 2560; time = 0.0125s; samplesPerSecond = 204489.2
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1801-1810, 41.23%]: CE.SM = 1.83999023 * 2560; Err = 0.50898438 * 2560; time = 0.0128s; samplesPerSecond = 200391.4
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1811-1820, 41.46%]: CE.SM = 1.86228027 * 2560; Err = 0.52226562 * 2560; time = 0.0127s; samplesPerSecond = 201527.2
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1821-1830, 41.69%]: CE.SM = 1.91003418 * 2560; Err = 0.53320312 * 2560; time = 0.0127s; samplesPerSecond = 201813.2
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1831-1840, 41.91%]: CE.SM = 1.85129395 * 2560; Err = 0.51914063 * 2560; time = 0.0129s; samplesPerSecond = 197729.2
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1841-1850, 42.14%]: CE.SM = 1.84382324 * 2560; Err = 0.52500000 * 2560; time = 0.0125s; samplesPerSecond = 204980.4
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1851-1860, 42.37%]: CE.SM = 1.91240234 * 2560; Err = 0.53320312 * 2560; time = 0.0128s; samplesPerSecond = 199485.7
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1861-1870, 42.60%]: CE.SM = 1.89465332 * 2560; Err = 0.53867188 * 2560; time = 0.0125s; samplesPerSecond = 204130.5
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1871-1880, 42.82%]: CE.SM = 1.83962402 * 2560; Err = 0.52578125 * 2560; time = 0.0129s; samplesPerSecond = 197698.7
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1881-1890, 43.05%]: CE.SM = 1.89062500 * 2560; Err = 0.55078125 * 2560; time = 0.0125s; samplesPerSecond = 204114.2
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1891-1900, 43.28%]: CE.SM = 1.87995605 * 2560; Err = 0.52226562 * 2560; time = 0.0124s; samplesPerSecond = 206036.2
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1901-1910, 43.51%]: CE.SM = 1.82639160 * 2560; Err = 0.52539062 * 2560; time = 0.0125s; samplesPerSecond = 204489.2
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1911-1920, 43.74%]: CE.SM = 1.83681641 * 2560; Err = 0.51796875 * 2560; time = 0.0125s; samplesPerSecond = 204081.6
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1921-1930, 43.96%]: CE.SM = 1.84453125 * 2560; Err = 0.52265625 * 2560; time = 0.0147s; samplesPerSecond = 174303.8
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1931-1940, 44.19%]: CE.SM = 1.90651855 * 2560; Err = 0.54804688 * 2560; time = 0.0126s; samplesPerSecond = 202836.5
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1941-1950, 44.42%]: CE.SM = 1.86655273 * 2560; Err = 0.51562500 * 2560; time = 0.0128s; samplesPerSecond = 199921.9
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1951-1960, 44.65%]: CE.SM = 1.86147461 * 2560; Err = 0.51562500 * 2560; time = 0.0124s; samplesPerSecond = 206534.9
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1961-1970, 44.87%]: CE.SM = 1.91000977 * 2560; Err = 0.52773437 * 2560; time = 0.0125s; samplesPerSecond = 204146.7
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1971-1980, 45.10%]: CE.SM = 1.97180176 * 2560; Err = 0.55429688 * 2560; time = 0.0124s; samplesPerSecond = 207186.8
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1981-1990, 45.33%]: CE.SM = 1.89184570 * 2560; Err = 0.53242188 * 2560; time = 0.0123s; samplesPerSecond = 208231.7
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[1991-2000, 45.56%]: CE.SM = 1.91662598 * 2560; Err = 0.53046875 * 2560; time = 0.0129s; samplesPerSecond = 199159.8
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[2001-2010, 45.79%]: CE.SM = 1.87573242 * 2560; Err = 0.53085938 * 2560; time = 0.0124s; samplesPerSecond = 206285.3
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[2011-2020, 46.01%]: CE.SM = 1.87846680 * 2560; Err = 0.53750000 * 2560; time = 0.0124s; samplesPerSecond = 205721.6
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[2021-2030, 46.24%]: CE.SM = 1.88742676 * 2560; Err = 0.53554687 * 2560; time = 0.0125s; samplesPerSecond = 205622.5
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[2031-2040, 46.47%]: CE.SM = 1.86494141 * 2560; Err = 0.51015625 * 2560; time = 0.0126s; samplesPerSecond = 202997.4
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[2041-2050, 46.70%]: CE.SM = 1.91333008 * 2560; Err = 0.52656250 * 2560; time = 0.0129s; samplesPerSecond = 199066.9
12/20/2016 15:26:53:  Epoch[ 2 of 2]-Minibatch[2051-2060, 46.92%]: CE.SM = 1.91572266 * 2560; Err = 0.53710938 * 2560; time = 0.0113s; samplesPerSecond = 226468.5
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2061-2070, 47.15%]: CE.SM = 1.90705566 * 2560; Err = 0.52304688 * 2560; time = 0.0120s; samplesPerSecond = 213814.4
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2071-2080, 47.38%]: CE.SM = 1.90310059 * 2560; Err = 0.53476563 * 2560; time = 0.0120s; samplesPerSecond = 212925.2
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2081-2090, 47.61%]: CE.SM = 1.87597656 * 2560; Err = 0.52148438 * 2560; time = 0.0124s; samplesPerSecond = 206952.3
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2091-2100, 47.84%]: CE.SM = 1.86469727 * 2560; Err = 0.53710938 * 2560; time = 0.0118s; samplesPerSecond = 217742.6
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2101-2110, 48.06%]: CE.SM = 1.83996582 * 2560; Err = 0.50859375 * 2560; time = 0.0115s; samplesPerSecond = 221894.8
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2111-2120, 48.29%]: CE.SM = 1.88588867 * 2560; Err = 0.52382812 * 2560; time = 0.0120s; samplesPerSecond = 213868.0
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2121-2130, 48.52%]: CE.SM = 1.83378906 * 2560; Err = 0.51289063 * 2560; time = 0.0121s; samplesPerSecond = 212131.3
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2131-2140, 48.75%]: CE.SM = 1.89682617 * 2560; Err = 0.53554687 * 2560; time = 0.0121s; samplesPerSecond = 212131.3
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2141-2150, 48.97%]: CE.SM = 1.84282227 * 2560; Err = 0.52656250 * 2560; time = 0.0123s; samplesPerSecond = 208877.3
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2151-2160, 49.20%]: CE.SM = 1.89228516 * 2560; Err = 0.53281250 * 2560; time = 0.0125s; samplesPerSecond = 204163.0
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2161-2170, 49.43%]: CE.SM = 1.88515625 * 2560; Err = 0.52539062 * 2560; time = 0.0127s; samplesPerSecond = 201400.4
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2171-2180, 49.66%]: CE.SM = 1.94418945 * 2560; Err = 0.54179687 * 2560; time = 0.0127s; samplesPerSecond = 201163.0
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2181-2190, 49.89%]: CE.SM = 1.86586914 * 2560; Err = 0.52343750 * 2560; time = 0.0127s; samplesPerSecond = 202084.0
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2191-2200, 50.11%]: CE.SM = 1.83012695 * 2560; Err = 0.51367188 * 2560; time = 0.0126s; samplesPerSecond = 202547.7
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2201-2210, 50.34%]: CE.SM = 1.89677734 * 2560; Err = 0.51289063 * 2560; time = 0.0126s; samplesPerSecond = 202483.6
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2211-2220, 50.57%]: CE.SM = 1.90009766 * 2560; Err = 0.53906250 * 2560; time = 0.0136s; samplesPerSecond = 188401.5
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2221-2230, 50.80%]: CE.SM = 1.87587891 * 2560; Err = 0.52812500 * 2560; time = 0.0125s; samplesPerSecond = 204211.9
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2231-2240, 51.03%]: CE.SM = 1.85366211 * 2560; Err = 0.52070313 * 2560; time = 0.0128s; samplesPerSecond = 199968.8
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2241-2250, 51.25%]: CE.SM = 1.88876953 * 2560; Err = 0.54218750 * 2560; time = 0.0129s; samplesPerSecond = 199206.3
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2251-2260, 51.48%]: CE.SM = 1.93066406 * 2560; Err = 0.53281250 * 2560; time = 0.0129s; samplesPerSecond = 198927.7
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2261-2270, 51.71%]: CE.SM = 1.86967773 * 2560; Err = 0.52851563 * 2560; time = 0.0127s; samplesPerSecond = 201257.9
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2271-2280, 51.94%]: CE.SM = 1.89311523 * 2560; Err = 0.52148438 * 2560; time = 0.0129s; samplesPerSecond = 198342.0
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2281-2290, 52.16%]: CE.SM = 1.88627930 * 2560; Err = 0.52851563 * 2560; time = 0.0134s; samplesPerSecond = 191717.2
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2291-2300, 52.39%]: CE.SM = 1.83872070 * 2560; Err = 0.51718750 * 2560; time = 0.0128s; samplesPerSecond = 199750.3
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2301-2310, 52.62%]: CE.SM = 1.86181641 * 2560; Err = 0.51484375 * 2560; time = 0.0126s; samplesPerSecond = 202435.6
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2311-2320, 52.85%]: CE.SM = 1.85615234 * 2560; Err = 0.52500000 * 2560; time = 0.0130s; samplesPerSecond = 196650.8
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2321-2330, 53.08%]: CE.SM = 1.83613281 * 2560; Err = 0.51679688 * 2560; time = 0.0124s; samplesPerSecond = 205754.7
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2331-2340, 53.30%]: CE.SM = 1.86040039 * 2560; Err = 0.51093750 * 2560; time = 0.0127s; samplesPerSecond = 201813.2
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2341-2350, 53.53%]: CE.SM = 1.86142578 * 2560; Err = 0.52695313 * 2560; time = 0.0128s; samplesPerSecond = 200579.8
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2351-2360, 53.76%]: CE.SM = 1.82827148 * 2560; Err = 0.51523438 * 2560; time = 0.0126s; samplesPerSecond = 203756.8
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2361-2370, 53.99%]: CE.SM = 1.83261719 * 2560; Err = 0.51601562 * 2560; time = 0.0125s; samplesPerSecond = 205408.0
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2371-2380, 54.21%]: CE.SM = 1.86826172 * 2560; Err = 0.51914063 * 2560; time = 0.0124s; samplesPerSecond = 207103.0
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2381-2390, 54.44%]: CE.SM = 1.91005859 * 2560; Err = 0.53007812 * 2560; time = 0.0122s; samplesPerSecond = 209270.0
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2391-2400, 54.67%]: CE.SM = 1.87724609 * 2560; Err = 0.51289063 * 2560; time = 0.0123s; samplesPerSecond = 207421.8
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2401-2410, 54.90%]: CE.SM = 1.80898438 * 2560; Err = 0.52148438 * 2560; time = 0.0120s; samplesPerSecond = 212695.2
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2411-2420, 55.13%]: CE.SM = 1.81557617 * 2560; Err = 0.52265625 * 2560; time = 0.0126s; samplesPerSecond = 202820.5
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2421-2430, 55.35%]: CE.SM = 1.88564453 * 2560; Err = 0.53984375 * 2560; time = 0.0124s; samplesPerSecond = 205655.5
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2431-2440, 55.58%]: CE.SM = 1.84516602 * 2560; Err = 0.51601562 * 2560; time = 0.0123s; samplesPerSecond = 208045.5
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2441-2450, 55.81%]: CE.SM = 1.84545898 * 2560; Err = 0.51406250 * 2560; time = 0.0123s; samplesPerSecond = 208622.0
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2451-2460, 56.04%]: CE.SM = 1.87768555 * 2560; Err = 0.50312500 * 2560; time = 0.0123s; samplesPerSecond = 207825.9
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2461-2470, 56.26%]: CE.SM = 1.84604492 * 2560; Err = 0.51015625 * 2560; time = 0.0123s; samplesPerSecond = 207489.1
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2471-2480, 56.49%]: CE.SM = 1.85810547 * 2560; Err = 0.52382812 * 2560; time = 0.0121s; samplesPerSecond = 211325.7
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2481-2490, 56.72%]: CE.SM = 1.86171875 * 2560; Err = 0.52343750 * 2560; time = 0.0123s; samplesPerSecond = 208265.5
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2491-2500, 56.95%]: CE.SM = 1.87363281 * 2560; Err = 0.52851563 * 2560; time = 0.0122s; samplesPerSecond = 209407.0
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2501-2510, 57.18%]: CE.SM = 1.83774414 * 2560; Err = 0.51367188 * 2560; time = 0.0126s; samplesPerSecond = 203854.1
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2511-2520, 57.40%]: CE.SM = 1.88388672 * 2560; Err = 0.53710938 * 2560; time = 0.0120s; samplesPerSecond = 212501.0
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2521-2530, 57.63%]: CE.SM = 1.86962891 * 2560; Err = 0.51640625 * 2560; time = 0.0122s; samplesPerSecond = 209561.2
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2531-2540, 57.86%]: CE.SM = 1.86962891 * 2560; Err = 0.52382812 * 2560; time = 0.0121s; samplesPerSecond = 212448.1
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2541-2550, 58.09%]: CE.SM = 1.85917969 * 2560; Err = 0.51523438 * 2560; time = 0.0120s; samplesPerSecond = 212571.6
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2551-2560, 58.31%]: CE.SM = 1.84125977 * 2560; Err = 0.52031250 * 2560; time = 0.0122s; samplesPerSecond = 209973.8
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2561-2570, 58.54%]: CE.SM = 1.87871094 * 2560; Err = 0.52578125 * 2560; time = 0.0123s; samplesPerSecond = 208096.2
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2571-2580, 58.77%]: CE.SM = 1.88837891 * 2560; Err = 0.53789062 * 2560; time = 0.0123s; samplesPerSecond = 208571.0
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2581-2590, 59.00%]: CE.SM = 1.84331055 * 2560; Err = 0.51757812 * 2560; time = 0.0122s; samplesPerSecond = 209939.3
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2591-2600, 59.23%]: CE.SM = 1.82324219 * 2560; Err = 0.51406250 * 2560; time = 0.0128s; samplesPerSecond = 200752.8
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2601-2610, 59.45%]: CE.SM = 1.87426758 * 2560; Err = 0.52656250 * 2560; time = 0.0123s; samplesPerSecond = 208486.0
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2611-2620, 59.68%]: CE.SM = 1.84555664 * 2560; Err = 0.52070313 * 2560; time = 0.0125s; samplesPerSecond = 204767.2
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2621-2630, 59.91%]: CE.SM = 1.85776367 * 2560; Err = 0.52656250 * 2560; time = 0.0126s; samplesPerSecond = 203142.4
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2631-2640, 60.14%]: CE.SM = 1.89091797 * 2560; Err = 0.53164062 * 2560; time = 0.0127s; samplesPerSecond = 201479.6
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2641-2650, 60.36%]: CE.SM = 1.89101563 * 2560; Err = 0.53671875 * 2560; time = 0.0117s; samplesPerSecond = 218897.0
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2651-2660, 60.59%]: CE.SM = 1.88398438 * 2560; Err = 0.53945312 * 2560; time = 0.0123s; samplesPerSecond = 207556.3
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2661-2670, 60.82%]: CE.SM = 1.86562500 * 2560; Err = 0.51054687 * 2560; time = 0.0118s; samplesPerSecond = 216949.2
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2671-2680, 61.05%]: CE.SM = 1.87128906 * 2560; Err = 0.52539062 * 2560; time = 0.0121s; samplesPerSecond = 212430.5
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2681-2690, 61.28%]: CE.SM = 1.84438477 * 2560; Err = 0.52578125 * 2560; time = 0.0119s; samplesPerSecond = 215924.4
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2691-2700, 61.50%]: CE.SM = 1.83740234 * 2560; Err = 0.51679688 * 2560; time = 0.0116s; samplesPerSecond = 220994.5
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2701-2710, 61.73%]: CE.SM = 1.89438477 * 2560; Err = 0.53085938 * 2560; time = 0.0114s; samplesPerSecond = 223952.4
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2711-2720, 61.96%]: CE.SM = 1.87695312 * 2560; Err = 0.52968750 * 2560; time = 0.0115s; samplesPerSecond = 223541.7
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2721-2730, 62.19%]: CE.SM = 1.82788086 * 2560; Err = 0.51445312 * 2560; time = 0.0129s; samplesPerSecond = 197897.3
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2731-2740, 62.41%]: CE.SM = 1.78891602 * 2560; Err = 0.50273437 * 2560; time = 0.0123s; samplesPerSecond = 208096.2
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2741-2750, 62.64%]: CE.SM = 1.85717773 * 2560; Err = 0.52734375 * 2560; time = 0.0120s; samplesPerSecond = 212659.9
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2751-2760, 62.87%]: CE.SM = 1.87075195 * 2560; Err = 0.53515625 * 2560; time = 0.0121s; samplesPerSecond = 211099.2
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2761-2770, 63.10%]: CE.SM = 1.86494141 * 2560; Err = 0.53593750 * 2560; time = 0.0123s; samplesPerSecond = 207758.5
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2771-2780, 63.33%]: CE.SM = 1.90820312 * 2560; Err = 0.54335937 * 2560; time = 0.0123s; samplesPerSecond = 208079.3
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2781-2790, 63.55%]: CE.SM = 1.86381836 * 2560; Err = 0.51640625 * 2560; time = 0.0122s; samplesPerSecond = 210215.1
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2791-2800, 63.78%]: CE.SM = 1.89003906 * 2560; Err = 0.53242188 * 2560; time = 0.0125s; samplesPerSecond = 205194.0
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2801-2810, 64.01%]: CE.SM = 1.88906250 * 2560; Err = 0.52968750 * 2560; time = 0.0128s; samplesPerSecond = 200517.0
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2811-2820, 64.24%]: CE.SM = 1.85834961 * 2560; Err = 0.52421875 * 2560; time = 0.0116s; samplesPerSecond = 220044.7
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2821-2830, 64.46%]: CE.SM = 1.84150391 * 2560; Err = 0.51914063 * 2560; time = 0.0110s; samplesPerSecond = 232304.9
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2831-2840, 64.69%]: CE.SM = 1.89248047 * 2560; Err = 0.51562500 * 2560; time = 0.0127s; samplesPerSecond = 201368.7
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2841-2850, 64.92%]: CE.SM = 1.86342773 * 2560; Err = 0.52851563 * 2560; time = 0.0125s; samplesPerSecond = 204440.2
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2851-2860, 65.15%]: CE.SM = 1.84472656 * 2560; Err = 0.51875000 * 2560; time = 0.0117s; samplesPerSecond = 217928.0
12/20/2016 15:26:54:  Epoch[ 2 of 2]-Minibatch[2861-2870, 65.38%]: CE.SM = 1.84960938 * 2560; Err = 0.51679688 * 2560; time = 0.0119s; samplesPerSecond = 215815.2
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[2871-2880, 65.60%]: CE.SM = 1.83886719 * 2560; Err = 0.52109375 * 2560; time = 0.0122s; samplesPerSecond = 209836.1
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[2881-2890, 65.83%]: CE.SM = 1.82075195 * 2560; Err = 0.51679688 * 2560; time = 0.0122s; samplesPerSecond = 209030.8
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[2891-2900, 66.06%]: CE.SM = 1.83129883 * 2560; Err = 0.51992187 * 2560; time = 0.0127s; samplesPerSecond = 201670.1
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[2901-2910, 66.29%]: CE.SM = 1.83378906 * 2560; Err = 0.51601562 * 2560; time = 0.0124s; samplesPerSecond = 205986.5
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[2911-2920, 66.51%]: CE.SM = 1.78813477 * 2560; Err = 0.51562500 * 2560; time = 0.0129s; samplesPerSecond = 199190.8
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[2921-2930, 66.74%]: CE.SM = 1.81762695 * 2560; Err = 0.51250000 * 2560; time = 0.0128s; samplesPerSecond = 200737.1
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[2931-2940, 66.97%]: CE.SM = 1.80634766 * 2560; Err = 0.50703125 * 2560; time = 0.0123s; samplesPerSecond = 208639.0
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[2941-2950, 67.20%]: CE.SM = 1.80576172 * 2560; Err = 0.51171875 * 2560; time = 0.0119s; samplesPerSecond = 216015.5
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[2951-2960, 67.43%]: CE.SM = 1.86713867 * 2560; Err = 0.51914063 * 2560; time = 0.0115s; samplesPerSecond = 223152.0
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[2961-2970, 67.65%]: CE.SM = 1.84252930 * 2560; Err = 0.53125000 * 2560; time = 0.0118s; samplesPerSecond = 217502.1
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[2971-2980, 67.88%]: CE.SM = 1.81816406 * 2560; Err = 0.51093750 * 2560; time = 0.0117s; samplesPerSecond = 218560.6
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[2981-2990, 68.11%]: CE.SM = 1.81328125 * 2560; Err = 0.50703125 * 2560; time = 0.0125s; samplesPerSecond = 205539.9
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[2991-3000, 68.34%]: CE.SM = 1.85893555 * 2560; Err = 0.51640625 * 2560; time = 0.0123s; samplesPerSecond = 207287.4
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3001-3010, 68.56%]: CE.SM = 1.82934570 * 2560; Err = 0.50898438 * 2560; time = 0.0122s; samplesPerSecond = 210318.8
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3011-3020, 68.79%]: CE.SM = 1.84033203 * 2560; Err = 0.51367188 * 2560; time = 0.0120s; samplesPerSecond = 212730.6
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3021-3030, 69.02%]: CE.SM = 1.87211914 * 2560; Err = 0.51484375 * 2560; time = 0.0117s; samplesPerSecond = 219441.1
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3031-3040, 69.25%]: CE.SM = 1.83349609 * 2560; Err = 0.52031250 * 2560; time = 0.0122s; samplesPerSecond = 210111.6
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3041-3050, 69.48%]: CE.SM = 1.85756836 * 2560; Err = 0.52382812 * 2560; time = 0.0113s; samplesPerSecond = 226348.4
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3051-3060, 69.70%]: CE.SM = 1.84912109 * 2560; Err = 0.51875000 * 2560; time = 0.0135s; samplesPerSecond = 189573.5
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3061-3070, 69.93%]: CE.SM = 1.80712891 * 2560; Err = 0.50546875 * 2560; time = 0.0112s; samplesPerSecond = 229596.4
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3071-3080, 70.16%]: CE.SM = 1.84218750 * 2560; Err = 0.51601562 * 2560; time = 0.0112s; samplesPerSecond = 227677.0
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3081-3090, 70.39%]: CE.SM = 1.84047852 * 2560; Err = 0.51289063 * 2560; time = 0.0114s; samplesPerSecond = 224640.2
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3091-3100, 70.62%]: CE.SM = 1.80136719 * 2560; Err = 0.50859375 * 2560; time = 0.0114s; samplesPerSecond = 223658.9
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3101-3110, 70.84%]: CE.SM = 1.84233398 * 2560; Err = 0.51523438 * 2560; time = 0.0113s; samplesPerSecond = 226268.3
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3111-3120, 71.07%]: CE.SM = 1.89243164 * 2560; Err = 0.53750000 * 2560; time = 0.0115s; samplesPerSecond = 222010.2
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3121-3130, 71.30%]: CE.SM = 1.85029297 * 2560; Err = 0.52343750 * 2560; time = 0.0113s; samplesPerSecond = 227515.1
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3131-3140, 71.53%]: CE.SM = 1.83510742 * 2560; Err = 0.53046875 * 2560; time = 0.0117s; samplesPerSecond = 219159.3
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3141-3150, 71.75%]: CE.SM = 1.82812500 * 2560; Err = 0.50937500 * 2560; time = 0.0113s; samplesPerSecond = 227192.0
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3151-3160, 71.98%]: CE.SM = 1.85263672 * 2560; Err = 0.51757812 * 2560; time = 0.0113s; samplesPerSecond = 226568.7
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3161-3170, 72.21%]: CE.SM = 1.83691406 * 2560; Err = 0.51015625 * 2560; time = 0.0114s; samplesPerSecond = 224168.1
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3171-3180, 72.44%]: CE.SM = 1.91420898 * 2560; Err = 0.52031250 * 2560; time = 0.0116s; samplesPerSecond = 220347.7
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3181-3190, 72.67%]: CE.SM = 1.85649414 * 2560; Err = 0.51640625 * 2560; time = 0.0118s; samplesPerSecond = 217428.2
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3191-3200, 72.89%]: CE.SM = 1.87006836 * 2560; Err = 0.53710938 * 2560; time = 0.0113s; samplesPerSecond = 226568.7
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3201-3210, 73.12%]: CE.SM = 1.81015625 * 2560; Err = 0.51406250 * 2560; time = 0.0123s; samplesPerSecond = 208028.6
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3211-3220, 73.35%]: CE.SM = 1.81064453 * 2560; Err = 0.51796875 * 2560; time = 0.0125s; samplesPerSecond = 204832.8
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3221-3230, 73.58%]: CE.SM = 1.83979492 * 2560; Err = 0.52695313 * 2560; time = 0.0128s; samplesPerSecond = 199330.4
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3231-3240, 73.80%]: CE.SM = 1.83408203 * 2560; Err = 0.51875000 * 2560; time = 0.0121s; samplesPerSecond = 211047.0
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3241-3250, 74.03%]: CE.SM = 1.80664062 * 2560; Err = 0.49843750 * 2560; time = 0.0127s; samplesPerSecond = 200941.9
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3251-3260, 74.26%]: CE.SM = 1.79487305 * 2560; Err = 0.52773437 * 2560; time = 0.0123s; samplesPerSecond = 208588.0
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3261-3270, 74.49%]: CE.SM = 1.81635742 * 2560; Err = 0.51679688 * 2560; time = 0.0126s; samplesPerSecond = 202692.0
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3271-3280, 74.72%]: CE.SM = 1.79697266 * 2560; Err = 0.49960938 * 2560; time = 0.0121s; samplesPerSecond = 210734.3
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3281-3290, 74.94%]: CE.SM = 1.85244141 * 2560; Err = 0.51953125 * 2560; time = 0.0129s; samplesPerSecond = 197820.9
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3291-3300, 75.17%]: CE.SM = 1.80913086 * 2560; Err = 0.52070313 * 2560; time = 0.0122s; samplesPerSecond = 210630.2
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3301-3310, 75.40%]: CE.SM = 1.83208008 * 2560; Err = 0.51367188 * 2560; time = 0.0124s; samplesPerSecond = 206218.8
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3311-3320, 75.63%]: CE.SM = 1.86264648 * 2560; Err = 0.52070313 * 2560; time = 0.0121s; samplesPerSecond = 212113.7
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3321-3330, 75.85%]: CE.SM = 1.84252930 * 2560; Err = 0.52304688 * 2560; time = 0.0124s; samplesPerSecond = 206935.6
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3331-3340, 76.08%]: CE.SM = 1.82617188 * 2560; Err = 0.51601562 * 2560; time = 0.0147s; samplesPerSecond = 174220.8
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3341-3350, 76.31%]: CE.SM = 1.86938477 * 2560; Err = 0.51835937 * 2560; time = 0.0127s; samplesPerSecond = 201876.8
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3351-3360, 76.54%]: CE.SM = 1.83022461 * 2560; Err = 0.51953125 * 2560; time = 0.0127s; samplesPerSecond = 201194.6
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3361-3370, 76.77%]: CE.SM = 1.81645508 * 2560; Err = 0.51210937 * 2560; time = 0.0127s; samplesPerSecond = 201558.9
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3371-3380, 76.99%]: CE.SM = 1.85385742 * 2560; Err = 0.51015625 * 2560; time = 0.0121s; samplesPerSecond = 212008.3
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3381-3390, 77.22%]: CE.SM = 1.79584961 * 2560; Err = 0.51015625 * 2560; time = 0.0121s; samplesPerSecond = 210769.0
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3391-3400, 77.45%]: CE.SM = 1.81440430 * 2560; Err = 0.51718750 * 2560; time = 0.0132s; samplesPerSecond = 194662.0
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3401-3410, 77.68%]: CE.SM = 1.84702148 * 2560; Err = 0.50351563 * 2560; time = 0.0130s; samplesPerSecond = 196877.6
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3411-3420, 77.90%]: CE.SM = 1.84355469 * 2560; Err = 0.51328125 * 2560; time = 0.0123s; samplesPerSecond = 207522.7
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3421-3430, 78.13%]: CE.SM = 1.85087891 * 2560; Err = 0.53242188 * 2560; time = 0.0127s; samplesPerSecond = 202084.0
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3431-3440, 78.36%]: CE.SM = 1.83476562 * 2560; Err = 0.51054687 * 2560; time = 0.0123s; samplesPerSecond = 207505.9
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3441-3450, 78.59%]: CE.SM = 1.83666992 * 2560; Err = 0.51484375 * 2560; time = 0.0127s; samplesPerSecond = 201511.3
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3451-3460, 78.82%]: CE.SM = 1.76933594 * 2560; Err = 0.49726562 * 2560; time = 0.0127s; samplesPerSecond = 201432.1
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3461-3470, 79.04%]: CE.SM = 1.77695312 * 2560; Err = 0.49882813 * 2560; time = 0.0124s; samplesPerSecond = 206202.2
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3471-3480, 79.27%]: CE.SM = 1.84511719 * 2560; Err = 0.51523438 * 2560; time = 0.0124s; samplesPerSecond = 205754.7
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3481-3490, 79.50%]: CE.SM = 1.85498047 * 2560; Err = 0.51171875 * 2560; time = 0.0126s; samplesPerSecond = 203935.3
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3491-3500, 79.73%]: CE.SM = 1.81713867 * 2560; Err = 0.51093750 * 2560; time = 0.0133s; samplesPerSecond = 192684.0
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3501-3510, 79.95%]: CE.SM = 1.80195313 * 2560; Err = 0.51523438 * 2560; time = 0.0123s; samplesPerSecond = 208639.0
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3511-3520, 80.18%]: CE.SM = 1.82128906 * 2560; Err = 0.51289063 * 2560; time = 0.0196s; samplesPerSecond = 130618.9
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3521-3530, 80.41%]: CE.SM = 1.86850586 * 2560; Err = 0.52304688 * 2560; time = 0.0153s; samplesPerSecond = 166894.8
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3531-3540, 80.64%]: CE.SM = 1.83325195 * 2560; Err = 0.50859375 * 2560; time = 0.0148s; samplesPerSecond = 172727.9
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3541-3550, 80.87%]: CE.SM = 1.85903320 * 2560; Err = 0.53125000 * 2560; time = 0.0132s; samplesPerSecond = 193602.1
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3551-3560, 81.09%]: CE.SM = 1.82094727 * 2560; Err = 0.50000000 * 2560; time = 0.0136s; samplesPerSecond = 187807.2
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3561-3570, 81.32%]: CE.SM = 1.82465820 * 2560; Err = 0.51367188 * 2560; time = 0.0133s; samplesPerSecond = 193178.4
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3571-3580, 81.55%]: CE.SM = 1.81044922 * 2560; Err = 0.50468750 * 2560; time = 0.0132s; samplesPerSecond = 193924.7
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3581-3590, 81.78%]: CE.SM = 1.81601562 * 2560; Err = 0.51406250 * 2560; time = 0.0123s; samplesPerSecond = 208079.3
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3591-3600, 82.00%]: CE.SM = 1.82875977 * 2560; Err = 0.53085938 * 2560; time = 0.0127s; samplesPerSecond = 201622.4
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3601-3610, 82.23%]: CE.SM = 1.86459961 * 2560; Err = 0.52148438 * 2560; time = 0.0127s; samplesPerSecond = 202131.9
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3611-3620, 82.46%]: CE.SM = 1.75166016 * 2560; Err = 0.48945312 * 2560; time = 0.0125s; samplesPerSecond = 204931.2
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3621-3630, 82.69%]: CE.SM = 1.81909180 * 2560; Err = 0.49609375 * 2560; time = 0.0127s; samplesPerSecond = 201178.8
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3631-3640, 82.92%]: CE.SM = 1.88437500 * 2560; Err = 0.53320312 * 2560; time = 0.0125s; samplesPerSecond = 204130.5
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3641-3650, 83.14%]: CE.SM = 1.82729492 * 2560; Err = 0.50546875 * 2560; time = 0.0122s; samplesPerSecond = 209526.9
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3651-3660, 83.37%]: CE.SM = 1.82304687 * 2560; Err = 0.51289063 * 2560; time = 0.0109s; samplesPerSecond = 235901.2
12/20/2016 15:26:55:  Epoch[ 2 of 2]-Minibatch[3661-3670, 83.60%]: CE.SM = 1.75488281 * 2560; Err = 0.49687500 * 2560; time = 0.0119s; samplesPerSecond = 215706.1
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3671-3680, 83.83%]: CE.SM = 1.79887695 * 2560; Err = 0.51875000 * 2560; time = 0.0124s; samplesPerSecond = 207220.3
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3681-3690, 84.05%]: CE.SM = 1.72993164 * 2560; Err = 0.48750000 * 2560; time = 0.0131s; samplesPerSecond = 196033.4
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3691-3700, 84.28%]: CE.SM = 1.80927734 * 2560; Err = 0.51601562 * 2560; time = 0.0121s; samplesPerSecond = 212377.6
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3701-3710, 84.51%]: CE.SM = 1.76040039 * 2560; Err = 0.50898438 * 2560; time = 0.0126s; samplesPerSecond = 203336.0
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3711-3720, 84.74%]: CE.SM = 1.76772461 * 2560; Err = 0.50507813 * 2560; time = 0.0122s; samplesPerSecond = 210526.3
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3721-3730, 84.97%]: CE.SM = 1.84091797 * 2560; Err = 0.52421875 * 2560; time = 0.0124s; samplesPerSecond = 207270.7
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3731-3740, 85.19%]: CE.SM = 1.83491211 * 2560; Err = 0.52265625 * 2560; time = 0.0120s; samplesPerSecond = 213297.8
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3741-3750, 85.42%]: CE.SM = 1.78100586 * 2560; Err = 0.50078125 * 2560; time = 0.0131s; samplesPerSecond = 195360.2
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3751-3760, 85.65%]: CE.SM = 1.78291016 * 2560; Err = 0.49726562 * 2560; time = 0.0121s; samplesPerSecond = 212236.8
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3761-3770, 85.88%]: CE.SM = 1.79794922 * 2560; Err = 0.51953125 * 2560; time = 0.0126s; samplesPerSecond = 203061.8
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3771-3780, 86.10%]: CE.SM = 1.82446289 * 2560; Err = 0.52304688 * 2560; time = 0.0118s; samplesPerSecond = 217077.9
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3781-3790, 86.33%]: CE.SM = 1.87031250 * 2560; Err = 0.52578125 * 2560; time = 0.0126s; samplesPerSecond = 202403.5
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3791-3800, 86.56%]: CE.SM = 1.80161133 * 2560; Err = 0.50781250 * 2560; time = 0.0124s; samplesPerSecond = 207170.0
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3801-3810, 86.79%]: CE.SM = 1.82685547 * 2560; Err = 0.52343750 * 2560; time = 0.0120s; samplesPerSecond = 213564.7
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3811-3820, 87.02%]: CE.SM = 1.78115234 * 2560; Err = 0.50351563 * 2560; time = 0.0126s; samplesPerSecond = 202740.2
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3821-3830, 87.24%]: CE.SM = 1.81323242 * 2560; Err = 0.51484375 * 2560; time = 0.0124s; samplesPerSecond = 206651.6
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3831-3840, 87.47%]: CE.SM = 1.79628906 * 2560; Err = 0.51054687 * 2560; time = 0.0126s; samplesPerSecond = 203756.8
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3841-3850, 87.70%]: CE.SM = 1.83515625 * 2560; Err = 0.50742188 * 2560; time = 0.0123s; samplesPerSecond = 207825.9
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3851-3860, 87.93%]: CE.SM = 1.82045898 * 2560; Err = 0.52304688 * 2560; time = 0.0119s; samplesPerSecond = 214585.1
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3861-3870, 88.15%]: CE.SM = 1.83813477 * 2560; Err = 0.51757812 * 2560; time = 0.0133s; samplesPerSecond = 192916.4
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3871-3880, 88.38%]: CE.SM = 1.83710938 * 2560; Err = 0.50664062 * 2560; time = 0.0130s; samplesPerSecond = 196514.9
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3881-3890, 88.61%]: CE.SM = 1.76499023 * 2560; Err = 0.49609375 * 2560; time = 0.0125s; samplesPerSecond = 205062.5
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3891-3900, 88.84%]: CE.SM = 1.73325195 * 2560; Err = 0.47929688 * 2560; time = 0.0136s; samplesPerSecond = 188179.9
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3901-3910, 89.07%]: CE.SM = 1.77685547 * 2560; Err = 0.50390625 * 2560; time = 0.0125s; samplesPerSecond = 204619.9
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3911-3920, 89.29%]: CE.SM = 1.82666016 * 2560; Err = 0.51132813 * 2560; time = 0.0126s; samplesPerSecond = 203239.1
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3921-3930, 89.52%]: CE.SM = 1.80292969 * 2560; Err = 0.49570313 * 2560; time = 0.0130s; samplesPerSecond = 196923.1
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3931-3940, 89.75%]: CE.SM = 1.80249023 * 2560; Err = 0.51054687 * 2560; time = 0.0134s; samplesPerSecond = 191358.9
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3941-3950, 89.98%]: CE.SM = 1.83979492 * 2560; Err = 0.52539062 * 2560; time = 0.0125s; samplesPerSecond = 204277.1
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3951-3960, 90.21%]: CE.SM = 1.84409180 * 2560; Err = 0.51289063 * 2560; time = 0.0128s; samplesPerSecond = 200281.6
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3961-3970, 90.43%]: CE.SM = 1.78017578 * 2560; Err = 0.50390625 * 2560; time = 0.0134s; samplesPerSecond = 190731.6
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3971-3980, 90.66%]: CE.SM = 1.82504883 * 2560; Err = 0.51171875 * 2560; time = 0.0131s; samplesPerSecond = 195270.8
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3981-3990, 90.89%]: CE.SM = 1.84252930 * 2560; Err = 0.50937500 * 2560; time = 0.0132s; samplesPerSecond = 193675.3
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[3991-4000, 91.12%]: CE.SM = 1.72265625 * 2560; Err = 0.48437500 * 2560; time = 0.0127s; samplesPerSecond = 201068.2
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4001-4010, 91.34%]: CE.SM = 1.79365234 * 2560; Err = 0.50078125 * 2560; time = 0.0130s; samplesPerSecond = 196469.7
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4011-4020, 91.57%]: CE.SM = 1.78505859 * 2560; Err = 0.50351563 * 2560; time = 0.0129s; samplesPerSecond = 198203.8
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4021-4030, 91.80%]: CE.SM = 1.76494141 * 2560; Err = 0.50039062 * 2560; time = 0.0130s; samplesPerSecond = 196258.8
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4031-4040, 92.03%]: CE.SM = 1.84604492 * 2560; Err = 0.52265625 * 2560; time = 0.0130s; samplesPerSecond = 196590.4
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4041-4050, 92.26%]: CE.SM = 1.78842773 * 2560; Err = 0.49648437 * 2560; time = 0.0140s; samplesPerSecond = 182401.1
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4051-4060, 92.48%]: CE.SM = 1.81274414 * 2560; Err = 0.50351563 * 2560; time = 0.0133s; samplesPerSecond = 191961.6
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4061-4070, 92.71%]: CE.SM = 1.77856445 * 2560; Err = 0.50390625 * 2560; time = 0.0129s; samplesPerSecond = 198603.6
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4071-4080, 92.94%]: CE.SM = 1.82216797 * 2560; Err = 0.51757812 * 2560; time = 0.0131s; samplesPerSecond = 194943.6
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4081-4090, 93.17%]: CE.SM = 1.79311523 * 2560; Err = 0.49570313 * 2560; time = 0.0132s; samplesPerSecond = 193543.5
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4091-4100, 93.39%]: CE.SM = 1.81718750 * 2560; Err = 0.51367188 * 2560; time = 0.0128s; samplesPerSecond = 200689.9
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4101-4110, 93.62%]: CE.SM = 1.77495117 * 2560; Err = 0.50156250 * 2560; time = 0.0292s; samplesPerSecond = 87710.3
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4111-4120, 93.85%]: CE.SM = 1.78623047 * 2560; Err = 0.49804688 * 2560; time = 0.0163s; samplesPerSecond = 156824.3
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4121-4130, 94.08%]: CE.SM = 1.82124023 * 2560; Err = 0.52109375 * 2560; time = 0.0156s; samplesPerSecond = 164577.3
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4131-4140, 94.31%]: CE.SM = 1.78481445 * 2560; Err = 0.52343750 * 2560; time = 0.0147s; samplesPerSecond = 173842.2
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4141-4150, 94.53%]: CE.SM = 1.80024414 * 2560; Err = 0.50546875 * 2560; time = 0.0149s; samplesPerSecond = 171248.9
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4151-4160, 94.76%]: CE.SM = 1.80478516 * 2560; Err = 0.51992187 * 2560; time = 0.0145s; samplesPerSecond = 176771.2
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4161-4170, 94.99%]: CE.SM = 1.75395508 * 2560; Err = 0.49492188 * 2560; time = 0.0144s; samplesPerSecond = 177605.1
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4171-4180, 95.22%]: CE.SM = 1.77470703 * 2560; Err = 0.50937500 * 2560; time = 0.0144s; samplesPerSecond = 177469.7
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4181-4190, 95.44%]: CE.SM = 1.80605469 * 2560; Err = 0.50898438 * 2560; time = 0.0136s; samplesPerSecond = 187573.3
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4191-4200, 95.67%]: CE.SM = 1.84560547 * 2560; Err = 0.50703125 * 2560; time = 0.0141s; samplesPerSecond = 181689.1
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4201-4210, 95.90%]: CE.SM = 1.75351562 * 2560; Err = 0.50976562 * 2560; time = 0.0137s; samplesPerSecond = 186330.9
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4211-4220, 96.13%]: CE.SM = 1.85898437 * 2560; Err = 0.51367188 * 2560; time = 0.0136s; samplesPerSecond = 188526.4
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4221-4230, 96.36%]: CE.SM = 1.77314453 * 2560; Err = 0.50820312 * 2560; time = 0.0141s; samplesPerSecond = 181021.1
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4231-4240, 96.58%]: CE.SM = 1.82387695 * 2560; Err = 0.51367188 * 2560; time = 0.0139s; samplesPerSecond = 184384.9
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4241-4250, 96.81%]: CE.SM = 1.81279297 * 2560; Err = 0.52265625 * 2560; time = 0.0137s; samplesPerSecond = 186466.6
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4251-4260, 97.04%]: CE.SM = 1.78925781 * 2560; Err = 0.51250000 * 2560; time = 0.0140s; samplesPerSecond = 182987.8
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4261-4270, 97.27%]: CE.SM = 1.79052734 * 2560; Err = 0.49921875 * 2560; time = 0.0142s; samplesPerSecond = 180002.8
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4271-4280, 97.49%]: CE.SM = 1.78276367 * 2560; Err = 0.50625000 * 2560; time = 0.0129s; samplesPerSecond = 198111.7
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4281-4290, 97.72%]: CE.SM = 1.81381836 * 2560; Err = 0.50742188 * 2560; time = 0.0140s; samplesPerSecond = 183394.2
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4291-4300, 97.95%]: CE.SM = 1.81865234 * 2560; Err = 0.51523438 * 2560; time = 0.0136s; samplesPerSecond = 188415.4
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4301-4310, 98.18%]: CE.SM = 1.83369141 * 2560; Err = 0.51757812 * 2560; time = 0.0137s; samplesPerSecond = 187025.1
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4311-4320, 98.41%]: CE.SM = 1.81455078 * 2560; Err = 0.51406250 * 2560; time = 0.0134s; samplesPerSecond = 191516.4
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4321-4330, 98.63%]: CE.SM = 1.81796875 * 2560; Err = 0.51250000 * 2560; time = 0.0136s; samplesPerSecond = 188276.8
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4331-4340, 98.86%]: CE.SM = 1.76904297 * 2560; Err = 0.49687500 * 2560; time = 0.0132s; samplesPerSecond = 193616.7
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4341-4350, 99.09%]: CE.SM = 1.83959961 * 2560; Err = 0.51796875 * 2560; time = 0.0132s; samplesPerSecond = 193411.9
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4351-4360, 99.32%]: CE.SM = 1.85507812 * 2560; Err = 0.51562500 * 2560; time = 0.0137s; samplesPerSecond = 187244.0
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4361-4370, 99.54%]: CE.SM = 1.79169922 * 2560; Err = 0.50078125 * 2560; time = 0.0137s; samplesPerSecond = 186317.3
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4371-4380, 99.77%]: CE.SM = 1.73535156 * 2560; Err = 0.48671875 * 2560; time = 0.0138s; samplesPerSecond = 185078.1
12/20/2016 15:26:56:  Epoch[ 2 of 2]-Minibatch[4381-4390, 100.00%]: CE.SM = 1.76074219 * 2560; Err = 0.49804688 * 2560; time = 0.0138s; samplesPerSecond = 185992.4
12/20/2016 15:26:56: Finished Epoch[ 2 of 2]: [Training] CE.SM = 1.87851578 * 1124823; Err = 0.52538933 * 1124823; totalSamplesSeen = 2249646; learningRatePerSample = 0.00039062501; epochTime=5.74425s
12/20/2016 15:26:56: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/dptmodel1/cntkSpeech.dnn'

12/20/2016 15:26:57: Action "train" complete.


12/20/2016 15:26:57: ##############################################################################
12/20/2016 15:26:57: #                                                                            #
12/20/2016 15:26:57: # TIMIT_AddLayer2 command (edit action)                                      #
12/20/2016 15:26:57: #                                                                            #
12/20/2016 15:26:57: ##############################################################################


12/20/2016 15:26:57: Action "edit" complete.


12/20/2016 15:26:57: ##############################################################################
12/20/2016 15:26:57: #                                                                            #
12/20/2016 15:26:57: # TIMIT_DiscrimPreTrain2 command (train action)                              #
12/20/2016 15:26:57: #                                                                            #
12/20/2016 15:26:57: ##############################################################################

12/20/2016 15:26:57: 
Starting from checkpoint. Loading network from '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/dptmodel2/cntkSpeech.dnn.0'.
NDLBuilder Using GPU 0
Reading script file /home/philly/data/CNTKTestData/Speech/ASR/TIMIT.train.scp.fbank.fullpath.rnn ... 3696 entries
HTKDataDeserializer::HTKDataDeserializer: selected 3696 utterances grouped into 13 chunks, average chunk size: 284.3 utterances, 86524.8 frames (for I/O: 284.3 utterances, 86524.8 frames)
HTKDataDeserializer::HTKDataDeserializer: determined feature kind as 72-dimensional 'FBANK_D_A_Z' with frame shift 10.0 ms
total 183 state names in state list /home/philly/data/CNTKTestData/Speech/ASR/TIMIT.statelist
htkmlfreader: reading MLF file /home/philly/data/CNTKTestData/Speech/ASR/TIMIT.train.align_cistate.mlf.cntk ... total 3696 entries
MLFDataDeserializer::MLFDataDeserializer: 3696 utterances with 1124823 frames in 183 classes
12/20/2016 15:26:57: 
Model has 24 nodes. Using GPU 0.

12/20/2016 15:26:57: Training criterion:   CE.SM = CrossEntropyWithSoftmax
12/20/2016 15:26:57: Evaluation criterion: Err = ClassificationError

12/20/2016 15:26:57: Training 762551 parameters in 6 out of 6 parameter tensors and 15 nodes with gradient:

12/20/2016 15:26:57: 	Node 'CE.BFF.B' (LearnableParameter operation) : [183]
12/20/2016 15:26:57: 	Node 'CE.BFF.W' (LearnableParameter operation) : [183 x 512]
12/20/2016 15:26:57: 	Node 'L1.BFF.B' (LearnableParameter operation) : [512]
12/20/2016 15:26:57: 	Node 'L1.BFF.W' (LearnableParameter operation) : [512 x 792]
12/20/2016 15:26:57: 	Node 'L2.BFF.B' (LearnableParameter operation) : [512]
12/20/2016 15:26:57: 	Node 'L2.BFF.W' (LearnableParameter operation) : [512 x 512]

12/20/2016 15:26:57: No PreCompute nodes found, or all already computed. Skipping pre-computation step.

12/20/2016 15:26:57: Starting Epoch 1: learning rate per sample = 0.000391  effective momentum = 0.900000  momentum as time constant = 2429.8 samples

12/20/2016 15:26:57: Starting minibatch loop.
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[   1-  10]: CE.SM = 5.94420013 * 2560; Err = 0.95195312 * 2560; time = 0.6606s; samplesPerSecond = 3875.4
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[  11-  20]: CE.SM = 4.67969666 * 2560; Err = 0.92460937 * 2560; time = 0.0138s; samplesPerSecond = 186168.3
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[  21-  30]: CE.SM = 4.16574402 * 2560; Err = 0.87304688 * 2560; time = 0.0137s; samplesPerSecond = 186970.5
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[  31-  40]: CE.SM = 3.85346680 * 2560; Err = 0.84843750 * 2560; time = 0.0135s; samplesPerSecond = 189826.5
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[  41-  50]: CE.SM = 3.62228851 * 2560; Err = 0.81328125 * 2560; time = 0.0137s; samplesPerSecond = 187134.5
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[  51-  60]: CE.SM = 3.47206726 * 2560; Err = 0.77460938 * 2560; time = 0.0138s; samplesPerSecond = 186168.3
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[  61-  70]: CE.SM = 3.34695740 * 2560; Err = 0.76171875 * 2560; time = 0.0139s; samplesPerSecond = 183960.9
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[  71-  80]: CE.SM = 3.25104675 * 2560; Err = 0.76054687 * 2560; time = 0.0138s; samplesPerSecond = 186114.1
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[  81-  90]: CE.SM = 3.11874390 * 2560; Err = 0.72421875 * 2560; time = 0.0139s; samplesPerSecond = 183657.4
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[  91- 100]: CE.SM = 3.10517273 * 2560; Err = 0.72656250 * 2560; time = 0.0141s; samplesPerSecond = 181856.9
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 101- 110]: CE.SM = 3.02628784 * 2560; Err = 0.71796875 * 2560; time = 0.0140s; samplesPerSecond = 182375.2
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 111- 120]: CE.SM = 3.00162964 * 2560; Err = 0.72148437 * 2560; time = 0.0138s; samplesPerSecond = 186073.6
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 121- 130]: CE.SM = 2.87215881 * 2560; Err = 0.69023437 * 2560; time = 0.0167s; samplesPerSecond = 153689.1
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 131- 140]: CE.SM = 2.84855347 * 2560; Err = 0.69140625 * 2560; time = 0.0221s; samplesPerSecond = 115596.5
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 141- 150]: CE.SM = 2.80492554 * 2560; Err = 0.67929688 * 2560; time = 0.0144s; samplesPerSecond = 177617.4
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 151- 160]: CE.SM = 2.82548828 * 2560; Err = 0.68476563 * 2560; time = 0.0157s; samplesPerSecond = 163473.8
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 161- 170]: CE.SM = 2.73877563 * 2560; Err = 0.65742188 * 2560; time = 0.0144s; samplesPerSecond = 178012.7
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 171- 180]: CE.SM = 2.72180786 * 2560; Err = 0.66875000 * 2560; time = 0.0148s; samplesPerSecond = 173253.9
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 181- 190]: CE.SM = 2.66038208 * 2560; Err = 0.66289062 * 2560; time = 0.0136s; samplesPerSecond = 187972.7
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 191- 200]: CE.SM = 2.66605835 * 2560; Err = 0.66875000 * 2560; time = 0.0143s; samplesPerSecond = 179083.6
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 201- 210]: CE.SM = 2.67689819 * 2560; Err = 0.67812500 * 2560; time = 0.0139s; samplesPerSecond = 183749.6
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 211- 220]: CE.SM = 2.60620117 * 2560; Err = 0.65117187 * 2560; time = 0.0137s; samplesPerSecond = 187161.9
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 221- 230]: CE.SM = 2.62480469 * 2560; Err = 0.64531250 * 2560; time = 0.0139s; samplesPerSecond = 184212.4
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 231- 240]: CE.SM = 2.61395264 * 2560; Err = 0.65312500 * 2560; time = 0.0139s; samplesPerSecond = 184159.4
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 241- 250]: CE.SM = 2.52625732 * 2560; Err = 0.63750000 * 2560; time = 0.0140s; samplesPerSecond = 183184.3
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 251- 260]: CE.SM = 2.58402710 * 2560; Err = 0.64570313 * 2560; time = 0.0138s; samplesPerSecond = 185051.3
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 261- 270]: CE.SM = 2.54647217 * 2560; Err = 0.64882812 * 2560; time = 0.0139s; samplesPerSecond = 184584.3
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 271- 280]: CE.SM = 2.56469727 * 2560; Err = 0.65898437 * 2560; time = 0.0138s; samplesPerSecond = 185158.4
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 281- 290]: CE.SM = 2.53161011 * 2560; Err = 0.63945312 * 2560; time = 0.0138s; samplesPerSecond = 184904.3
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 291- 300]: CE.SM = 2.50845337 * 2560; Err = 0.63945312 * 2560; time = 0.0137s; samplesPerSecond = 186806.8
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 301- 310]: CE.SM = 2.48307495 * 2560; Err = 0.63945312 * 2560; time = 0.0138s; samplesPerSecond = 185185.2
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 311- 320]: CE.SM = 2.43105469 * 2560; Err = 0.62578125 * 2560; time = 0.0137s; samplesPerSecond = 187093.5
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 321- 330]: CE.SM = 2.49989014 * 2560; Err = 0.62968750 * 2560; time = 0.0138s; samplesPerSecond = 185776.5
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 331- 340]: CE.SM = 2.47990112 * 2560; Err = 0.62304688 * 2560; time = 0.0137s; samplesPerSecond = 186493.8
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 341- 350]: CE.SM = 2.41032715 * 2560; Err = 0.63281250 * 2560; time = 0.0136s; samplesPerSecond = 187697.0
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 351- 360]: CE.SM = 2.43629150 * 2560; Err = 0.63554687 * 2560; time = 0.0137s; samplesPerSecond = 186793.1
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 361- 370]: CE.SM = 2.37458496 * 2560; Err = 0.60781250 * 2560; time = 0.0137s; samplesPerSecond = 187120.8
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 371- 380]: CE.SM = 2.39515381 * 2560; Err = 0.61132812 * 2560; time = 0.0138s; samplesPerSecond = 185051.3
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 381- 390]: CE.SM = 2.38558350 * 2560; Err = 0.62382812 * 2560; time = 0.0139s; samplesPerSecond = 184331.8
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 391- 400]: CE.SM = 2.42677002 * 2560; Err = 0.61796875 * 2560; time = 0.0138s; samplesPerSecond = 185749.5
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 401- 410]: CE.SM = 2.41041260 * 2560; Err = 0.62382812 * 2560; time = 0.0139s; samplesPerSecond = 184398.2
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 411- 420]: CE.SM = 2.36721191 * 2560; Err = 0.61015625 * 2560; time = 0.0138s; samplesPerSecond = 186141.2
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 421- 430]: CE.SM = 2.34991455 * 2560; Err = 0.61132812 * 2560; time = 0.0138s; samplesPerSecond = 185924.9
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 431- 440]: CE.SM = 2.32735596 * 2560; Err = 0.60976562 * 2560; time = 0.0138s; samplesPerSecond = 185830.4
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 441- 450]: CE.SM = 2.36285400 * 2560; Err = 0.62382812 * 2560; time = 0.0135s; samplesPerSecond = 189713.9
12/20/2016 15:26:58:  Epoch[ 1 of 2]-Minibatch[ 451- 460]: CE.SM = 2.30509033 * 2560; Err = 0.60351562 * 2560; time = 0.0136s; samplesPerSecond = 188762.7
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 461- 470]: CE.SM = 2.33947754 * 2560; Err = 0.60781250 * 2560; time = 0.0136s; samplesPerSecond = 187614.5
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 471- 480]: CE.SM = 2.33045654 * 2560; Err = 0.62460938 * 2560; time = 0.0138s; samplesPerSecond = 185628.3
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 481- 490]: CE.SM = 2.27999268 * 2560; Err = 0.59257812 * 2560; time = 0.0137s; samplesPerSecond = 187504.6
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 491- 500]: CE.SM = 2.28487549 * 2560; Err = 0.59179688 * 2560; time = 0.0138s; samplesPerSecond = 185857.4
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 501- 510]: CE.SM = 2.26192627 * 2560; Err = 0.58945313 * 2560; time = 0.0493s; samplesPerSecond = 51934.4
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 511- 520]: CE.SM = 2.26481934 * 2560; Err = 0.60585937 * 2560; time = 0.0146s; samplesPerSecond = 175920.8
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 521- 530]: CE.SM = 2.28381348 * 2560; Err = 0.59453125 * 2560; time = 0.0141s; samplesPerSecond = 181740.7
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 531- 540]: CE.SM = 2.26817627 * 2560; Err = 0.60117188 * 2560; time = 0.0146s; samplesPerSecond = 174744.0
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 541- 550]: CE.SM = 2.29996338 * 2560; Err = 0.61054688 * 2560; time = 0.0148s; samplesPerSecond = 173277.4
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 551- 560]: CE.SM = 2.30961914 * 2560; Err = 0.60703125 * 2560; time = 0.0151s; samplesPerSecond = 170043.2
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 561- 570]: CE.SM = 2.28759766 * 2560; Err = 0.61289063 * 2560; time = 0.0148s; samplesPerSecond = 173535.8
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 571- 580]: CE.SM = 2.26955566 * 2560; Err = 0.60312500 * 2560; time = 0.0147s; samplesPerSecond = 174493.9
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 581- 590]: CE.SM = 2.23947754 * 2560; Err = 0.58710938 * 2560; time = 0.0144s; samplesPerSecond = 178087.0
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 591- 600]: CE.SM = 2.20615234 * 2560; Err = 0.58515625 * 2560; time = 0.0143s; samplesPerSecond = 179046.0
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 601- 610]: CE.SM = 2.27614746 * 2560; Err = 0.61210937 * 2560; time = 0.0139s; samplesPerSecond = 184704.2
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 611- 620]: CE.SM = 2.23916016 * 2560; Err = 0.60859375 * 2560; time = 0.0140s; samplesPerSecond = 182349.2
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 621- 630]: CE.SM = 2.23623047 * 2560; Err = 0.59023437 * 2560; time = 0.0137s; samplesPerSecond = 186943.2
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 631- 640]: CE.SM = 2.17432861 * 2560; Err = 0.59023437 * 2560; time = 0.0139s; samplesPerSecond = 183934.5
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 641- 650]: CE.SM = 2.25935059 * 2560; Err = 0.59687500 * 2560; time = 0.0139s; samplesPerSecond = 184557.7
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 651- 660]: CE.SM = 2.20894775 * 2560; Err = 0.59257812 * 2560; time = 0.0140s; samplesPerSecond = 183144.9
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 661- 670]: CE.SM = 2.21094971 * 2560; Err = 0.60078125 * 2560; time = 0.0139s; samplesPerSecond = 183881.6
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 671- 680]: CE.SM = 2.22462158 * 2560; Err = 0.60000000 * 2560; time = 0.0140s; samplesPerSecond = 183236.7
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 681- 690]: CE.SM = 2.21235352 * 2560; Err = 0.59882813 * 2560; time = 0.0139s; samplesPerSecond = 184066.7
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 691- 700]: CE.SM = 2.18731689 * 2560; Err = 0.59023437 * 2560; time = 0.0137s; samplesPerSecond = 187463.4
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 701- 710]: CE.SM = 2.25867920 * 2560; Err = 0.60703125 * 2560; time = 0.0140s; samplesPerSecond = 182674.5
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 711- 720]: CE.SM = 2.21640625 * 2560; Err = 0.58906250 * 2560; time = 0.0139s; samplesPerSecond = 184027.0
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 721- 730]: CE.SM = 2.18303223 * 2560; Err = 0.58632812 * 2560; time = 0.0139s; samplesPerSecond = 184411.5
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 731- 740]: CE.SM = 2.23302002 * 2560; Err = 0.60664063 * 2560; time = 0.0137s; samplesPerSecond = 187298.8
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 741- 750]: CE.SM = 2.17429199 * 2560; Err = 0.59882813 * 2560; time = 0.0139s; samplesPerSecond = 184610.9
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 751- 760]: CE.SM = 2.16752930 * 2560; Err = 0.58085937 * 2560; time = 0.0139s; samplesPerSecond = 184624.3
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 761- 770]: CE.SM = 2.13688965 * 2560; Err = 0.56718750 * 2560; time = 0.0139s; samplesPerSecond = 184677.5
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 771- 780]: CE.SM = 2.16594238 * 2560; Err = 0.59765625 * 2560; time = 0.0138s; samplesPerSecond = 185776.5
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 781- 790]: CE.SM = 2.18470459 * 2560; Err = 0.58867187 * 2560; time = 0.0138s; samplesPerSecond = 185897.9
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 791- 800]: CE.SM = 2.17609863 * 2560; Err = 0.59140625 * 2560; time = 0.0138s; samplesPerSecond = 185198.6
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 801- 810]: CE.SM = 2.18056641 * 2560; Err = 0.58710938 * 2560; time = 0.0139s; samplesPerSecond = 184797.5
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 811- 820]: CE.SM = 2.17788086 * 2560; Err = 0.59609375 * 2560; time = 0.0138s; samplesPerSecond = 185722.6
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 821- 830]: CE.SM = 2.18115234 * 2560; Err = 0.59023437 * 2560; time = 0.0137s; samplesPerSecond = 187408.5
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 831- 840]: CE.SM = 2.14001465 * 2560; Err = 0.58085937 * 2560; time = 0.0138s; samplesPerSecond = 185413.2
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 841- 850]: CE.SM = 2.15212402 * 2560; Err = 0.58046875 * 2560; time = 0.0159s; samplesPerSecond = 161188.8
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 851- 860]: CE.SM = 2.15566406 * 2560; Err = 0.57734375 * 2560; time = 0.0139s; samplesPerSecond = 184265.5
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 861- 870]: CE.SM = 2.18437500 * 2560; Err = 0.59570312 * 2560; time = 0.0137s; samplesPerSecond = 187257.7
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 871- 880]: CE.SM = 2.17077637 * 2560; Err = 0.59140625 * 2560; time = 0.0156s; samplesPerSecond = 163829.5
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 881- 890]: CE.SM = 2.15415039 * 2560; Err = 0.58437500 * 2560; time = 0.0136s; samplesPerSecond = 188512.5
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 891- 900]: CE.SM = 2.13859863 * 2560; Err = 0.57500000 * 2560; time = 0.0138s; samplesPerSecond = 184864.2
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 901- 910]: CE.SM = 2.22949219 * 2560; Err = 0.59921875 * 2560; time = 0.0139s; samplesPerSecond = 183670.5
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 911- 920]: CE.SM = 2.13901367 * 2560; Err = 0.57773438 * 2560; time = 0.0137s; samplesPerSecond = 186236.0
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 921- 930]: CE.SM = 2.18354492 * 2560; Err = 0.59296875 * 2560; time = 0.0135s; samplesPerSecond = 189391.1
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 931- 940]: CE.SM = 2.12163086 * 2560; Err = 0.58007812 * 2560; time = 0.0134s; samplesPerSecond = 190973.5
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 941- 950]: CE.SM = 2.12070312 * 2560; Err = 0.56640625 * 2560; time = 0.0135s; samplesPerSecond = 189503.3
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 951- 960]: CE.SM = 2.15219727 * 2560; Err = 0.57773438 * 2560; time = 0.0133s; samplesPerSecond = 191875.3
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 961- 970]: CE.SM = 2.13505859 * 2560; Err = 0.57578125 * 2560; time = 0.0135s; samplesPerSecond = 189433.2
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 971- 980]: CE.SM = 2.12956543 * 2560; Err = 0.58554688 * 2560; time = 0.0135s; samplesPerSecond = 189671.8
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 981- 990]: CE.SM = 2.10236816 * 2560; Err = 0.54648438 * 2560; time = 0.0134s; samplesPerSecond = 190603.8
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[ 991-1000]: CE.SM = 2.12192383 * 2560; Err = 0.58437500 * 2560; time = 0.0135s; samplesPerSecond = 190066.1
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[1001-1010]: CE.SM = 2.11430664 * 2560; Err = 0.57187500 * 2560; time = 0.0133s; samplesPerSecond = 192945.4
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[1011-1020]: CE.SM = 2.12680664 * 2560; Err = 0.57343750 * 2560; time = 0.0132s; samplesPerSecond = 193807.3
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[1021-1030]: CE.SM = 2.09177246 * 2560; Err = 0.57226562 * 2560; time = 0.0134s; samplesPerSecond = 191087.6
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[1031-1040]: CE.SM = 2.06386719 * 2560; Err = 0.57773438 * 2560; time = 0.0133s; samplesPerSecond = 192843.7
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[1041-1050]: CE.SM = 2.11679687 * 2560; Err = 0.57070312 * 2560; time = 0.0134s; samplesPerSecond = 190362.9
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[1051-1060]: CE.SM = 2.11418457 * 2560; Err = 0.57070312 * 2560; time = 0.0135s; samplesPerSecond = 189573.5
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[1061-1070]: CE.SM = 2.10070801 * 2560; Err = 0.58398438 * 2560; time = 0.0136s; samplesPerSecond = 188124.6
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[1071-1080]: CE.SM = 2.11284180 * 2560; Err = 0.57617188 * 2560; time = 0.0137s; samplesPerSecond = 187381.1
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[1081-1090]: CE.SM = 2.09321289 * 2560; Err = 0.56835938 * 2560; time = 0.0136s; samplesPerSecond = 188526.4
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[1091-1100]: CE.SM = 2.14072266 * 2560; Err = 0.59765625 * 2560; time = 0.0135s; samplesPerSecond = 189069.4
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[1101-1110]: CE.SM = 2.08286133 * 2560; Err = 0.56835938 * 2560; time = 0.0133s; samplesPerSecond = 192495.7
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[1111-1120]: CE.SM = 2.11921387 * 2560; Err = 0.56718750 * 2560; time = 0.0136s; samplesPerSecond = 188152.3
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[1121-1130]: CE.SM = 2.07326660 * 2560; Err = 0.55976563 * 2560; time = 0.0136s; samplesPerSecond = 187614.5
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[1131-1140]: CE.SM = 2.10241699 * 2560; Err = 0.56445312 * 2560; time = 0.0135s; samplesPerSecond = 189475.2
12/20/2016 15:26:59:  Epoch[ 1 of 2]-Minibatch[1141-1150]: CE.SM = 2.03896484 * 2560; Err = 0.56367188 * 2560; time = 0.0133s; samplesPerSecond = 192163.3
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1151-1160]: CE.SM = 2.03527832 * 2560; Err = 0.57109375 * 2560; time = 0.0137s; samplesPerSecond = 187312.5
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1161-1170]: CE.SM = 2.06286621 * 2560; Err = 0.56367188 * 2560; time = 0.0136s; samplesPerSecond = 188263.0
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1171-1180]: CE.SM = 2.08144531 * 2560; Err = 0.56835938 * 2560; time = 0.0135s; samplesPerSecond = 189461.2
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1181-1190]: CE.SM = 2.10759277 * 2560; Err = 0.57617188 * 2560; time = 0.0134s; samplesPerSecond = 190377.0
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1191-1200]: CE.SM = 2.09802246 * 2560; Err = 0.57343750 * 2560; time = 0.0137s; samplesPerSecond = 187435.9
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1201-1210]: CE.SM = 2.08498535 * 2560; Err = 0.57617188 * 2560; time = 0.0134s; samplesPerSecond = 190377.0
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1211-1220]: CE.SM = 2.09631348 * 2560; Err = 0.56796875 * 2560; time = 0.0136s; samplesPerSecond = 188693.2
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1221-1230]: CE.SM = 2.10070801 * 2560; Err = 0.58281250 * 2560; time = 0.0144s; samplesPerSecond = 177273.0
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1231-1240]: CE.SM = 2.06872559 * 2560; Err = 0.56953125 * 2560; time = 0.0135s; samplesPerSecond = 190094.3
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1241-1250]: CE.SM = 2.08120117 * 2560; Err = 0.56640625 * 2560; time = 0.0134s; samplesPerSecond = 190575.4
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1251-1260]: CE.SM = 2.05219727 * 2560; Err = 0.56718750 * 2560; time = 0.0133s; samplesPerSecond = 191789.0
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1261-1270]: CE.SM = 2.06970215 * 2560; Err = 0.55898437 * 2560; time = 0.0136s; samplesPerSecond = 187752.1
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1271-1280]: CE.SM = 2.09816895 * 2560; Err = 0.57265625 * 2560; time = 0.0135s; samplesPerSecond = 189545.4
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1281-1290]: CE.SM = 2.08830566 * 2560; Err = 0.57500000 * 2560; time = 0.0136s; samplesPerSecond = 188679.2
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1291-1300]: CE.SM = 2.09643555 * 2560; Err = 0.57695312 * 2560; time = 0.0134s; samplesPerSecond = 190532.9
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1301-1310]: CE.SM = 2.05429687 * 2560; Err = 0.57031250 * 2560; time = 0.0135s; samplesPerSecond = 189475.2
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1311-1320]: CE.SM = 2.06188965 * 2560; Err = 0.55664062 * 2560; time = 0.0137s; samplesPerSecond = 187285.1
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1321-1330]: CE.SM = 2.07668457 * 2560; Err = 0.56796875 * 2560; time = 0.0132s; samplesPerSecond = 194277.9
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1331-1340]: CE.SM = 2.08168945 * 2560; Err = 0.57968750 * 2560; time = 0.0134s; samplesPerSecond = 190405.4
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1341-1350]: CE.SM = 2.02194824 * 2560; Err = 0.54687500 * 2560; time = 0.0136s; samplesPerSecond = 187573.3
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1351-1360]: CE.SM = 2.02275391 * 2560; Err = 0.55625000 * 2560; time = 0.0135s; samplesPerSecond = 189489.3
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1361-1370]: CE.SM = 2.06604004 * 2560; Err = 0.57226562 * 2560; time = 0.0136s; samplesPerSecond = 188776.6
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1371-1380]: CE.SM = 2.05793457 * 2560; Err = 0.57304687 * 2560; time = 0.0134s; samplesPerSecond = 191559.4
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1381-1390]: CE.SM = 2.09367676 * 2560; Err = 0.57109375 * 2560; time = 0.0135s; samplesPerSecond = 189125.3
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1391-1400]: CE.SM = 1.99501953 * 2560; Err = 0.56093750 * 2560; time = 0.0134s; samplesPerSecond = 191201.7
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1401-1410]: CE.SM = 2.07678223 * 2560; Err = 0.56796875 * 2560; time = 0.0135s; samplesPerSecond = 189995.5
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1411-1420]: CE.SM = 2.07834473 * 2560; Err = 0.58476562 * 2560; time = 0.0136s; samplesPerSecond = 188290.7
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1421-1430]: CE.SM = 2.04689941 * 2560; Err = 0.57460937 * 2560; time = 0.0138s; samplesPerSecond = 185145.0
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1431-1440]: CE.SM = 2.06831055 * 2560; Err = 0.56835938 * 2560; time = 0.0139s; samplesPerSecond = 184584.3
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1441-1450]: CE.SM = 2.01337891 * 2560; Err = 0.56484375 * 2560; time = 0.0139s; samplesPerSecond = 184292.0
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1451-1460]: CE.SM = 2.07126465 * 2560; Err = 0.58046875 * 2560; time = 0.0138s; samplesPerSecond = 185776.5
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1461-1470]: CE.SM = 2.06589355 * 2560; Err = 0.56796875 * 2560; time = 0.0136s; samplesPerSecond = 187545.8
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1471-1480]: CE.SM = 2.01569824 * 2560; Err = 0.55312500 * 2560; time = 0.0139s; samplesPerSecond = 184531.1
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1481-1490]: CE.SM = 2.05886230 * 2560; Err = 0.55859375 * 2560; time = 0.0139s; samplesPerSecond = 184637.6
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1491-1500]: CE.SM = 2.04626465 * 2560; Err = 0.57265625 * 2560; time = 0.0167s; samplesPerSecond = 153027.7
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1501-1510]: CE.SM = 2.00869141 * 2560; Err = 0.55585938 * 2560; time = 0.0135s; samplesPerSecond = 190037.9
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1511-1520]: CE.SM = 2.01970215 * 2560; Err = 0.55195313 * 2560; time = 0.0141s; samplesPerSecond = 181418.8
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1521-1530]: CE.SM = 2.06591797 * 2560; Err = 0.55898437 * 2560; time = 0.0141s; samplesPerSecond = 181341.6
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1531-1540]: CE.SM = 2.07675781 * 2560; Err = 0.57070312 * 2560; time = 0.0139s; samplesPerSecond = 184159.4
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1541-1550]: CE.SM = 2.12395020 * 2560; Err = 0.59179688 * 2560; time = 0.0139s; samplesPerSecond = 183828.8
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1551-1560]: CE.SM = 2.02360840 * 2560; Err = 0.57226562 * 2560; time = 0.0139s; samplesPerSecond = 184531.1
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1561-1570]: CE.SM = 2.01040039 * 2560; Err = 0.56914062 * 2560; time = 0.0137s; samplesPerSecond = 186847.7
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1571-1580]: CE.SM = 2.00563965 * 2560; Err = 0.55351562 * 2560; time = 0.0148s; samplesPerSecond = 172716.2
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1581-1590]: CE.SM = 2.05007324 * 2560; Err = 0.55156250 * 2560; time = 0.0141s; samplesPerSecond = 181740.7
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1591-1600]: CE.SM = 2.03186035 * 2560; Err = 0.54960937 * 2560; time = 0.0140s; samplesPerSecond = 182284.2
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1601-1610]: CE.SM = 2.03540039 * 2560; Err = 0.57968750 * 2560; time = 0.0137s; samplesPerSecond = 187312.5
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1611-1620]: CE.SM = 2.04838867 * 2560; Err = 0.57031250 * 2560; time = 0.0136s; samplesPerSecond = 187752.1
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1621-1630]: CE.SM = 2.02785645 * 2560; Err = 0.56718750 * 2560; time = 0.0138s; samplesPerSecond = 185763.0
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1631-1640]: CE.SM = 2.01467285 * 2560; Err = 0.56093750 * 2560; time = 0.0138s; samplesPerSecond = 185938.4
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1641-1650]: CE.SM = 2.00605469 * 2560; Err = 0.56015625 * 2560; time = 0.0140s; samplesPerSecond = 183486.2
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1651-1660]: CE.SM = 2.02138672 * 2560; Err = 0.56171875 * 2560; time = 0.0138s; samplesPerSecond = 185198.6
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1661-1670]: CE.SM = 2.02915039 * 2560; Err = 0.56640625 * 2560; time = 0.0140s; samplesPerSecond = 182479.2
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1671-1680]: CE.SM = 2.04030762 * 2560; Err = 0.56875000 * 2560; time = 0.0139s; samplesPerSecond = 184278.7
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1681-1690]: CE.SM = 2.03925781 * 2560; Err = 0.57265625 * 2560; time = 0.0136s; samplesPerSecond = 188041.7
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1691-1700]: CE.SM = 2.00346680 * 2560; Err = 0.56406250 * 2560; time = 0.0139s; samplesPerSecond = 183552.0
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1701-1710]: CE.SM = 1.94975586 * 2560; Err = 0.52773437 * 2560; time = 0.0140s; samplesPerSecond = 182505.2
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1711-1720]: CE.SM = 2.04074707 * 2560; Err = 0.56757813 * 2560; time = 0.0138s; samplesPerSecond = 185279.0
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1721-1730]: CE.SM = 2.05895996 * 2560; Err = 0.56914062 * 2560; time = 0.0137s; samplesPerSecond = 187093.5
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1731-1740]: CE.SM = 2.02551270 * 2560; Err = 0.56015625 * 2560; time = 0.0139s; samplesPerSecond = 184345.1
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1741-1750]: CE.SM = 2.03193359 * 2560; Err = 0.56640625 * 2560; time = 0.0139s; samplesPerSecond = 184132.9
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1751-1760]: CE.SM = 2.01364746 * 2560; Err = 0.54882812 * 2560; time = 0.0137s; samplesPerSecond = 186888.6
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1761-1770]: CE.SM = 1.97338867 * 2560; Err = 0.54648438 * 2560; time = 0.0138s; samplesPerSecond = 185440.1
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1771-1780]: CE.SM = 2.05146484 * 2560; Err = 0.56679687 * 2560; time = 0.0134s; samplesPerSecond = 191344.6
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1781-1790]: CE.SM = 1.99951172 * 2560; Err = 0.55390625 * 2560; time = 0.0136s; samplesPerSecond = 188207.6
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1791-1800]: CE.SM = 2.05146484 * 2560; Err = 0.56171875 * 2560; time = 0.0136s; samplesPerSecond = 188832.3
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1801-1810]: CE.SM = 2.02548828 * 2560; Err = 0.55585938 * 2560; time = 0.0136s; samplesPerSecond = 188818.4
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1811-1820]: CE.SM = 1.97587891 * 2560; Err = 0.56562500 * 2560; time = 0.0136s; samplesPerSecond = 188332.2
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1821-1830]: CE.SM = 1.97197266 * 2560; Err = 0.55820313 * 2560; time = 0.0132s; samplesPerSecond = 193222.1
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1831-1840]: CE.SM = 2.01259766 * 2560; Err = 0.56796875 * 2560; time = 0.0135s; samplesPerSecond = 190249.7
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1841-1850]: CE.SM = 2.01948242 * 2560; Err = 0.57031250 * 2560; time = 0.0135s; samplesPerSecond = 189293.1
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1851-1860]: CE.SM = 1.95937500 * 2560; Err = 0.54765625 * 2560; time = 0.0135s; samplesPerSecond = 188971.7
12/20/2016 15:27:00:  Epoch[ 1 of 2]-Minibatch[1861-1870]: CE.SM = 2.00395508 * 2560; Err = 0.55273438 * 2560; time = 0.0150s; samplesPerSecond = 170894.5
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[1871-1880]: CE.SM = 2.03676758 * 2560; Err = 0.56640625 * 2560; time = 0.0172s; samplesPerSecond = 148586.7
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[1881-1890]: CE.SM = 2.04804687 * 2560; Err = 0.56835938 * 2560; time = 0.0144s; samplesPerSecond = 177432.8
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[1891-1900]: CE.SM = 1.99121094 * 2560; Err = 0.55234375 * 2560; time = 0.0144s; samplesPerSecond = 177506.6
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[1901-1910]: CE.SM = 1.96035156 * 2560; Err = 0.55390625 * 2560; time = 0.0143s; samplesPerSecond = 179021.0
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[1911-1920]: CE.SM = 2.00434570 * 2560; Err = 0.54960937 * 2560; time = 0.0137s; samplesPerSecond = 186520.9
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[1921-1930]: CE.SM = 2.00698242 * 2560; Err = 0.56210938 * 2560; time = 0.0140s; samplesPerSecond = 182570.2
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[1931-1940]: CE.SM = 1.99907227 * 2560; Err = 0.55625000 * 2560; time = 0.0137s; samplesPerSecond = 186752.3
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[1941-1950]: CE.SM = 1.98598633 * 2560; Err = 0.55937500 * 2560; time = 0.0139s; samplesPerSecond = 183525.7
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[1951-1960]: CE.SM = 1.97285156 * 2560; Err = 0.54531250 * 2560; time = 0.0140s; samplesPerSecond = 183210.5
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[1961-1970]: CE.SM = 2.03979492 * 2560; Err = 0.56250000 * 2560; time = 0.0139s; samplesPerSecond = 184331.8
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[1971-1980]: CE.SM = 1.98227539 * 2560; Err = 0.55312500 * 2560; time = 0.0138s; samplesPerSecond = 185319.2
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[1981-1990]: CE.SM = 1.96958008 * 2560; Err = 0.55156250 * 2560; time = 0.0138s; samplesPerSecond = 185803.5
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[1991-2000]: CE.SM = 1.98637695 * 2560; Err = 0.55351562 * 2560; time = 0.0137s; samplesPerSecond = 187161.9
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2001-2010]: CE.SM = 1.97304688 * 2560; Err = 0.55117187 * 2560; time = 0.0137s; samplesPerSecond = 186861.3
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2011-2020]: CE.SM = 2.01562500 * 2560; Err = 0.55664062 * 2560; time = 0.0139s; samplesPerSecond = 183696.9
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2021-2030]: CE.SM = 1.97363281 * 2560; Err = 0.54882812 * 2560; time = 0.0136s; samplesPerSecond = 187600.8
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2031-2040]: CE.SM = 2.00390625 * 2560; Err = 0.56718750 * 2560; time = 0.0139s; samplesPerSecond = 183842.0
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2041-2050]: CE.SM = 2.03706055 * 2560; Err = 0.57343750 * 2560; time = 0.0139s; samplesPerSecond = 183789.2
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2051-2060]: CE.SM = 1.98881836 * 2560; Err = 0.55625000 * 2560; time = 0.0139s; samplesPerSecond = 184784.2
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2061-2070]: CE.SM = 1.99233398 * 2560; Err = 0.55664062 * 2560; time = 0.0138s; samplesPerSecond = 186154.7
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2071-2080]: CE.SM = 1.98857422 * 2560; Err = 0.55390625 * 2560; time = 0.0137s; samplesPerSecond = 186398.7
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2081-2090]: CE.SM = 2.00820313 * 2560; Err = 0.56210938 * 2560; time = 0.0138s; samplesPerSecond = 185951.9
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2091-2100]: CE.SM = 1.96430664 * 2560; Err = 0.53750000 * 2560; time = 0.0139s; samplesPerSecond = 184784.2
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2101-2110]: CE.SM = 1.96435547 * 2560; Err = 0.56015625 * 2560; time = 0.0139s; samplesPerSecond = 183815.6
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2111-2120]: CE.SM = 2.01323242 * 2560; Err = 0.56640625 * 2560; time = 0.0137s; samplesPerSecond = 186588.9
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2121-2130]: CE.SM = 1.94594727 * 2560; Err = 0.55195313 * 2560; time = 0.0141s; samplesPerSecond = 181123.5
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2131-2140]: CE.SM = 1.98330078 * 2560; Err = 0.56054688 * 2560; time = 0.0139s; samplesPerSecond = 184797.5
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2141-2150]: CE.SM = 1.97622070 * 2560; Err = 0.56015625 * 2560; time = 0.0140s; samplesPerSecond = 183118.7
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2151-2160]: CE.SM = 1.94619141 * 2560; Err = 0.54492188 * 2560; time = 0.0140s; samplesPerSecond = 183420.5
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2161-2170]: CE.SM = 1.96440430 * 2560; Err = 0.53515625 * 2560; time = 0.0139s; samplesPerSecond = 183604.7
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2171-2180]: CE.SM = 1.99257813 * 2560; Err = 0.55546875 * 2560; time = 0.0139s; samplesPerSecond = 184517.8
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2181-2190]: CE.SM = 1.95458984 * 2560; Err = 0.55312500 * 2560; time = 0.0136s; samplesPerSecond = 188041.7
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2191-2200]: CE.SM = 1.92900391 * 2560; Err = 0.53671875 * 2560; time = 0.0138s; samplesPerSecond = 185709.1
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2201-2210]: CE.SM = 1.96098633 * 2560; Err = 0.55312500 * 2560; time = 0.0150s; samplesPerSecond = 171145.9
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2211-2220]: CE.SM = 1.93588867 * 2560; Err = 0.53750000 * 2560; time = 0.0140s; samplesPerSecond = 183368.0
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2221-2230]: CE.SM = 1.96728516 * 2560; Err = 0.54453125 * 2560; time = 0.0139s; samplesPerSecond = 184212.4
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2231-2240]: CE.SM = 1.94555664 * 2560; Err = 0.52734375 * 2560; time = 0.0138s; samplesPerSecond = 185037.9
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2241-2250]: CE.SM = 1.95527344 * 2560; Err = 0.54023438 * 2560; time = 0.0138s; samplesPerSecond = 184890.9
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2251-2260]: CE.SM = 2.02788086 * 2560; Err = 0.55820313 * 2560; time = 0.0137s; samplesPerSecond = 187449.7
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2261-2270]: CE.SM = 1.90000000 * 2560; Err = 0.54257813 * 2560; time = 0.0139s; samplesPerSecond = 184571.0
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2271-2280]: CE.SM = 1.92939453 * 2560; Err = 0.53867188 * 2560; time = 0.0140s; samplesPerSecond = 182271.3
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2281-2290]: CE.SM = 1.94399414 * 2560; Err = 0.53476563 * 2560; time = 0.0144s; samplesPerSecond = 177199.4
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2291-2300]: CE.SM = 1.91708984 * 2560; Err = 0.55429688 * 2560; time = 0.0139s; samplesPerSecond = 184345.1
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2301-2310]: CE.SM = 1.95463867 * 2560; Err = 0.54023438 * 2560; time = 0.0139s; samplesPerSecond = 183828.8
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2311-2320]: CE.SM = 1.99580078 * 2560; Err = 0.55859375 * 2560; time = 0.0138s; samplesPerSecond = 185372.9
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2321-2330]: CE.SM = 1.96337891 * 2560; Err = 0.55742187 * 2560; time = 0.0137s; samplesPerSecond = 186507.4
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2331-2340]: CE.SM = 1.92661133 * 2560; Err = 0.54648438 * 2560; time = 0.0138s; samplesPerSecond = 185480.4
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2341-2350]: CE.SM = 1.93349609 * 2560; Err = 0.53945312 * 2560; time = 0.0137s; samplesPerSecond = 186684.2
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2351-2360]: CE.SM = 1.95673828 * 2560; Err = 0.54296875 * 2560; time = 0.0137s; samplesPerSecond = 187285.1
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2361-2370]: CE.SM = 2.01811523 * 2560; Err = 0.57539063 * 2560; time = 0.0137s; samplesPerSecond = 186412.3
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2371-2380]: CE.SM = 1.93569336 * 2560; Err = 0.53750000 * 2560; time = 0.0138s; samplesPerSecond = 185453.5
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2381-2390]: CE.SM = 1.96171875 * 2560; Err = 0.54687500 * 2560; time = 0.0137s; samplesPerSecond = 186602.5
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2391-2400]: CE.SM = 1.98891602 * 2560; Err = 0.55507812 * 2560; time = 0.0148s; samplesPerSecond = 173347.8
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2401-2410]: CE.SM = 1.94653320 * 2560; Err = 0.54492188 * 2560; time = 0.0136s; samplesPerSecond = 188540.3
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2411-2420]: CE.SM = 1.95698242 * 2560; Err = 0.55820313 * 2560; time = 0.0139s; samplesPerSecond = 183710.1
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2421-2430]: CE.SM = 1.96635742 * 2560; Err = 0.55117187 * 2560; time = 0.0139s; samplesPerSecond = 183828.8
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2431-2440]: CE.SM = 1.90522461 * 2560; Err = 0.53398437 * 2560; time = 0.0139s; samplesPerSecond = 183604.7
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2441-2450]: CE.SM = 1.94438477 * 2560; Err = 0.54414063 * 2560; time = 0.0137s; samplesPerSecond = 187107.1
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2451-2460]: CE.SM = 1.99086914 * 2560; Err = 0.55234375 * 2560; time = 0.0139s; samplesPerSecond = 184013.8
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2461-2470]: CE.SM = 1.97045898 * 2560; Err = 0.54375000 * 2560; time = 0.0138s; samplesPerSecond = 184971.1
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2471-2480]: CE.SM = 1.97177734 * 2560; Err = 0.54179687 * 2560; time = 0.0136s; samplesPerSecond = 188512.5
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2481-2490]: CE.SM = 1.92241211 * 2560; Err = 0.53632813 * 2560; time = 0.0139s; samplesPerSecond = 184650.9
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2491-2500]: CE.SM = 1.95019531 * 2560; Err = 0.55468750 * 2560; time = 0.0138s; samplesPerSecond = 185426.6
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2501-2510]: CE.SM = 1.92841797 * 2560; Err = 0.54804688 * 2560; time = 0.0137s; samplesPerSecond = 186249.5
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2511-2520]: CE.SM = 1.92260742 * 2560; Err = 0.53945312 * 2560; time = 0.0138s; samplesPerSecond = 186087.1
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2521-2530]: CE.SM = 1.96210938 * 2560; Err = 0.55234375 * 2560; time = 0.0138s; samplesPerSecond = 185668.7
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2531-2540]: CE.SM = 1.92211914 * 2560; Err = 0.53750000 * 2560; time = 0.0135s; samplesPerSecond = 189083.4
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2541-2550]: CE.SM = 1.96440430 * 2560; Err = 0.55585938 * 2560; time = 0.0135s; samplesPerSecond = 190249.7
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2551-2560]: CE.SM = 1.98139648 * 2560; Err = 0.55273438 * 2560; time = 0.0138s; samplesPerSecond = 185158.4
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2561-2570]: CE.SM = 1.97363281 * 2560; Err = 0.55937500 * 2560; time = 0.0140s; samplesPerSecond = 183302.3
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2571-2580]: CE.SM = 1.96342773 * 2560; Err = 0.55000000 * 2560; time = 0.0139s; samplesPerSecond = 183565.2
12/20/2016 15:27:01:  Epoch[ 1 of 2]-Minibatch[2581-2590]: CE.SM = 1.91293945 * 2560; Err = 0.52539062 * 2560; time = 0.0139s; samplesPerSecond = 184119.7
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2591-2600]: CE.SM = 1.98510742 * 2560; Err = 0.55742187 * 2560; time = 0.0138s; samplesPerSecond = 185252.2
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2601-2610]: CE.SM = 1.98271484 * 2560; Err = 0.54960937 * 2560; time = 0.0138s; samplesPerSecond = 184837.5
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2611-2620]: CE.SM = 1.94370117 * 2560; Err = 0.54960937 * 2560; time = 0.0137s; samplesPerSecond = 186847.7
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2621-2630]: CE.SM = 1.91777344 * 2560; Err = 0.54531250 * 2560; time = 0.0139s; samplesPerSecond = 184650.9
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2631-2640]: CE.SM = 1.96645508 * 2560; Err = 0.54414063 * 2560; time = 0.0144s; samplesPerSecond = 177518.9
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2641-2650]: CE.SM = 1.96372070 * 2560; Err = 0.54453125 * 2560; time = 0.0139s; samplesPerSecond = 184371.6
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2651-2660]: CE.SM = 1.97504883 * 2560; Err = 0.55429688 * 2560; time = 0.0139s; samplesPerSecond = 184504.5
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2661-2670]: CE.SM = 1.93710937 * 2560; Err = 0.54921875 * 2560; time = 0.0139s; samplesPerSecond = 183696.9
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2671-2680]: CE.SM = 1.91523438 * 2560; Err = 0.54218750 * 2560; time = 0.0138s; samplesPerSecond = 185238.8
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2681-2690]: CE.SM = 1.86660156 * 2560; Err = 0.53320312 * 2560; time = 0.0136s; samplesPerSecond = 187587.0
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2691-2700]: CE.SM = 1.94663086 * 2560; Err = 0.55468750 * 2560; time = 0.0136s; samplesPerSecond = 187642.0
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2701-2710]: CE.SM = 1.98691406 * 2560; Err = 0.55468750 * 2560; time = 0.0138s; samplesPerSecond = 185655.2
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2711-2720]: CE.SM = 1.90556641 * 2560; Err = 0.53046875 * 2560; time = 0.0139s; samplesPerSecond = 183776.0
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2721-2730]: CE.SM = 1.94096680 * 2560; Err = 0.55117187 * 2560; time = 0.0137s; samplesPerSecond = 186398.7
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2731-2740]: CE.SM = 1.98857422 * 2560; Err = 0.55039063 * 2560; time = 0.0144s; samplesPerSecond = 177224.0
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2741-2750]: CE.SM = 1.92099609 * 2560; Err = 0.53750000 * 2560; time = 0.0138s; samplesPerSecond = 185641.8
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2751-2760]: CE.SM = 1.99335938 * 2560; Err = 0.55898437 * 2560; time = 0.0138s; samplesPerSecond = 186114.1
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2761-2770]: CE.SM = 1.91083984 * 2560; Err = 0.53750000 * 2560; time = 0.0139s; samplesPerSecond = 184810.9
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2771-2780]: CE.SM = 1.90512695 * 2560; Err = 0.54531250 * 2560; time = 0.0139s; samplesPerSecond = 183868.4
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2781-2790]: CE.SM = 1.96254883 * 2560; Err = 0.55078125 * 2560; time = 0.0138s; samplesPerSecond = 185816.9
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2791-2800]: CE.SM = 1.92402344 * 2560; Err = 0.53710938 * 2560; time = 0.0166s; samplesPerSecond = 153920.2
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2801-2810]: CE.SM = 1.95649414 * 2560; Err = 0.54179687 * 2560; time = 0.0206s; samplesPerSecond = 124247.7
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2811-2820]: CE.SM = 1.89633789 * 2560; Err = 0.52656250 * 2560; time = 0.0152s; samplesPerSecond = 168177.6
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2821-2830]: CE.SM = 1.97568359 * 2560; Err = 0.55156250 * 2560; time = 0.0144s; samplesPerSecond = 177518.9
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2831-2840]: CE.SM = 1.94858398 * 2560; Err = 0.54882812 * 2560; time = 0.0141s; samplesPerSecond = 181869.8
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2841-2850]: CE.SM = 1.95322266 * 2560; Err = 0.54882812 * 2560; time = 0.0139s; samplesPerSecond = 183921.3
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2851-2860]: CE.SM = 1.97250977 * 2560; Err = 0.55664062 * 2560; time = 0.0141s; samplesPerSecond = 181174.8
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2861-2870]: CE.SM = 1.95932617 * 2560; Err = 0.54023438 * 2560; time = 0.0139s; samplesPerSecond = 184757.5
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2871-2880]: CE.SM = 2.00102539 * 2560; Err = 0.55781250 * 2560; time = 0.0140s; samplesPerSecond = 182518.2
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2881-2890]: CE.SM = 1.90146484 * 2560; Err = 0.54218750 * 2560; time = 0.0140s; samplesPerSecond = 182922.5
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2891-2900]: CE.SM = 1.86699219 * 2560; Err = 0.52304688 * 2560; time = 0.0138s; samplesPerSecond = 185292.4
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2901-2910]: CE.SM = 1.92338867 * 2560; Err = 0.54296875 * 2560; time = 0.0140s; samplesPerSecond = 182948.6
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2911-2920]: CE.SM = 1.92846680 * 2560; Err = 0.54453125 * 2560; time = 0.0138s; samplesPerSecond = 185790.0
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2921-2930]: CE.SM = 1.89741211 * 2560; Err = 0.53593750 * 2560; time = 0.0139s; samplesPerSecond = 183591.5
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2931-2940]: CE.SM = 1.95268555 * 2560; Err = 0.54843750 * 2560; time = 0.0139s; samplesPerSecond = 184477.9
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2941-2950]: CE.SM = 1.90903320 * 2560; Err = 0.53828125 * 2560; time = 0.0140s; samplesPerSecond = 182870.2
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2951-2960]: CE.SM = 1.92890625 * 2560; Err = 0.54843750 * 2560; time = 0.0138s; samplesPerSecond = 185426.6
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2961-2970]: CE.SM = 1.93525391 * 2560; Err = 0.53750000 * 2560; time = 0.0138s; samplesPerSecond = 185897.9
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2971-2980]: CE.SM = 1.92534180 * 2560; Err = 0.53867188 * 2560; time = 0.0138s; samplesPerSecond = 185547.6
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2981-2990]: CE.SM = 1.95014648 * 2560; Err = 0.54296875 * 2560; time = 0.0142s; samplesPerSecond = 180663.4
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[2991-3000]: CE.SM = 1.91274414 * 2560; Err = 0.54179687 * 2560; time = 0.0140s; samplesPerSecond = 182948.6
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3001-3010]: CE.SM = 1.96337891 * 2560; Err = 0.55429688 * 2560; time = 0.0136s; samplesPerSecond = 187972.7
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3011-3020]: CE.SM = 1.94189453 * 2560; Err = 0.52265625 * 2560; time = 0.0137s; samplesPerSecond = 186902.2
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3021-3030]: CE.SM = 1.91362305 * 2560; Err = 0.55156250 * 2560; time = 0.0138s; samplesPerSecond = 185614.8
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3031-3040]: CE.SM = 1.89428711 * 2560; Err = 0.52812500 * 2560; time = 0.0140s; samplesPerSecond = 183302.3
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3041-3050]: CE.SM = 1.94248047 * 2560; Err = 0.54804688 * 2560; time = 0.0137s; samplesPerSecond = 186834.0
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3051-3060]: CE.SM = 1.90336914 * 2560; Err = 0.53984375 * 2560; time = 0.0141s; samplesPerSecond = 181650.5
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3061-3070]: CE.SM = 1.95395508 * 2560; Err = 0.53710938 * 2560; time = 0.0138s; samplesPerSecond = 186006.0
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3071-3080]: CE.SM = 1.92797852 * 2560; Err = 0.53007812 * 2560; time = 0.0138s; samplesPerSecond = 185870.9
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3081-3090]: CE.SM = 1.85307617 * 2560; Err = 0.51796875 * 2560; time = 0.0137s; samplesPerSecond = 186480.2
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3091-3100]: CE.SM = 1.91708984 * 2560; Err = 0.53710938 * 2560; time = 0.0139s; samplesPerSecond = 184252.2
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3101-3110]: CE.SM = 1.92670898 * 2560; Err = 0.55117187 * 2560; time = 0.0139s; samplesPerSecond = 184650.9
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3111-3120]: CE.SM = 1.92255859 * 2560; Err = 0.55078125 * 2560; time = 0.0136s; samplesPerSecond = 188540.3
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3121-3130]: CE.SM = 1.93789063 * 2560; Err = 0.53476563 * 2560; time = 0.0138s; samplesPerSecond = 185011.2
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3131-3140]: CE.SM = 1.92001953 * 2560; Err = 0.53671875 * 2560; time = 0.0139s; samplesPerSecond = 184119.7
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3141-3150]: CE.SM = 1.93608398 * 2560; Err = 0.54296875 * 2560; time = 0.0140s; samplesPerSecond = 183499.4
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3151-3160]: CE.SM = 1.95009766 * 2560; Err = 0.53632813 * 2560; time = 0.0138s; samplesPerSecond = 186073.6
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3161-3170]: CE.SM = 1.89521484 * 2560; Err = 0.52890625 * 2560; time = 0.0157s; samplesPerSecond = 163505.1
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3171-3180]: CE.SM = 1.82788086 * 2560; Err = 0.52343750 * 2560; time = 0.0161s; samplesPerSecond = 158858.2
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3181-3190]: CE.SM = 1.90927734 * 2560; Err = 0.53320312 * 2560; time = 0.0154s; samplesPerSecond = 165856.8
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3191-3200]: CE.SM = 1.88593750 * 2560; Err = 0.54023438 * 2560; time = 0.0143s; samplesPerSecond = 178571.4
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3201-3210]: CE.SM = 1.85766602 * 2560; Err = 0.52382812 * 2560; time = 0.0142s; samplesPerSecond = 180637.9
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3211-3220]: CE.SM = 1.90019531 * 2560; Err = 0.52343750 * 2560; time = 0.0145s; samplesPerSecond = 176187.2
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3221-3230]: CE.SM = 1.90996094 * 2560; Err = 0.53125000 * 2560; time = 0.0139s; samplesPerSecond = 183828.8
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3231-3240]: CE.SM = 1.91191406 * 2560; Err = 0.53984375 * 2560; time = 0.0141s; samplesPerSecond = 181960.3
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3241-3250]: CE.SM = 1.93149414 * 2560; Err = 0.54492188 * 2560; time = 0.0142s; samplesPerSecond = 180752.7
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3251-3260]: CE.SM = 1.89628906 * 2560; Err = 0.55039063 * 2560; time = 0.0138s; samplesPerSecond = 185938.4
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3261-3270]: CE.SM = 1.94335938 * 2560; Err = 0.55546875 * 2560; time = 0.0139s; samplesPerSecond = 184172.7
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3271-3280]: CE.SM = 1.88012695 * 2560; Err = 0.53710938 * 2560; time = 0.0138s; samplesPerSecond = 185051.3
12/20/2016 15:27:02:  Epoch[ 1 of 2]-Minibatch[3281-3290]: CE.SM = 1.92460937 * 2560; Err = 0.54140625 * 2560; time = 0.0135s; samplesPerSecond = 190009.6
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3291-3300]: CE.SM = 1.93857422 * 2560; Err = 0.55000000 * 2560; time = 0.0136s; samplesPerSecond = 188554.2
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3301-3310]: CE.SM = 1.94765625 * 2560; Err = 0.54492188 * 2560; time = 0.0139s; samplesPerSecond = 184185.9
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3311-3320]: CE.SM = 1.91728516 * 2560; Err = 0.54179687 * 2560; time = 0.0138s; samplesPerSecond = 185426.6
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3321-3330]: CE.SM = 1.93071289 * 2560; Err = 0.55273438 * 2560; time = 0.0137s; samplesPerSecond = 187312.5
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3331-3340]: CE.SM = 1.88588867 * 2560; Err = 0.54257813 * 2560; time = 0.0143s; samplesPerSecond = 179158.8
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3341-3350]: CE.SM = 1.88554687 * 2560; Err = 0.53281250 * 2560; time = 0.0137s; samplesPerSecond = 186793.1
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3351-3360]: CE.SM = 1.90917969 * 2560; Err = 0.54062500 * 2560; time = 0.0139s; samplesPerSecond = 184345.1
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3361-3370]: CE.SM = 1.91723633 * 2560; Err = 0.54921875 * 2560; time = 0.0139s; samplesPerSecond = 184398.2
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3371-3380]: CE.SM = 1.89375000 * 2560; Err = 0.52968750 * 2560; time = 0.0140s; samplesPerSecond = 183394.2
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3381-3390]: CE.SM = 1.92807617 * 2560; Err = 0.54335937 * 2560; time = 0.0137s; samplesPerSecond = 186480.2
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3391-3400]: CE.SM = 1.94638672 * 2560; Err = 0.54843750 * 2560; time = 0.0139s; samplesPerSecond = 184624.3
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3401-3410]: CE.SM = 1.83686523 * 2560; Err = 0.52500000 * 2560; time = 0.0140s; samplesPerSecond = 183158.0
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3411-3420]: CE.SM = 1.92480469 * 2560; Err = 0.54062500 * 2560; time = 0.0140s; samplesPerSecond = 183289.2
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3421-3430]: CE.SM = 1.90415039 * 2560; Err = 0.53750000 * 2560; time = 0.0135s; samplesPerSecond = 189363.1
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3431-3440]: CE.SM = 1.90375977 * 2560; Err = 0.52070313 * 2560; time = 0.0136s; samplesPerSecond = 187765.9
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3441-3450]: CE.SM = 1.90517578 * 2560; Err = 0.54609375 * 2560; time = 0.0136s; samplesPerSecond = 187807.2
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3451-3460]: CE.SM = 1.89497070 * 2560; Err = 0.53046875 * 2560; time = 0.0136s; samplesPerSecond = 188318.4
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3461-3470]: CE.SM = 1.90297852 * 2560; Err = 0.53554687 * 2560; time = 0.0135s; samplesPerSecond = 189433.2
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3471-3480]: CE.SM = 1.87651367 * 2560; Err = 0.53359375 * 2560; time = 0.0135s; samplesPerSecond = 190207.3
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3481-3490]: CE.SM = 1.91694336 * 2560; Err = 0.54375000 * 2560; time = 0.0137s; samplesPerSecond = 186412.3
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3491-3500]: CE.SM = 1.87861328 * 2560; Err = 0.52734375 * 2560; time = 0.0135s; samplesPerSecond = 189391.1
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3501-3510]: CE.SM = 1.90092773 * 2560; Err = 0.53125000 * 2560; time = 0.0152s; samplesPerSecond = 168798.6
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3511-3520]: CE.SM = 1.89189453 * 2560; Err = 0.53164062 * 2560; time = 0.0138s; samplesPerSecond = 185453.5
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3521-3530]: CE.SM = 1.93027344 * 2560; Err = 0.54140625 * 2560; time = 0.0138s; samplesPerSecond = 185628.3
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3531-3540]: CE.SM = 1.86113281 * 2560; Err = 0.52031250 * 2560; time = 0.0136s; samplesPerSecond = 188484.8
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3541-3550]: CE.SM = 1.95087891 * 2560; Err = 0.56250000 * 2560; time = 0.0134s; samplesPerSecond = 191044.8
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3551-3560]: CE.SM = 1.90029297 * 2560; Err = 0.53281250 * 2560; time = 0.0134s; samplesPerSecond = 191016.3
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3561-3570]: CE.SM = 1.88647461 * 2560; Err = 0.52304688 * 2560; time = 0.0136s; samplesPerSecond = 188193.8
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3571-3580]: CE.SM = 1.85664063 * 2560; Err = 0.53320312 * 2560; time = 0.0136s; samplesPerSecond = 188152.3
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3581-3590]: CE.SM = 1.92524414 * 2560; Err = 0.54726562 * 2560; time = 0.0137s; samplesPerSecond = 187353.6
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3591-3600]: CE.SM = 1.86660156 * 2560; Err = 0.51875000 * 2560; time = 0.0136s; samplesPerSecond = 188457.0
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3601-3610]: CE.SM = 1.94765625 * 2560; Err = 0.55507812 * 2560; time = 0.0137s; samplesPerSecond = 187408.5
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3611-3620]: CE.SM = 1.86611328 * 2560; Err = 0.52695313 * 2560; time = 0.0135s; samplesPerSecond = 189391.1
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3621-3630]: CE.SM = 1.90703125 * 2560; Err = 0.53203125 * 2560; time = 0.0136s; samplesPerSecond = 188346.1
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3631-3640]: CE.SM = 1.92260742 * 2560; Err = 0.54531250 * 2560; time = 0.0135s; samplesPerSecond = 189139.3
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3641-3650]: CE.SM = 1.89340820 * 2560; Err = 0.53984375 * 2560; time = 0.0137s; samplesPerSecond = 187202.9
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3651-3660]: CE.SM = 1.83222656 * 2560; Err = 0.51640625 * 2560; time = 0.0134s; samplesPerSecond = 190589.6
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3661-3670]: CE.SM = 1.88227539 * 2560; Err = 0.53476563 * 2560; time = 0.0135s; samplesPerSecond = 189097.4
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3671-3680]: CE.SM = 1.89931641 * 2560; Err = 0.54492188 * 2560; time = 0.0135s; samplesPerSecond = 189475.2
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3681-3690]: CE.SM = 1.89799805 * 2560; Err = 0.54804688 * 2560; time = 0.0136s; samplesPerSecond = 188721.0
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3691-3700]: CE.SM = 1.83325195 * 2560; Err = 0.51523438 * 2560; time = 0.0135s; samplesPerSecond = 189405.1
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3701-3710]: CE.SM = 1.96923828 * 2560; Err = 0.55585938 * 2560; time = 0.0135s; samplesPerSecond = 189503.3
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3711-3720]: CE.SM = 1.91289062 * 2560; Err = 0.53515625 * 2560; time = 0.0136s; samplesPerSecond = 188679.2
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3721-3730]: CE.SM = 1.87954102 * 2560; Err = 0.53125000 * 2560; time = 0.0136s; samplesPerSecond = 188804.5
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3731-3740]: CE.SM = 1.87856445 * 2560; Err = 0.54726562 * 2560; time = 0.0136s; samplesPerSecond = 188069.4
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3741-3750]: CE.SM = 1.83803711 * 2560; Err = 0.52382812 * 2560; time = 0.0136s; samplesPerSecond = 188124.6
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3751-3760]: CE.SM = 1.90791016 * 2560; Err = 0.54570312 * 2560; time = 0.0137s; samplesPerSecond = 187545.8
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3761-3770]: CE.SM = 1.87988281 * 2560; Err = 0.53242188 * 2560; time = 0.0134s; samplesPerSecond = 191273.2
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3771-3780]: CE.SM = 1.88076172 * 2560; Err = 0.53242188 * 2560; time = 0.0136s; samplesPerSecond = 187683.3
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3781-3790]: CE.SM = 1.93720703 * 2560; Err = 0.54531250 * 2560; time = 0.0136s; samplesPerSecond = 188804.5
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3791-3800]: CE.SM = 1.94760742 * 2560; Err = 0.54296875 * 2560; time = 0.0135s; samplesPerSecond = 189097.4
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3801-3810]: CE.SM = 1.90834961 * 2560; Err = 0.54492188 * 2560; time = 0.0136s; samplesPerSecond = 188138.5
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3811-3820]: CE.SM = 1.90668945 * 2560; Err = 0.54453125 * 2560; time = 0.0137s; samplesPerSecond = 187052.5
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3821-3830]: CE.SM = 1.90273438 * 2560; Err = 0.54375000 * 2560; time = 0.0136s; samplesPerSecond = 188290.7
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3831-3840]: CE.SM = 1.87407227 * 2560; Err = 0.53007812 * 2560; time = 0.0135s; samplesPerSecond = 189195.2
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3841-3850]: CE.SM = 1.88706055 * 2560; Err = 0.53750000 * 2560; time = 0.0137s; samplesPerSecond = 187052.5
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3851-3860]: CE.SM = 1.84047852 * 2560; Err = 0.52187500 * 2560; time = 0.0136s; samplesPerSecond = 187559.5
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3861-3870]: CE.SM = 1.84833984 * 2560; Err = 0.53007812 * 2560; time = 0.0135s; samplesPerSecond = 189939.2
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3871-3880]: CE.SM = 1.87617188 * 2560; Err = 0.53203125 * 2560; time = 0.0136s; samplesPerSecond = 187903.7
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3881-3890]: CE.SM = 1.89052734 * 2560; Err = 0.53476563 * 2560; time = 0.0136s; samplesPerSecond = 187917.5
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3891-3900]: CE.SM = 1.89599609 * 2560; Err = 0.53710938 * 2560; time = 0.0137s; samplesPerSecond = 187326.2
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3901-3910]: CE.SM = 1.93281250 * 2560; Err = 0.55273438 * 2560; time = 0.0133s; samplesPerSecond = 191947.2
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3911-3920]: CE.SM = 1.90908203 * 2560; Err = 0.54062500 * 2560; time = 0.0136s; samplesPerSecond = 188484.8
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3921-3930]: CE.SM = 1.89228516 * 2560; Err = 0.53437500 * 2560; time = 0.0137s; samplesPerSecond = 186997.8
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3931-3940]: CE.SM = 1.87070312 * 2560; Err = 0.52382812 * 2560; time = 0.0134s; samplesPerSecond = 190419.5
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3941-3950]: CE.SM = 1.85498047 * 2560; Err = 0.52421875 * 2560; time = 0.0136s; samplesPerSecond = 188110.8
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3951-3960]: CE.SM = 1.86435547 * 2560; Err = 0.52187500 * 2560; time = 0.0136s; samplesPerSecond = 188804.5
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3961-3970]: CE.SM = 1.88056641 * 2560; Err = 0.53671875 * 2560; time = 0.0136s; samplesPerSecond = 187628.3
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3971-3980]: CE.SM = 1.91367187 * 2560; Err = 0.54531250 * 2560; time = 0.0136s; samplesPerSecond = 188707.1
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3981-3990]: CE.SM = 1.87060547 * 2560; Err = 0.51953125 * 2560; time = 0.0138s; samplesPerSecond = 185709.1
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[3991-4000]: CE.SM = 1.87324219 * 2560; Err = 0.52929688 * 2560; time = 0.0137s; samplesPerSecond = 186656.9
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[4001-4010]: CE.SM = 1.90332031 * 2560; Err = 0.54648438 * 2560; time = 0.0139s; samplesPerSecond = 184544.4
12/20/2016 15:27:03:  Epoch[ 1 of 2]-Minibatch[4011-4020]: CE.SM = 1.86650391 * 2560; Err = 0.52070313 * 2560; time = 0.0135s; samplesPerSecond = 189896.9
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4021-4030]: CE.SM = 1.87138672 * 2560; Err = 0.53046875 * 2560; time = 0.0139s; samplesPerSecond = 183974.1
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4031-4040]: CE.SM = 1.88515625 * 2560; Err = 0.52890625 * 2560; time = 0.0137s; samplesPerSecond = 186575.3
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4041-4050]: CE.SM = 1.89375000 * 2560; Err = 0.52812500 * 2560; time = 0.0160s; samplesPerSecond = 160150.1
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4051-4060]: CE.SM = 1.86728516 * 2560; Err = 0.52695313 * 2560; time = 0.0136s; samplesPerSecond = 188373.8
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4061-4070]: CE.SM = 1.91240234 * 2560; Err = 0.53867188 * 2560; time = 0.0138s; samplesPerSecond = 185843.9
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4071-4080]: CE.SM = 1.84580078 * 2560; Err = 0.52968750 * 2560; time = 0.0138s; samplesPerSecond = 185386.3
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4081-4090]: CE.SM = 1.82929688 * 2560; Err = 0.51757812 * 2560; time = 0.0137s; samplesPerSecond = 186629.7
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4091-4100]: CE.SM = 1.84414063 * 2560; Err = 0.52343750 * 2560; time = 0.0138s; samplesPerSecond = 184984.5
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4101-4110]: CE.SM = 1.85585938 * 2560; Err = 0.52460938 * 2560; time = 0.0190s; samplesPerSecond = 134397.3
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4111-4120]: CE.SM = 1.88847656 * 2560; Err = 0.54101562 * 2560; time = 0.0149s; samplesPerSecond = 171685.3
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4121-4130]: CE.SM = 1.84257812 * 2560; Err = 0.52148438 * 2560; time = 0.0147s; samplesPerSecond = 174208.9
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4131-4140]: CE.SM = 1.87197266 * 2560; Err = 0.52539062 * 2560; time = 0.0143s; samplesPerSecond = 178895.9
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4141-4150]: CE.SM = 1.87421875 * 2560; Err = 0.51562500 * 2560; time = 0.0144s; samplesPerSecond = 177765.4
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4151-4160]: CE.SM = 1.94804687 * 2560; Err = 0.55351562 * 2560; time = 0.0138s; samplesPerSecond = 185319.2
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4161-4170]: CE.SM = 1.87666016 * 2560; Err = 0.53554687 * 2560; time = 0.0139s; samplesPerSecond = 184238.9
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4171-4180]: CE.SM = 1.81767578 * 2560; Err = 0.52109375 * 2560; time = 0.0139s; samplesPerSecond = 184770.8
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4181-4190]: CE.SM = 1.86787109 * 2560; Err = 0.51171875 * 2560; time = 0.0140s; samplesPerSecond = 183368.0
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4191-4200]: CE.SM = 1.86191406 * 2560; Err = 0.52539062 * 2560; time = 0.0136s; samplesPerSecond = 188000.3
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4201-4210]: CE.SM = 1.85234375 * 2560; Err = 0.51328125 * 2560; time = 0.0138s; samplesPerSecond = 185520.7
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4211-4220]: CE.SM = 1.90498047 * 2560; Err = 0.54921875 * 2560; time = 0.0138s; samplesPerSecond = 185574.5
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4221-4230]: CE.SM = 1.85625000 * 2560; Err = 0.52578125 * 2560; time = 0.0136s; samplesPerSecond = 187807.2
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4231-4240]: CE.SM = 1.83476562 * 2560; Err = 0.50585938 * 2560; time = 0.0139s; samplesPerSecond = 184358.3
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4241-4250]: CE.SM = 1.88623047 * 2560; Err = 0.50820312 * 2560; time = 0.0137s; samplesPerSecond = 186466.6
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4251-4260]: CE.SM = 1.88984375 * 2560; Err = 0.52539062 * 2560; time = 0.0138s; samplesPerSecond = 185790.0
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4261-4270]: CE.SM = 1.90039062 * 2560; Err = 0.53710938 * 2560; time = 0.0139s; samplesPerSecond = 183631.0
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4271-4280]: CE.SM = 1.86679687 * 2560; Err = 0.53242188 * 2560; time = 0.0138s; samplesPerSecond = 185104.8
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4281-4290]: CE.SM = 1.90302734 * 2560; Err = 0.53242188 * 2560; time = 0.0138s; samplesPerSecond = 185776.5
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4291-4300]: CE.SM = 1.84619141 * 2560; Err = 0.53750000 * 2560; time = 0.0137s; samplesPerSecond = 187038.8
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4301-4310]: CE.SM = 1.88281250 * 2560; Err = 0.53554687 * 2560; time = 0.0135s; samplesPerSecond = 189728.0
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4311-4320]: CE.SM = 1.87568359 * 2560; Err = 0.53437500 * 2560; time = 0.0134s; samplesPerSecond = 191444.8
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4321-4330]: CE.SM = 1.84062500 * 2560; Err = 0.52734375 * 2560; time = 0.0136s; samplesPerSecond = 187738.3
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4331-4340]: CE.SM = 1.85576172 * 2560; Err = 0.53046875 * 2560; time = 0.0135s; samplesPerSecond = 190207.3
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4341-4350]: CE.SM = 1.90996094 * 2560; Err = 0.55429688 * 2560; time = 0.0136s; samplesPerSecond = 188263.0
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4351-4360]: CE.SM = 1.84404297 * 2560; Err = 0.52304688 * 2560; time = 0.0136s; samplesPerSecond = 188665.3
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4361-4370]: CE.SM = 1.86757813 * 2560; Err = 0.53671875 * 2560; time = 0.0135s; samplesPerSecond = 189391.1
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4371-4380]: CE.SM = 1.86113281 * 2560; Err = 0.52812500 * 2560; time = 0.0136s; samplesPerSecond = 188443.1
12/20/2016 15:27:04:  Epoch[ 1 of 2]-Minibatch[4381-4390]: CE.SM = 1.85566406 * 2560; Err = 0.52187500 * 2560; time = 0.0136s; samplesPerSecond = 187738.3
12/20/2016 15:27:04: Finished Epoch[ 1 of 2]: [Training] CE.SM = 2.07966809 * 1124823; Err = 0.56705544 * 1124823; totalSamplesSeen = 1124823; learningRatePerSample = 0.00039062501; epochTime=6.99866s
12/20/2016 15:27:04: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/dptmodel2/cntkSpeech.dnn.1'

12/20/2016 15:27:04: Starting Epoch 2: learning rate per sample = 0.000391  effective momentum = 0.900000  momentum as time constant = 2429.8 samples

12/20/2016 15:27:04: Starting minibatch loop.
12/20/2016 15:27:04:  Epoch[ 2 of 2]-Minibatch[   1-  10, 0.23%]: CE.SM = 1.88272705 * 2560; Err = 0.53046875 * 2560; time = 0.0250s; samplesPerSecond = 102535.3
12/20/2016 15:27:04:  Epoch[ 2 of 2]-Minibatch[  11-  20, 0.46%]: CE.SM = 1.83230095 * 2560; Err = 0.52968750 * 2560; time = 0.0136s; samplesPerSecond = 188138.5
12/20/2016 15:27:04:  Epoch[ 2 of 2]-Minibatch[  21-  30, 0.68%]: CE.SM = 1.90064125 * 2560; Err = 0.53632813 * 2560; time = 0.0151s; samplesPerSecond = 169783.8
12/20/2016 15:27:04:  Epoch[ 2 of 2]-Minibatch[  31-  40, 0.91%]: CE.SM = 1.87685852 * 2560; Err = 0.52226562 * 2560; time = 0.0134s; samplesPerSecond = 190476.2
12/20/2016 15:27:04:  Epoch[ 2 of 2]-Minibatch[  41-  50, 1.14%]: CE.SM = 1.81960068 * 2560; Err = 0.51601562 * 2560; time = 0.0131s; samplesPerSecond = 195196.3
12/20/2016 15:27:04:  Epoch[ 2 of 2]-Minibatch[  51-  60, 1.37%]: CE.SM = 1.84842758 * 2560; Err = 0.53085938 * 2560; time = 0.0135s; samplesPerSecond = 189083.4
12/20/2016 15:27:04:  Epoch[ 2 of 2]-Minibatch[  61-  70, 1.59%]: CE.SM = 1.91667633 * 2560; Err = 0.53867188 * 2560; time = 0.0138s; samplesPerSecond = 185185.2
12/20/2016 15:27:04:  Epoch[ 2 of 2]-Minibatch[  71-  80, 1.82%]: CE.SM = 1.81425629 * 2560; Err = 0.52460938 * 2560; time = 0.0141s; samplesPerSecond = 181844.0
12/20/2016 15:27:04:  Epoch[ 2 of 2]-Minibatch[  81-  90, 2.05%]: CE.SM = 1.79854126 * 2560; Err = 0.52343750 * 2560; time = 0.0148s; samplesPerSecond = 173089.9
12/20/2016 15:27:04:  Epoch[ 2 of 2]-Minibatch[  91- 100, 2.28%]: CE.SM = 1.84776001 * 2560; Err = 0.53476563 * 2560; time = 0.0138s; samplesPerSecond = 186141.2
12/20/2016 15:27:04:  Epoch[ 2 of 2]-Minibatch[ 101- 110, 2.51%]: CE.SM = 1.83752441 * 2560; Err = 0.53789062 * 2560; time = 0.0138s; samplesPerSecond = 184864.2
12/20/2016 15:27:04:  Epoch[ 2 of 2]-Minibatch[ 111- 120, 2.73%]: CE.SM = 1.84032440 * 2560; Err = 0.53710938 * 2560; time = 0.0138s; samplesPerSecond = 185965.4
12/20/2016 15:27:04:  Epoch[ 2 of 2]-Minibatch[ 121- 130, 2.96%]: CE.SM = 1.87525177 * 2560; Err = 0.52695313 * 2560; time = 0.0139s; samplesPerSecond = 184757.5
12/20/2016 15:27:04:  Epoch[ 2 of 2]-Minibatch[ 131- 140, 3.19%]: CE.SM = 1.87568054 * 2560; Err = 0.52265625 * 2560; time = 0.0137s; samplesPerSecond = 186943.2
12/20/2016 15:27:04:  Epoch[ 2 of 2]-Minibatch[ 141- 150, 3.42%]: CE.SM = 1.80683289 * 2560; Err = 0.50312500 * 2560; time = 0.0137s; samplesPerSecond = 186875.0
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 151- 160, 3.64%]: CE.SM = 1.86755981 * 2560; Err = 0.53359375 * 2560; time = 0.0137s; samplesPerSecond = 186697.8
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 161- 170, 3.87%]: CE.SM = 1.89820251 * 2560; Err = 0.53125000 * 2560; time = 0.0138s; samplesPerSecond = 184957.7
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 171- 180, 4.10%]: CE.SM = 1.89616699 * 2560; Err = 0.52812500 * 2560; time = 0.0136s; samplesPerSecond = 187779.7
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 181- 190, 4.33%]: CE.SM = 1.87921448 * 2560; Err = 0.54648438 * 2560; time = 0.0136s; samplesPerSecond = 188707.1
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 191- 200, 4.56%]: CE.SM = 1.84592896 * 2560; Err = 0.52929688 * 2560; time = 0.0138s; samplesPerSecond = 185938.4
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 201- 210, 4.78%]: CE.SM = 1.87736511 * 2560; Err = 0.52500000 * 2560; time = 0.0138s; samplesPerSecond = 184837.5
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 211- 220, 5.01%]: CE.SM = 1.83711243 * 2560; Err = 0.51875000 * 2560; time = 0.0138s; samplesPerSecond = 185924.9
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 221- 230, 5.24%]: CE.SM = 1.88039856 * 2560; Err = 0.53710938 * 2560; time = 0.0139s; samplesPerSecond = 184027.0
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 231- 240, 5.47%]: CE.SM = 1.88231812 * 2560; Err = 0.52851563 * 2560; time = 0.0136s; samplesPerSecond = 187779.7
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 241- 250, 5.69%]: CE.SM = 1.86379089 * 2560; Err = 0.53007812 * 2560; time = 0.0138s; samplesPerSecond = 184957.7
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 251- 260, 5.92%]: CE.SM = 1.86029663 * 2560; Err = 0.53007812 * 2560; time = 0.0136s; samplesPerSecond = 188526.4
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 261- 270, 6.15%]: CE.SM = 1.86083374 * 2560; Err = 0.53632813 * 2560; time = 0.0134s; samplesPerSecond = 190916.5
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 271- 280, 6.38%]: CE.SM = 1.78636169 * 2560; Err = 0.51054687 * 2560; time = 0.0135s; samplesPerSecond = 189265.1
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 281- 290, 6.61%]: CE.SM = 1.85677490 * 2560; Err = 0.52617187 * 2560; time = 0.0136s; samplesPerSecond = 188041.7
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 291- 300, 6.83%]: CE.SM = 1.90758667 * 2560; Err = 0.53593750 * 2560; time = 0.0136s; samplesPerSecond = 188000.3
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 301- 310, 7.06%]: CE.SM = 1.86036377 * 2560; Err = 0.52539062 * 2560; time = 0.0135s; samplesPerSecond = 189643.7
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 311- 320, 7.29%]: CE.SM = 1.84779663 * 2560; Err = 0.53242188 * 2560; time = 0.0134s; samplesPerSecond = 190575.4
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 321- 330, 7.52%]: CE.SM = 1.87850342 * 2560; Err = 0.53085938 * 2560; time = 0.0136s; samplesPerSecond = 187876.1
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 331- 340, 7.74%]: CE.SM = 1.83889771 * 2560; Err = 0.52031250 * 2560; time = 0.0136s; samplesPerSecond = 187683.3
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 341- 350, 7.97%]: CE.SM = 1.88981323 * 2560; Err = 0.54023438 * 2560; time = 0.0138s; samplesPerSecond = 185037.9
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 351- 360, 8.20%]: CE.SM = 1.86324463 * 2560; Err = 0.52304688 * 2560; time = 0.0138s; samplesPerSecond = 184917.7
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 361- 370, 8.43%]: CE.SM = 1.85560303 * 2560; Err = 0.52421875 * 2560; time = 0.0135s; samplesPerSecond = 189447.2
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 371- 380, 8.66%]: CE.SM = 1.89002075 * 2560; Err = 0.53398437 * 2560; time = 0.0139s; samplesPerSecond = 184557.7
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 381- 390, 8.88%]: CE.SM = 1.79342651 * 2560; Err = 0.52187500 * 2560; time = 0.0135s; samplesPerSecond = 188957.8
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 391- 400, 9.11%]: CE.SM = 1.84598999 * 2560; Err = 0.51992187 * 2560; time = 0.0137s; samplesPerSecond = 187244.0
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 401- 410, 9.34%]: CE.SM = 1.88099976 * 2560; Err = 0.52382812 * 2560; time = 0.0137s; samplesPerSecond = 186317.3
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 411- 420, 9.57%]: CE.SM = 1.91538086 * 2560; Err = 0.54179687 * 2560; time = 0.0139s; samplesPerSecond = 184384.9
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 421- 430, 9.79%]: CE.SM = 1.86830444 * 2560; Err = 0.53281250 * 2560; time = 0.0138s; samplesPerSecond = 185507.2
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 431- 440, 10.02%]: CE.SM = 1.86873169 * 2560; Err = 0.52812500 * 2560; time = 0.0137s; samplesPerSecond = 186222.4
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 441- 450, 10.25%]: CE.SM = 1.80231934 * 2560; Err = 0.51015625 * 2560; time = 0.0137s; samplesPerSecond = 186765.9
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 451- 460, 10.48%]: CE.SM = 1.82319946 * 2560; Err = 0.52656250 * 2560; time = 0.0137s; samplesPerSecond = 186236.0
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 461- 470, 10.71%]: CE.SM = 1.78828125 * 2560; Err = 0.51718750 * 2560; time = 0.0138s; samplesPerSecond = 185520.7
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 471- 480, 10.93%]: CE.SM = 1.83073120 * 2560; Err = 0.52031250 * 2560; time = 0.0138s; samplesPerSecond = 185870.9
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 481- 490, 11.16%]: CE.SM = 1.86691284 * 2560; Err = 0.52539062 * 2560; time = 0.0136s; samplesPerSecond = 187958.9
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 491- 500, 11.39%]: CE.SM = 1.84379883 * 2560; Err = 0.52656250 * 2560; time = 0.0137s; samplesPerSecond = 186330.9
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 501- 510, 11.62%]: CE.SM = 1.80936890 * 2560; Err = 0.51562500 * 2560; time = 0.0137s; samplesPerSecond = 186861.3
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 511- 520, 11.85%]: CE.SM = 1.86774902 * 2560; Err = 0.52539062 * 2560; time = 0.0137s; samplesPerSecond = 186344.4
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 521- 530, 12.07%]: CE.SM = 1.81527710 * 2560; Err = 0.52070313 * 2560; time = 0.0139s; samplesPerSecond = 183591.5
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 531- 540, 12.30%]: CE.SM = 1.81638184 * 2560; Err = 0.51718750 * 2560; time = 0.0139s; samplesPerSecond = 184677.5
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 541- 550, 12.53%]: CE.SM = 1.84501953 * 2560; Err = 0.52929688 * 2560; time = 0.0138s; samplesPerSecond = 185453.5
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 551- 560, 12.76%]: CE.SM = 1.84465332 * 2560; Err = 0.52187500 * 2560; time = 0.0137s; samplesPerSecond = 187463.4
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 561- 570, 12.98%]: CE.SM = 1.87416992 * 2560; Err = 0.54648438 * 2560; time = 0.0138s; samplesPerSecond = 186154.7
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 571- 580, 13.21%]: CE.SM = 1.86470947 * 2560; Err = 0.52968750 * 2560; time = 0.0139s; samplesPerSecond = 184199.2
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 581- 590, 13.44%]: CE.SM = 1.82187500 * 2560; Err = 0.51718750 * 2560; time = 0.0138s; samplesPerSecond = 185587.9
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 591- 600, 13.67%]: CE.SM = 1.84006348 * 2560; Err = 0.53242188 * 2560; time = 0.0138s; samplesPerSecond = 185466.9
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 601- 610, 13.90%]: CE.SM = 1.89018555 * 2560; Err = 0.53125000 * 2560; time = 0.0138s; samplesPerSecond = 185790.0
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 611- 620, 14.12%]: CE.SM = 1.83461914 * 2560; Err = 0.52617187 * 2560; time = 0.0138s; samplesPerSecond = 184931.0
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 621- 630, 14.35%]: CE.SM = 1.82160645 * 2560; Err = 0.51171875 * 2560; time = 0.0136s; samplesPerSecond = 188540.3
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 631- 640, 14.58%]: CE.SM = 1.87395020 * 2560; Err = 0.53437500 * 2560; time = 0.0138s; samplesPerSecond = 186033.0
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 641- 650, 14.81%]: CE.SM = 1.81381836 * 2560; Err = 0.51484375 * 2560; time = 0.0138s; samplesPerSecond = 185574.5
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 651- 660, 15.03%]: CE.SM = 1.86229248 * 2560; Err = 0.52695313 * 2560; time = 0.0137s; samplesPerSecond = 186344.4
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 661- 670, 15.26%]: CE.SM = 1.83739014 * 2560; Err = 0.51562500 * 2560; time = 0.0138s; samplesPerSecond = 184984.5
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 671- 680, 15.49%]: CE.SM = 1.84979248 * 2560; Err = 0.53085938 * 2560; time = 0.0138s; samplesPerSecond = 185426.6
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 681- 690, 15.72%]: CE.SM = 1.81499023 * 2560; Err = 0.52265625 * 2560; time = 0.0182s; samplesPerSecond = 140543.5
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 691- 700, 15.95%]: CE.SM = 1.80056152 * 2560; Err = 0.51523438 * 2560; time = 0.0149s; samplesPerSecond = 171616.3
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 701- 710, 16.17%]: CE.SM = 1.80628662 * 2560; Err = 0.52265625 * 2560; time = 0.0143s; samplesPerSecond = 178721.0
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 711- 720, 16.40%]: CE.SM = 1.91826172 * 2560; Err = 0.53281250 * 2560; time = 0.0139s; samplesPerSecond = 184650.9
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 721- 730, 16.63%]: CE.SM = 1.82839355 * 2560; Err = 0.50273437 * 2560; time = 0.0139s; samplesPerSecond = 183987.4
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 731- 740, 16.86%]: CE.SM = 1.82174072 * 2560; Err = 0.53125000 * 2560; time = 0.0136s; samplesPerSecond = 187972.7
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 741- 750, 17.08%]: CE.SM = 1.83923340 * 2560; Err = 0.53945312 * 2560; time = 0.0136s; samplesPerSecond = 187972.7
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 751- 760, 17.31%]: CE.SM = 1.83717041 * 2560; Err = 0.52226562 * 2560; time = 0.0141s; samplesPerSecond = 181573.2
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 761- 770, 17.54%]: CE.SM = 1.89501953 * 2560; Err = 0.53945312 * 2560; time = 0.0139s; samplesPerSecond = 183776.0
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 771- 780, 17.77%]: CE.SM = 1.82335205 * 2560; Err = 0.52421875 * 2560; time = 0.0137s; samplesPerSecond = 186738.6
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 781- 790, 18.00%]: CE.SM = 1.83837891 * 2560; Err = 0.52460938 * 2560; time = 0.0138s; samplesPerSecond = 185225.4
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 791- 800, 18.22%]: CE.SM = 1.82939453 * 2560; Err = 0.52617187 * 2560; time = 0.0138s; samplesPerSecond = 185359.5
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 801- 810, 18.45%]: CE.SM = 1.85537109 * 2560; Err = 0.52382812 * 2560; time = 0.0137s; samplesPerSecond = 186970.5
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 811- 820, 18.68%]: CE.SM = 1.83397217 * 2560; Err = 0.53476563 * 2560; time = 0.0138s; samplesPerSecond = 185118.2
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 821- 830, 18.91%]: CE.SM = 1.89172363 * 2560; Err = 0.53789062 * 2560; time = 0.0139s; samplesPerSecond = 184093.2
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 831- 840, 19.13%]: CE.SM = 1.82828369 * 2560; Err = 0.50937500 * 2560; time = 0.0139s; samplesPerSecond = 184637.6
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 841- 850, 19.36%]: CE.SM = 1.82821045 * 2560; Err = 0.51914063 * 2560; time = 0.0139s; samplesPerSecond = 183552.0
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 851- 860, 19.59%]: CE.SM = 1.81584473 * 2560; Err = 0.50976562 * 2560; time = 0.0140s; samplesPerSecond = 182818.0
12/20/2016 15:27:05:  Epoch[ 2 of 2]-Minibatch[ 861- 870, 19.82%]: CE.SM = 1.83343506 * 2560; Err = 0.52226562 * 2560; time = 0.0138s; samplesPerSecond = 185803.5
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[ 871- 880, 20.05%]: CE.SM = 1.81276855 * 2560; Err = 0.50351563 * 2560; time = 0.0138s; samplesPerSecond = 185561.0
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[ 881- 890, 20.27%]: CE.SM = 1.88613281 * 2560; Err = 0.53125000 * 2560; time = 0.0137s; samplesPerSecond = 186861.3
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[ 891- 900, 20.50%]: CE.SM = 1.85693359 * 2560; Err = 0.52929688 * 2560; time = 0.0138s; samplesPerSecond = 184931.0
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[ 901- 910, 20.73%]: CE.SM = 1.82524414 * 2560; Err = 0.52109375 * 2560; time = 0.0135s; samplesPerSecond = 189335.1
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[ 911- 920, 20.96%]: CE.SM = 1.81971436 * 2560; Err = 0.51054687 * 2560; time = 0.0139s; samplesPerSecond = 183894.8
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[ 921- 930, 21.18%]: CE.SM = 1.86175537 * 2560; Err = 0.52773437 * 2560; time = 0.0137s; samplesPerSecond = 187477.1
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[ 931- 940, 21.41%]: CE.SM = 1.82818604 * 2560; Err = 0.52773437 * 2560; time = 0.0136s; samplesPerSecond = 187614.5
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[ 941- 950, 21.64%]: CE.SM = 1.85286865 * 2560; Err = 0.52421875 * 2560; time = 0.0139s; samplesPerSecond = 184040.3
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[ 951- 960, 21.87%]: CE.SM = 1.82493896 * 2560; Err = 0.51718750 * 2560; time = 0.0138s; samplesPerSecond = 184971.1
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[ 961- 970, 22.10%]: CE.SM = 1.79885254 * 2560; Err = 0.50664062 * 2560; time = 0.0138s; samplesPerSecond = 185709.1
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[ 971- 980, 22.32%]: CE.SM = 1.83464355 * 2560; Err = 0.50898438 * 2560; time = 0.0139s; samplesPerSecond = 184770.8
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[ 981- 990, 22.55%]: CE.SM = 1.84475098 * 2560; Err = 0.52382812 * 2560; time = 0.0139s; samplesPerSecond = 184424.8
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[ 991-1000, 22.78%]: CE.SM = 1.83276367 * 2560; Err = 0.53085938 * 2560; time = 0.0139s; samplesPerSecond = 184544.4
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1001-1010, 23.01%]: CE.SM = 1.82891846 * 2560; Err = 0.52148438 * 2560; time = 0.0139s; samplesPerSecond = 184544.4
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1011-1020, 23.23%]: CE.SM = 1.82109375 * 2560; Err = 0.51640625 * 2560; time = 0.0147s; samplesPerSecond = 174517.7
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1021-1030, 23.46%]: CE.SM = 1.81221924 * 2560; Err = 0.51757812 * 2560; time = 0.0189s; samplesPerSecond = 135708.2
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1031-1040, 23.69%]: CE.SM = 1.80135498 * 2560; Err = 0.52421875 * 2560; time = 0.0157s; samplesPerSecond = 162570.6
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1041-1050, 23.92%]: CE.SM = 1.86221924 * 2560; Err = 0.53750000 * 2560; time = 0.0140s; samplesPerSecond = 182987.8
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1051-1060, 24.15%]: CE.SM = 1.83395996 * 2560; Err = 0.52226562 * 2560; time = 0.0143s; samplesPerSecond = 178434.5
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1061-1070, 24.37%]: CE.SM = 1.85870361 * 2560; Err = 0.52031250 * 2560; time = 0.0142s; samplesPerSecond = 180472.3
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1071-1080, 24.60%]: CE.SM = 1.87565918 * 2560; Err = 0.54726562 * 2560; time = 0.0138s; samplesPerSecond = 185198.6
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1081-1090, 24.83%]: CE.SM = 1.81761475 * 2560; Err = 0.51757812 * 2560; time = 0.0138s; samplesPerSecond = 185372.9
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1091-1100, 25.06%]: CE.SM = 1.86761475 * 2560; Err = 0.52265625 * 2560; time = 0.0138s; samplesPerSecond = 185252.2
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1101-1110, 25.28%]: CE.SM = 1.82877197 * 2560; Err = 0.52656250 * 2560; time = 0.0136s; samplesPerSecond = 187642.0
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1111-1120, 25.51%]: CE.SM = 1.81489258 * 2560; Err = 0.52656250 * 2560; time = 0.0136s; samplesPerSecond = 188401.5
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1121-1130, 25.74%]: CE.SM = 1.83378906 * 2560; Err = 0.52343750 * 2560; time = 0.0136s; samplesPerSecond = 188498.6
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1131-1140, 25.97%]: CE.SM = 1.83747559 * 2560; Err = 0.51992187 * 2560; time = 0.0136s; samplesPerSecond = 188359.9
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1141-1150, 26.20%]: CE.SM = 1.84008789 * 2560; Err = 0.52734375 * 2560; time = 0.0136s; samplesPerSecond = 188582.0
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1151-1160, 26.42%]: CE.SM = 1.85507812 * 2560; Err = 0.51875000 * 2560; time = 0.0135s; samplesPerSecond = 189559.4
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1161-1170, 26.65%]: CE.SM = 1.73210449 * 2560; Err = 0.49804688 * 2560; time = 0.0135s; samplesPerSecond = 189587.5
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1171-1180, 26.88%]: CE.SM = 1.81088867 * 2560; Err = 0.51328125 * 2560; time = 0.0138s; samplesPerSecond = 185965.4
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1181-1190, 27.11%]: CE.SM = 1.86279297 * 2560; Err = 0.53046875 * 2560; time = 0.0138s; samplesPerSecond = 185547.6
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1191-1200, 27.33%]: CE.SM = 1.80781250 * 2560; Err = 0.52226562 * 2560; time = 0.0136s; samplesPerSecond = 188069.4
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1201-1210, 27.56%]: CE.SM = 1.88918457 * 2560; Err = 0.54609375 * 2560; time = 0.0138s; samplesPerSecond = 185870.9
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1211-1220, 27.79%]: CE.SM = 1.84211426 * 2560; Err = 0.52773437 * 2560; time = 0.0137s; samplesPerSecond = 186915.9
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1221-1230, 28.02%]: CE.SM = 1.78420410 * 2560; Err = 0.50664062 * 2560; time = 0.0138s; samplesPerSecond = 185534.1
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1231-1240, 28.25%]: CE.SM = 1.84794922 * 2560; Err = 0.52187500 * 2560; time = 0.0138s; samplesPerSecond = 185574.5
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1241-1250, 28.47%]: CE.SM = 1.80800781 * 2560; Err = 0.51718750 * 2560; time = 0.0139s; samplesPerSecond = 184571.0
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1251-1260, 28.70%]: CE.SM = 1.82592773 * 2560; Err = 0.52031250 * 2560; time = 0.0140s; samplesPerSecond = 183512.5
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1261-1270, 28.93%]: CE.SM = 1.80766602 * 2560; Err = 0.51210937 * 2560; time = 0.0137s; samplesPerSecond = 187230.3
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1271-1280, 29.16%]: CE.SM = 1.86494141 * 2560; Err = 0.53476563 * 2560; time = 0.0137s; samplesPerSecond = 186779.5
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1281-1290, 29.38%]: CE.SM = 1.86328125 * 2560; Err = 0.51562500 * 2560; time = 0.0134s; samplesPerSecond = 191459.1
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1291-1300, 29.61%]: CE.SM = 1.80695801 * 2560; Err = 0.51992187 * 2560; time = 0.0133s; samplesPerSecond = 192742.1
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1301-1310, 29.84%]: CE.SM = 1.87536621 * 2560; Err = 0.53242188 * 2560; time = 0.0137s; samplesPerSecond = 187490.8
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1311-1320, 30.07%]: CE.SM = 1.83908691 * 2560; Err = 0.52109375 * 2560; time = 0.0136s; samplesPerSecond = 188429.3
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1321-1330, 30.30%]: CE.SM = 1.79235840 * 2560; Err = 0.51250000 * 2560; time = 0.0138s; samplesPerSecond = 184984.5
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1331-1340, 30.52%]: CE.SM = 1.83505859 * 2560; Err = 0.51367188 * 2560; time = 0.0136s; samplesPerSecond = 188000.3
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1341-1350, 30.75%]: CE.SM = 1.86105957 * 2560; Err = 0.53476563 * 2560; time = 0.0139s; samplesPerSecond = 184624.3
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1351-1360, 30.98%]: CE.SM = 1.78269043 * 2560; Err = 0.50820312 * 2560; time = 0.0154s; samplesPerSecond = 165889.1
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1361-1370, 31.21%]: CE.SM = 1.81145020 * 2560; Err = 0.51718750 * 2560; time = 0.0141s; samplesPerSecond = 181174.8
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1371-1380, 31.44%]: CE.SM = 1.80651855 * 2560; Err = 0.50820312 * 2560; time = 0.0138s; samplesPerSecond = 185078.1
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1381-1390, 31.66%]: CE.SM = 1.82094727 * 2560; Err = 0.51523438 * 2560; time = 0.0138s; samplesPerSecond = 184997.8
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1391-1400, 31.89%]: CE.SM = 1.81860352 * 2560; Err = 0.51054687 * 2560; time = 0.0137s; samplesPerSecond = 186330.9
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1401-1410, 32.12%]: CE.SM = 1.81735840 * 2560; Err = 0.51992187 * 2560; time = 0.0138s; samplesPerSecond = 186127.7
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1411-1420, 32.35%]: CE.SM = 1.86137695 * 2560; Err = 0.53476563 * 2560; time = 0.0138s; samplesPerSecond = 185104.8
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1421-1430, 32.57%]: CE.SM = 1.86530762 * 2560; Err = 0.54101562 * 2560; time = 0.0139s; samplesPerSecond = 184265.5
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1431-1440, 32.80%]: CE.SM = 1.86884766 * 2560; Err = 0.53203125 * 2560; time = 0.0138s; samplesPerSecond = 185305.8
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1441-1450, 33.03%]: CE.SM = 1.78491211 * 2560; Err = 0.51640625 * 2560; time = 0.0139s; samplesPerSecond = 183789.2
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1451-1460, 33.26%]: CE.SM = 1.85546875 * 2560; Err = 0.52187500 * 2560; time = 0.0138s; samplesPerSecond = 185466.9
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1461-1470, 33.49%]: CE.SM = 1.84155273 * 2560; Err = 0.51875000 * 2560; time = 0.0138s; samplesPerSecond = 185749.5
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1471-1480, 33.71%]: CE.SM = 1.86093750 * 2560; Err = 0.52304688 * 2560; time = 0.0138s; samplesPerSecond = 184957.7
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1481-1490, 33.94%]: CE.SM = 1.82763672 * 2560; Err = 0.52148438 * 2560; time = 0.0137s; samplesPerSecond = 186602.5
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1491-1500, 34.17%]: CE.SM = 1.84399414 * 2560; Err = 0.52148438 * 2560; time = 0.0138s; samplesPerSecond = 185965.4
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1501-1510, 34.40%]: CE.SM = 1.86037598 * 2560; Err = 0.53632813 * 2560; time = 0.0139s; samplesPerSecond = 184292.0
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1511-1520, 34.62%]: CE.SM = 1.79291992 * 2560; Err = 0.50976562 * 2560; time = 0.0137s; samplesPerSecond = 186779.5
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1521-1530, 34.85%]: CE.SM = 1.79309082 * 2560; Err = 0.50546875 * 2560; time = 0.0139s; samplesPerSecond = 184610.9
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1531-1540, 35.08%]: CE.SM = 1.77463379 * 2560; Err = 0.51718750 * 2560; time = 0.0138s; samplesPerSecond = 186168.3
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1541-1550, 35.31%]: CE.SM = 1.79702148 * 2560; Err = 0.50625000 * 2560; time = 0.0139s; samplesPerSecond = 184810.9
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1551-1560, 35.54%]: CE.SM = 1.83161621 * 2560; Err = 0.52421875 * 2560; time = 0.0136s; samplesPerSecond = 188568.1
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1561-1570, 35.76%]: CE.SM = 1.77692871 * 2560; Err = 0.50937500 * 2560; time = 0.0140s; samplesPerSecond = 183407.4
12/20/2016 15:27:06:  Epoch[ 2 of 2]-Minibatch[1571-1580, 35.99%]: CE.SM = 1.85539551 * 2560; Err = 0.53203125 * 2560; time = 0.0146s; samplesPerSecond = 174911.2
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1581-1590, 36.22%]: CE.SM = 1.84521484 * 2560; Err = 0.51601562 * 2560; time = 0.0137s; samplesPerSecond = 186493.8
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1591-1600, 36.45%]: CE.SM = 1.84174805 * 2560; Err = 0.53007812 * 2560; time = 0.0137s; samplesPerSecond = 186371.6
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1601-1610, 36.67%]: CE.SM = 1.78303223 * 2560; Err = 0.52226562 * 2560; time = 0.0137s; samplesPerSecond = 187422.2
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1611-1620, 36.90%]: CE.SM = 1.78288574 * 2560; Err = 0.51796875 * 2560; time = 0.0137s; samplesPerSecond = 186222.4
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1621-1630, 37.13%]: CE.SM = 1.82502441 * 2560; Err = 0.51445312 * 2560; time = 0.0134s; samplesPerSecond = 191444.8
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1631-1640, 37.36%]: CE.SM = 1.88908691 * 2560; Err = 0.54179687 * 2560; time = 0.0137s; samplesPerSecond = 186915.9
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1641-1650, 37.59%]: CE.SM = 1.83730469 * 2560; Err = 0.51328125 * 2560; time = 0.0138s; samplesPerSecond = 185171.8
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1651-1660, 37.81%]: CE.SM = 1.81320801 * 2560; Err = 0.51796875 * 2560; time = 0.0138s; samplesPerSecond = 185011.2
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1661-1670, 38.04%]: CE.SM = 1.80607910 * 2560; Err = 0.51484375 * 2560; time = 0.0136s; samplesPerSecond = 188429.3
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1671-1680, 38.27%]: CE.SM = 1.77026367 * 2560; Err = 0.49531250 * 2560; time = 0.0138s; samplesPerSecond = 185493.8
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1681-1690, 38.50%]: CE.SM = 1.83120117 * 2560; Err = 0.51796875 * 2560; time = 0.0138s; samplesPerSecond = 185440.1
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1691-1700, 38.72%]: CE.SM = 1.84626465 * 2560; Err = 0.54414063 * 2560; time = 0.0135s; samplesPerSecond = 189377.1
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1701-1710, 38.95%]: CE.SM = 1.77441406 * 2560; Err = 0.50039062 * 2560; time = 0.0138s; samplesPerSecond = 185816.9
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1711-1720, 39.18%]: CE.SM = 1.80747070 * 2560; Err = 0.52734375 * 2560; time = 0.0138s; samplesPerSecond = 185185.2
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1721-1730, 39.41%]: CE.SM = 1.83874512 * 2560; Err = 0.51835937 * 2560; time = 0.0138s; samplesPerSecond = 185453.5
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1731-1740, 39.64%]: CE.SM = 1.82116699 * 2560; Err = 0.53085938 * 2560; time = 0.0139s; samplesPerSecond = 184757.5
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1741-1750, 39.86%]: CE.SM = 1.82763672 * 2560; Err = 0.53632813 * 2560; time = 0.0139s; samplesPerSecond = 184438.0
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1751-1760, 40.09%]: CE.SM = 1.85583496 * 2560; Err = 0.52968750 * 2560; time = 0.0136s; samplesPerSecond = 188207.6
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1761-1770, 40.32%]: CE.SM = 1.80878906 * 2560; Err = 0.51796875 * 2560; time = 0.0138s; samplesPerSecond = 184957.7
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1771-1780, 40.55%]: CE.SM = 1.79667969 * 2560; Err = 0.51601562 * 2560; time = 0.0136s; samplesPerSecond = 188470.9
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1781-1790, 40.77%]: CE.SM = 1.84548340 * 2560; Err = 0.51015625 * 2560; time = 0.0137s; samplesPerSecond = 186208.9
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1791-1800, 41.00%]: CE.SM = 1.80117188 * 2560; Err = 0.50859375 * 2560; time = 0.0137s; samplesPerSecond = 186398.7
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1801-1810, 41.23%]: CE.SM = 1.76982422 * 2560; Err = 0.50117188 * 2560; time = 0.0138s; samplesPerSecond = 185722.6
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1811-1820, 41.46%]: CE.SM = 1.79731445 * 2560; Err = 0.51679688 * 2560; time = 0.0138s; samplesPerSecond = 185480.4
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1821-1830, 41.69%]: CE.SM = 1.83469238 * 2560; Err = 0.52148438 * 2560; time = 0.0137s; samplesPerSecond = 187449.7
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1831-1840, 41.91%]: CE.SM = 1.77836914 * 2560; Err = 0.51484375 * 2560; time = 0.0137s; samplesPerSecond = 186290.2
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1841-1850, 42.14%]: CE.SM = 1.76325684 * 2560; Err = 0.50781250 * 2560; time = 0.0136s; samplesPerSecond = 187917.5
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1851-1860, 42.37%]: CE.SM = 1.83273926 * 2560; Err = 0.51914063 * 2560; time = 0.0138s; samplesPerSecond = 186087.1
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1861-1870, 42.60%]: CE.SM = 1.81718750 * 2560; Err = 0.52578125 * 2560; time = 0.0138s; samplesPerSecond = 185709.1
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1871-1880, 42.82%]: CE.SM = 1.76718750 * 2560; Err = 0.51679688 * 2560; time = 0.0138s; samplesPerSecond = 185978.9
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1881-1890, 43.05%]: CE.SM = 1.81447754 * 2560; Err = 0.52656250 * 2560; time = 0.0138s; samplesPerSecond = 186181.8
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1891-1900, 43.28%]: CE.SM = 1.80461426 * 2560; Err = 0.51757812 * 2560; time = 0.0137s; samplesPerSecond = 186439.4
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1901-1910, 43.51%]: CE.SM = 1.74865723 * 2560; Err = 0.51484375 * 2560; time = 0.0138s; samplesPerSecond = 184957.7
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1911-1920, 43.74%]: CE.SM = 1.75905762 * 2560; Err = 0.50000000 * 2560; time = 0.0137s; samplesPerSecond = 186249.5
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1921-1930, 43.96%]: CE.SM = 1.77099609 * 2560; Err = 0.50937500 * 2560; time = 0.0165s; samplesPerSecond = 155292.7
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1931-1940, 44.19%]: CE.SM = 1.83259277 * 2560; Err = 0.53164062 * 2560; time = 0.0138s; samplesPerSecond = 185091.5
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1941-1950, 44.42%]: CE.SM = 1.78125000 * 2560; Err = 0.50390625 * 2560; time = 0.0137s; samplesPerSecond = 186330.9
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1951-1960, 44.65%]: CE.SM = 1.79167480 * 2560; Err = 0.50390625 * 2560; time = 0.0137s; samplesPerSecond = 186861.3
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1961-1970, 44.87%]: CE.SM = 1.83625488 * 2560; Err = 0.51328125 * 2560; time = 0.0137s; samplesPerSecond = 187422.2
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1971-1980, 45.10%]: CE.SM = 1.90090332 * 2560; Err = 0.54335937 * 2560; time = 0.0138s; samplesPerSecond = 185843.9
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1981-1990, 45.33%]: CE.SM = 1.82241211 * 2560; Err = 0.52304688 * 2560; time = 0.0153s; samplesPerSecond = 167681.9
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[1991-2000, 45.56%]: CE.SM = 1.84067383 * 2560; Err = 0.51757812 * 2560; time = 0.0172s; samplesPerSecond = 148759.4
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2001-2010, 45.79%]: CE.SM = 1.80270996 * 2560; Err = 0.52070313 * 2560; time = 0.0152s; samplesPerSecond = 168487.6
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2011-2020, 46.01%]: CE.SM = 1.79523926 * 2560; Err = 0.51953125 * 2560; time = 0.0141s; samplesPerSecond = 181328.8
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2021-2030, 46.24%]: CE.SM = 1.81750488 * 2560; Err = 0.52070313 * 2560; time = 0.0142s; samplesPerSecond = 179687.0
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2031-2040, 46.47%]: CE.SM = 1.79704590 * 2560; Err = 0.50273437 * 2560; time = 0.0141s; samplesPerSecond = 182193.4
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2041-2050, 46.70%]: CE.SM = 1.84082031 * 2560; Err = 0.52382812 * 2560; time = 0.0140s; samplesPerSecond = 182739.7
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2051-2060, 46.92%]: CE.SM = 1.85314941 * 2560; Err = 0.53437500 * 2560; time = 0.0137s; samplesPerSecond = 187011.5
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2061-2070, 47.15%]: CE.SM = 1.83657227 * 2560; Err = 0.51679688 * 2560; time = 0.0141s; samplesPerSecond = 181689.1
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2071-2080, 47.38%]: CE.SM = 1.83085937 * 2560; Err = 0.51835937 * 2560; time = 0.0139s; samplesPerSecond = 184637.6
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2081-2090, 47.61%]: CE.SM = 1.81198730 * 2560; Err = 0.51210937 * 2560; time = 0.0141s; samplesPerSecond = 181226.1
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2091-2100, 47.84%]: CE.SM = 1.78872070 * 2560; Err = 0.52265625 * 2560; time = 0.0134s; samplesPerSecond = 190405.4
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2101-2110, 48.06%]: CE.SM = 1.76713867 * 2560; Err = 0.49453125 * 2560; time = 0.0138s; samplesPerSecond = 185507.2
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2111-2120, 48.29%]: CE.SM = 1.81279297 * 2560; Err = 0.51953125 * 2560; time = 0.0139s; samplesPerSecond = 184770.8
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2121-2130, 48.52%]: CE.SM = 1.76801758 * 2560; Err = 0.50429687 * 2560; time = 0.0136s; samplesPerSecond = 187931.3
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2131-2140, 48.75%]: CE.SM = 1.82717285 * 2560; Err = 0.53164062 * 2560; time = 0.0139s; samplesPerSecond = 183947.7
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2141-2150, 48.97%]: CE.SM = 1.76989746 * 2560; Err = 0.51757812 * 2560; time = 0.0139s; samplesPerSecond = 184411.5
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2151-2160, 49.20%]: CE.SM = 1.82436523 * 2560; Err = 0.52265625 * 2560; time = 0.0139s; samplesPerSecond = 184704.2
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2161-2170, 49.43%]: CE.SM = 1.81520996 * 2560; Err = 0.51210937 * 2560; time = 0.0138s; samplesPerSecond = 186006.0
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2171-2180, 49.66%]: CE.SM = 1.87741699 * 2560; Err = 0.52460938 * 2560; time = 0.0137s; samplesPerSecond = 186806.8
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2181-2190, 49.89%]: CE.SM = 1.80192871 * 2560; Err = 0.52539062 * 2560; time = 0.0136s; samplesPerSecond = 188540.3
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2191-2200, 50.11%]: CE.SM = 1.75341797 * 2560; Err = 0.50351563 * 2560; time = 0.0135s; samplesPerSecond = 189981.4
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2201-2210, 50.34%]: CE.SM = 1.83098145 * 2560; Err = 0.50898438 * 2560; time = 0.0138s; samplesPerSecond = 185265.6
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2211-2220, 50.57%]: CE.SM = 1.82724609 * 2560; Err = 0.52968750 * 2560; time = 0.0139s; samplesPerSecond = 183617.8
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2221-2230, 50.80%]: CE.SM = 1.80654297 * 2560; Err = 0.52109375 * 2560; time = 0.0137s; samplesPerSecond = 186344.4
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2231-2240, 51.03%]: CE.SM = 1.79345703 * 2560; Err = 0.51132813 * 2560; time = 0.0140s; samplesPerSecond = 183420.5
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2241-2250, 51.25%]: CE.SM = 1.81787109 * 2560; Err = 0.53984375 * 2560; time = 0.0137s; samplesPerSecond = 187545.8
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2251-2260, 51.48%]: CE.SM = 1.86425781 * 2560; Err = 0.52539062 * 2560; time = 0.0137s; samplesPerSecond = 186263.1
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2261-2270, 51.71%]: CE.SM = 1.80136719 * 2560; Err = 0.51875000 * 2560; time = 0.0139s; samplesPerSecond = 184345.1
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2271-2280, 51.94%]: CE.SM = 1.81835938 * 2560; Err = 0.52109375 * 2560; time = 0.0138s; samplesPerSecond = 185897.9
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2281-2290, 52.16%]: CE.SM = 1.82402344 * 2560; Err = 0.52656250 * 2560; time = 0.0162s; samplesPerSecond = 157878.5
12/20/2016 15:27:07:  Epoch[ 2 of 2]-Minibatch[2291-2300, 52.39%]: CE.SM = 1.76269531 * 2560; Err = 0.50546875 * 2560; time = 0.0134s; samplesPerSecond = 191287.5
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2301-2310, 52.62%]: CE.SM = 1.80419922 * 2560; Err = 0.51835937 * 2560; time = 0.0136s; samplesPerSecond = 188484.8
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2311-2320, 52.85%]: CE.SM = 1.78476562 * 2560; Err = 0.52226562 * 2560; time = 0.0137s; samplesPerSecond = 186793.1
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2321-2330, 53.08%]: CE.SM = 1.77646484 * 2560; Err = 0.51718750 * 2560; time = 0.0138s; samplesPerSecond = 185104.8
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2331-2340, 53.30%]: CE.SM = 1.80302734 * 2560; Err = 0.51054687 * 2560; time = 0.0139s; samplesPerSecond = 184664.2
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2341-2350, 53.53%]: CE.SM = 1.79965820 * 2560; Err = 0.51875000 * 2560; time = 0.0136s; samplesPerSecond = 188221.5
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2351-2360, 53.76%]: CE.SM = 1.76601563 * 2560; Err = 0.51171875 * 2560; time = 0.0138s; samplesPerSecond = 185171.8
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2361-2370, 53.99%]: CE.SM = 1.76928711 * 2560; Err = 0.50507813 * 2560; time = 0.0137s; samplesPerSecond = 186875.0
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2371-2380, 54.21%]: CE.SM = 1.80229492 * 2560; Err = 0.51679688 * 2560; time = 0.0138s; samplesPerSecond = 185695.6
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2381-2390, 54.44%]: CE.SM = 1.84970703 * 2560; Err = 0.51718750 * 2560; time = 0.0135s; samplesPerSecond = 189981.4
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2391-2400, 54.67%]: CE.SM = 1.81137695 * 2560; Err = 0.50781250 * 2560; time = 0.0137s; samplesPerSecond = 186575.3
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2401-2410, 54.90%]: CE.SM = 1.75463867 * 2560; Err = 0.51328125 * 2560; time = 0.0137s; samplesPerSecond = 187202.9
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2411-2420, 55.13%]: CE.SM = 1.74296875 * 2560; Err = 0.51015625 * 2560; time = 0.0138s; samplesPerSecond = 185843.9
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2421-2430, 55.35%]: CE.SM = 1.82133789 * 2560; Err = 0.53085938 * 2560; time = 0.0138s; samplesPerSecond = 185292.4
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2431-2440, 55.58%]: CE.SM = 1.77197266 * 2560; Err = 0.50078125 * 2560; time = 0.0137s; samplesPerSecond = 187216.6
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2441-2450, 55.81%]: CE.SM = 1.77700195 * 2560; Err = 0.51054687 * 2560; time = 0.0137s; samplesPerSecond = 186765.9
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2451-2460, 56.04%]: CE.SM = 1.81181641 * 2560; Err = 0.50390625 * 2560; time = 0.0138s; samplesPerSecond = 184877.6
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2461-2470, 56.26%]: CE.SM = 1.78803711 * 2560; Err = 0.50234375 * 2560; time = 0.0137s; samplesPerSecond = 186317.3
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2471-2480, 56.49%]: CE.SM = 1.79492188 * 2560; Err = 0.51171875 * 2560; time = 0.0139s; samplesPerSecond = 183617.8
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2481-2490, 56.72%]: CE.SM = 1.79804688 * 2560; Err = 0.51093750 * 2560; time = 0.0136s; samplesPerSecond = 188179.9
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2491-2500, 56.95%]: CE.SM = 1.80859375 * 2560; Err = 0.51796875 * 2560; time = 0.0139s; samplesPerSecond = 184517.8
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2501-2510, 57.18%]: CE.SM = 1.77778320 * 2560; Err = 0.51289063 * 2560; time = 0.0138s; samplesPerSecond = 185790.0
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2511-2520, 57.40%]: CE.SM = 1.82363281 * 2560; Err = 0.52304688 * 2560; time = 0.0141s; samplesPerSecond = 181934.5
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2521-2530, 57.63%]: CE.SM = 1.80781250 * 2560; Err = 0.51328125 * 2560; time = 0.0137s; samplesPerSecond = 187093.5
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2531-2540, 57.86%]: CE.SM = 1.80737305 * 2560; Err = 0.52187500 * 2560; time = 0.0138s; samplesPerSecond = 186087.1
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2541-2550, 58.09%]: CE.SM = 1.79594727 * 2560; Err = 0.51132813 * 2560; time = 0.0139s; samplesPerSecond = 184597.6
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2551-2560, 58.31%]: CE.SM = 1.78164063 * 2560; Err = 0.50664062 * 2560; time = 0.0138s; samplesPerSecond = 185064.7
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2561-2570, 58.54%]: CE.SM = 1.81914062 * 2560; Err = 0.51992187 * 2560; time = 0.0136s; samplesPerSecond = 188179.9
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2571-2580, 58.77%]: CE.SM = 1.82685547 * 2560; Err = 0.52617187 * 2560; time = 0.0138s; samplesPerSecond = 186046.5
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2581-2590, 59.00%]: CE.SM = 1.77846680 * 2560; Err = 0.51054687 * 2560; time = 0.0139s; samplesPerSecond = 184597.6
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2591-2600, 59.23%]: CE.SM = 1.76464844 * 2560; Err = 0.50390625 * 2560; time = 0.0138s; samplesPerSecond = 186154.7
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2601-2610, 59.45%]: CE.SM = 1.81499023 * 2560; Err = 0.52148438 * 2560; time = 0.0138s; samplesPerSecond = 185171.8
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2611-2620, 59.68%]: CE.SM = 1.79052734 * 2560; Err = 0.51562500 * 2560; time = 0.0148s; samplesPerSecond = 173324.3
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2621-2630, 59.91%]: CE.SM = 1.80439453 * 2560; Err = 0.51328125 * 2560; time = 0.0149s; samplesPerSecond = 171558.8
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2631-2640, 60.14%]: CE.SM = 1.83305664 * 2560; Err = 0.52109375 * 2560; time = 0.0160s; samplesPerSecond = 160060.0
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2641-2650, 60.36%]: CE.SM = 1.83461914 * 2560; Err = 0.53476563 * 2560; time = 0.0141s; samplesPerSecond = 181316.0
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2651-2660, 60.59%]: CE.SM = 1.81669922 * 2560; Err = 0.53007812 * 2560; time = 0.0143s; samplesPerSecond = 178608.8
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2661-2670, 60.82%]: CE.SM = 1.80463867 * 2560; Err = 0.50117188 * 2560; time = 0.0141s; samplesPerSecond = 181341.6
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2671-2680, 61.05%]: CE.SM = 1.81782227 * 2560; Err = 0.52382812 * 2560; time = 0.0138s; samplesPerSecond = 185938.4
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2681-2690, 61.28%]: CE.SM = 1.78100586 * 2560; Err = 0.51250000 * 2560; time = 0.0143s; samplesPerSecond = 179623.9
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2691-2700, 61.50%]: CE.SM = 1.78720703 * 2560; Err = 0.51835937 * 2560; time = 0.0139s; samplesPerSecond = 184292.0
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2701-2710, 61.73%]: CE.SM = 1.83315430 * 2560; Err = 0.52460938 * 2560; time = 0.0141s; samplesPerSecond = 181908.6
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2711-2720, 61.96%]: CE.SM = 1.81992187 * 2560; Err = 0.52421875 * 2560; time = 0.0139s; samplesPerSecond = 184106.4
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2721-2730, 62.19%]: CE.SM = 1.77675781 * 2560; Err = 0.50781250 * 2560; time = 0.0137s; samplesPerSecond = 187257.7
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2731-2740, 62.41%]: CE.SM = 1.73662109 * 2560; Err = 0.49687500 * 2560; time = 0.0139s; samplesPerSecond = 184531.1
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2741-2750, 62.64%]: CE.SM = 1.79941406 * 2560; Err = 0.52226562 * 2560; time = 0.0138s; samplesPerSecond = 184957.7
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2751-2760, 62.87%]: CE.SM = 1.81865234 * 2560; Err = 0.51953125 * 2560; time = 0.0138s; samplesPerSecond = 185790.0
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2761-2770, 63.10%]: CE.SM = 1.80224609 * 2560; Err = 0.52148438 * 2560; time = 0.0138s; samplesPerSecond = 186073.6
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2771-2780, 63.33%]: CE.SM = 1.85053711 * 2560; Err = 0.54062500 * 2560; time = 0.0140s; samplesPerSecond = 182974.8
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2781-2790, 63.55%]: CE.SM = 1.80678711 * 2560; Err = 0.51015625 * 2560; time = 0.0138s; samplesPerSecond = 185884.4
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2791-2800, 63.78%]: CE.SM = 1.83393555 * 2560; Err = 0.52500000 * 2560; time = 0.0137s; samplesPerSecond = 187257.7
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2801-2810, 64.01%]: CE.SM = 1.82866211 * 2560; Err = 0.52695313 * 2560; time = 0.0140s; samplesPerSecond = 183184.3
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2811-2820, 64.24%]: CE.SM = 1.80908203 * 2560; Err = 0.51953125 * 2560; time = 0.0136s; samplesPerSecond = 188874.1
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2821-2830, 64.46%]: CE.SM = 1.78676758 * 2560; Err = 0.51718750 * 2560; time = 0.0139s; samplesPerSecond = 184531.1
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2831-2840, 64.69%]: CE.SM = 1.83452148 * 2560; Err = 0.51679688 * 2560; time = 0.0140s; samplesPerSecond = 183236.7
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2841-2850, 64.92%]: CE.SM = 1.80112305 * 2560; Err = 0.51757812 * 2560; time = 0.0138s; samplesPerSecond = 186181.8
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2851-2860, 65.15%]: CE.SM = 1.78598633 * 2560; Err = 0.50781250 * 2560; time = 0.0137s; samplesPerSecond = 186956.8
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2861-2870, 65.38%]: CE.SM = 1.79272461 * 2560; Err = 0.50625000 * 2560; time = 0.0139s; samplesPerSecond = 183974.1
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2871-2880, 65.60%]: CE.SM = 1.77548828 * 2560; Err = 0.51171875 * 2560; time = 0.0140s; samplesPerSecond = 182219.4
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2881-2890, 65.83%]: CE.SM = 1.77128906 * 2560; Err = 0.50976562 * 2560; time = 0.0138s; samplesPerSecond = 186006.0
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2891-2900, 66.06%]: CE.SM = 1.77797852 * 2560; Err = 0.49960938 * 2560; time = 0.0136s; samplesPerSecond = 188526.4
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2901-2910, 66.29%]: CE.SM = 1.77050781 * 2560; Err = 0.50468750 * 2560; time = 0.0136s; samplesPerSecond = 188693.2
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2911-2920, 66.51%]: CE.SM = 1.73720703 * 2560; Err = 0.50859375 * 2560; time = 0.0135s; samplesPerSecond = 190080.2
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2921-2930, 66.74%]: CE.SM = 1.76489258 * 2560; Err = 0.50117188 * 2560; time = 0.0134s; samplesPerSecond = 190462.0
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2931-2940, 66.97%]: CE.SM = 1.74692383 * 2560; Err = 0.49570313 * 2560; time = 0.0137s; samplesPerSecond = 186588.9
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2941-2950, 67.20%]: CE.SM = 1.75654297 * 2560; Err = 0.50664062 * 2560; time = 0.0135s; samplesPerSecond = 189826.5
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2951-2960, 67.43%]: CE.SM = 1.81645508 * 2560; Err = 0.51367188 * 2560; time = 0.0145s; samplesPerSecond = 177150.4
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2961-2970, 67.65%]: CE.SM = 1.78159180 * 2560; Err = 0.52421875 * 2560; time = 0.0137s; samplesPerSecond = 186779.5
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2971-2980, 67.88%]: CE.SM = 1.75991211 * 2560; Err = 0.50781250 * 2560; time = 0.0135s; samplesPerSecond = 189195.2
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2981-2990, 68.11%]: CE.SM = 1.75834961 * 2560; Err = 0.50234375 * 2560; time = 0.0136s; samplesPerSecond = 188235.3
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[2991-3000, 68.34%]: CE.SM = 1.80170898 * 2560; Err = 0.49843750 * 2560; time = 0.0135s; samplesPerSecond = 190278.0
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[3001-3010, 68.56%]: CE.SM = 1.77182617 * 2560; Err = 0.50742188 * 2560; time = 0.0134s; samplesPerSecond = 190377.0
12/20/2016 15:27:08:  Epoch[ 2 of 2]-Minibatch[3011-3020, 68.79%]: CE.SM = 1.78193359 * 2560; Err = 0.51015625 * 2560; time = 0.0134s; samplesPerSecond = 191330.3
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3021-3030, 69.02%]: CE.SM = 1.81240234 * 2560; Err = 0.51015625 * 2560; time = 0.0135s; samplesPerSecond = 188957.8
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3031-3040, 69.25%]: CE.SM = 1.78305664 * 2560; Err = 0.52226562 * 2560; time = 0.0135s; samplesPerSecond = 189321.1
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3041-3050, 69.48%]: CE.SM = 1.80869141 * 2560; Err = 0.51992187 * 2560; time = 0.0135s; samplesPerSecond = 189125.3
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3051-3060, 69.70%]: CE.SM = 1.80141602 * 2560; Err = 0.51640625 * 2560; time = 0.0137s; samplesPerSecond = 186575.3
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3061-3070, 69.93%]: CE.SM = 1.75258789 * 2560; Err = 0.50039062 * 2560; time = 0.0135s; samplesPerSecond = 189601.5
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3071-3080, 70.16%]: CE.SM = 1.79697266 * 2560; Err = 0.51367188 * 2560; time = 0.0132s; samplesPerSecond = 193382.7
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3081-3090, 70.39%]: CE.SM = 1.77890625 * 2560; Err = 0.50234375 * 2560; time = 0.0134s; samplesPerSecond = 191216.0
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3091-3100, 70.62%]: CE.SM = 1.74965820 * 2560; Err = 0.50195312 * 2560; time = 0.0136s; samplesPerSecond = 188679.2
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3101-3110, 70.84%]: CE.SM = 1.79072266 * 2560; Err = 0.51054687 * 2560; time = 0.0136s; samplesPerSecond = 188207.6
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3111-3120, 71.07%]: CE.SM = 1.84160156 * 2560; Err = 0.53046875 * 2560; time = 0.0136s; samplesPerSecond = 188179.9
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3121-3130, 71.30%]: CE.SM = 1.79755859 * 2560; Err = 0.51679688 * 2560; time = 0.0137s; samplesPerSecond = 187490.8
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3131-3140, 71.53%]: CE.SM = 1.77983398 * 2560; Err = 0.51718750 * 2560; time = 0.0135s; samplesPerSecond = 190150.8
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3141-3150, 71.75%]: CE.SM = 1.77705078 * 2560; Err = 0.50429687 * 2560; time = 0.0134s; samplesPerSecond = 191144.6
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3151-3160, 71.98%]: CE.SM = 1.79316406 * 2560; Err = 0.51015625 * 2560; time = 0.0137s; samplesPerSecond = 187504.6
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3161-3170, 72.21%]: CE.SM = 1.78403320 * 2560; Err = 0.50625000 * 2560; time = 0.0136s; samplesPerSecond = 187931.3
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3171-3180, 72.44%]: CE.SM = 1.86621094 * 2560; Err = 0.52070313 * 2560; time = 0.0136s; samplesPerSecond = 188832.3
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3181-3190, 72.67%]: CE.SM = 1.81772461 * 2560; Err = 0.51484375 * 2560; time = 0.0134s; samplesPerSecond = 191173.2
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3191-3200, 72.89%]: CE.SM = 1.81977539 * 2560; Err = 0.52460938 * 2560; time = 0.0135s; samplesPerSecond = 190037.9
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3201-3210, 73.12%]: CE.SM = 1.75942383 * 2560; Err = 0.51210937 * 2560; time = 0.0137s; samplesPerSecond = 187134.5
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3211-3220, 73.35%]: CE.SM = 1.76064453 * 2560; Err = 0.51484375 * 2560; time = 0.0136s; samplesPerSecond = 188249.1
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3221-3230, 73.58%]: CE.SM = 1.78642578 * 2560; Err = 0.52226562 * 2560; time = 0.0136s; samplesPerSecond = 188097.0
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3231-3240, 73.80%]: CE.SM = 1.77788086 * 2560; Err = 0.50429687 * 2560; time = 0.0135s; samplesPerSecond = 190320.4
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3241-3250, 74.03%]: CE.SM = 1.75961914 * 2560; Err = 0.50195312 * 2560; time = 0.0135s; samplesPerSecond = 190179.0
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3251-3260, 74.26%]: CE.SM = 1.74335938 * 2560; Err = 0.51562500 * 2560; time = 0.0136s; samplesPerSecond = 188484.8
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3261-3270, 74.49%]: CE.SM = 1.76113281 * 2560; Err = 0.51132813 * 2560; time = 0.0136s; samplesPerSecond = 188498.6
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3271-3280, 74.72%]: CE.SM = 1.74448242 * 2560; Err = 0.49921875 * 2560; time = 0.0138s; samplesPerSecond = 186060.0
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3281-3290, 74.94%]: CE.SM = 1.80136719 * 2560; Err = 0.51953125 * 2560; time = 0.0135s; samplesPerSecond = 189657.7
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3291-3300, 75.17%]: CE.SM = 1.76225586 * 2560; Err = 0.51796875 * 2560; time = 0.0135s; samplesPerSecond = 189251.1
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3301-3310, 75.40%]: CE.SM = 1.78129883 * 2560; Err = 0.51093750 * 2560; time = 0.0137s; samplesPerSecond = 186888.6
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3311-3320, 75.63%]: CE.SM = 1.80507813 * 2560; Err = 0.51171875 * 2560; time = 0.0136s; samplesPerSecond = 187614.5
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3321-3330, 75.85%]: CE.SM = 1.79355469 * 2560; Err = 0.52070313 * 2560; time = 0.0136s; samplesPerSecond = 188263.0
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3331-3340, 76.08%]: CE.SM = 1.77890625 * 2560; Err = 0.50664062 * 2560; time = 0.0156s; samplesPerSecond = 164609.1
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3341-3350, 76.31%]: CE.SM = 1.81684570 * 2560; Err = 0.51679688 * 2560; time = 0.0137s; samplesPerSecond = 187025.1
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3351-3360, 76.54%]: CE.SM = 1.78208008 * 2560; Err = 0.51562500 * 2560; time = 0.0136s; samplesPerSecond = 188762.7
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3361-3370, 76.77%]: CE.SM = 1.76938477 * 2560; Err = 0.50859375 * 2560; time = 0.0137s; samplesPerSecond = 187120.8
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3371-3380, 76.99%]: CE.SM = 1.80146484 * 2560; Err = 0.50312500 * 2560; time = 0.0136s; samplesPerSecond = 187600.8
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3381-3390, 77.22%]: CE.SM = 1.75029297 * 2560; Err = 0.50976562 * 2560; time = 0.0138s; samplesPerSecond = 185749.5
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3391-3400, 77.45%]: CE.SM = 1.76694336 * 2560; Err = 0.50546875 * 2560; time = 0.0138s; samplesPerSecond = 184877.6
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3401-3410, 77.68%]: CE.SM = 1.79418945 * 2560; Err = 0.50234375 * 2560; time = 0.0136s; samplesPerSecond = 187862.3
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3411-3420, 77.90%]: CE.SM = 1.79321289 * 2560; Err = 0.50703125 * 2560; time = 0.0138s; samplesPerSecond = 184997.8
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3421-3430, 78.13%]: CE.SM = 1.79912109 * 2560; Err = 0.52460938 * 2560; time = 0.0138s; samplesPerSecond = 185561.0
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3431-3440, 78.36%]: CE.SM = 1.78559570 * 2560; Err = 0.50273437 * 2560; time = 0.0137s; samplesPerSecond = 187257.7
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3441-3450, 78.59%]: CE.SM = 1.78613281 * 2560; Err = 0.50546875 * 2560; time = 0.0138s; samplesPerSecond = 186087.1
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3451-3460, 78.82%]: CE.SM = 1.71635742 * 2560; Err = 0.49296875 * 2560; time = 0.0136s; samplesPerSecond = 188707.1
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3461-3470, 79.04%]: CE.SM = 1.73120117 * 2560; Err = 0.49179688 * 2560; time = 0.0138s; samplesPerSecond = 185790.0
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3471-3480, 79.27%]: CE.SM = 1.79658203 * 2560; Err = 0.51562500 * 2560; time = 0.0138s; samplesPerSecond = 185238.8
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3481-3490, 79.50%]: CE.SM = 1.80585938 * 2560; Err = 0.50507813 * 2560; time = 0.0140s; samplesPerSecond = 183066.4
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3491-3500, 79.73%]: CE.SM = 1.77304687 * 2560; Err = 0.51093750 * 2560; time = 0.0138s; samplesPerSecond = 185399.8
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3501-3510, 79.95%]: CE.SM = 1.75615234 * 2560; Err = 0.51250000 * 2560; time = 0.0138s; samplesPerSecond = 185426.6
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3511-3520, 80.18%]: CE.SM = 1.77211914 * 2560; Err = 0.50820312 * 2560; time = 0.0140s; samplesPerSecond = 183000.9
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3521-3530, 80.41%]: CE.SM = 1.81816406 * 2560; Err = 0.51523438 * 2560; time = 0.0139s; samplesPerSecond = 184650.9
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3531-3540, 80.64%]: CE.SM = 1.79106445 * 2560; Err = 0.50820312 * 2560; time = 0.0137s; samplesPerSecond = 186779.5
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3541-3550, 80.87%]: CE.SM = 1.81113281 * 2560; Err = 0.52578125 * 2560; time = 0.0139s; samplesPerSecond = 183736.5
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3551-3560, 81.09%]: CE.SM = 1.77749023 * 2560; Err = 0.50156250 * 2560; time = 0.0134s; samplesPerSecond = 190873.8
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3561-3570, 81.32%]: CE.SM = 1.78422852 * 2560; Err = 0.51406250 * 2560; time = 0.0188s; samplesPerSecond = 136278.9
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3571-3580, 81.55%]: CE.SM = 1.76440430 * 2560; Err = 0.51132813 * 2560; time = 0.0134s; samplesPerSecond = 190959.3
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3581-3590, 81.78%]: CE.SM = 1.76381836 * 2560; Err = 0.50976562 * 2560; time = 0.0141s; samplesPerSecond = 181431.6
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3591-3600, 82.00%]: CE.SM = 1.78237305 * 2560; Err = 0.51953125 * 2560; time = 0.0135s; samplesPerSecond = 189111.3
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3601-3610, 82.23%]: CE.SM = 1.82558594 * 2560; Err = 0.51601562 * 2560; time = 0.0133s; samplesPerSecond = 192134.5
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3611-3620, 82.46%]: CE.SM = 1.70458984 * 2560; Err = 0.48906250 * 2560; time = 0.0133s; samplesPerSecond = 192800.1
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3621-3630, 82.69%]: CE.SM = 1.77993164 * 2560; Err = 0.49414062 * 2560; time = 0.0138s; samplesPerSecond = 185763.0
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3631-3640, 82.92%]: CE.SM = 1.84184570 * 2560; Err = 0.52617187 * 2560; time = 0.0138s; samplesPerSecond = 186127.7
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3641-3650, 83.14%]: CE.SM = 1.78476562 * 2560; Err = 0.50195312 * 2560; time = 0.0135s; samplesPerSecond = 189167.2
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3651-3660, 83.37%]: CE.SM = 1.77846680 * 2560; Err = 0.50390625 * 2560; time = 0.0136s; samplesPerSecond = 188470.9
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3661-3670, 83.60%]: CE.SM = 1.71289062 * 2560; Err = 0.50156250 * 2560; time = 0.0134s; samplesPerSecond = 190532.9
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3671-3680, 83.83%]: CE.SM = 1.74873047 * 2560; Err = 0.51132813 * 2560; time = 0.0135s; samplesPerSecond = 188971.7
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3681-3690, 84.05%]: CE.SM = 1.68398438 * 2560; Err = 0.48632812 * 2560; time = 0.0138s; samplesPerSecond = 184850.9
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3691-3700, 84.28%]: CE.SM = 1.76884766 * 2560; Err = 0.50781250 * 2560; time = 0.0136s; samplesPerSecond = 187752.1
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3701-3710, 84.51%]: CE.SM = 1.71567383 * 2560; Err = 0.50585938 * 2560; time = 0.0136s; samplesPerSecond = 188359.9
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3711-3720, 84.74%]: CE.SM = 1.72695313 * 2560; Err = 0.50468750 * 2560; time = 0.0135s; samplesPerSecond = 189854.6
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3721-3730, 84.97%]: CE.SM = 1.79985352 * 2560; Err = 0.51679688 * 2560; time = 0.0135s; samplesPerSecond = 190037.9
12/20/2016 15:27:09:  Epoch[ 2 of 2]-Minibatch[3731-3740, 85.19%]: CE.SM = 1.79199219 * 2560; Err = 0.52304688 * 2560; time = 0.0134s; samplesPerSecond = 190774.3
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[3741-3750, 85.42%]: CE.SM = 1.73603516 * 2560; Err = 0.49960938 * 2560; time = 0.0139s; samplesPerSecond = 184517.8
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[3751-3760, 85.65%]: CE.SM = 1.74492187 * 2560; Err = 0.49062500 * 2560; time = 0.0139s; samplesPerSecond = 184637.6
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[3761-3770, 85.88%]: CE.SM = 1.75126953 * 2560; Err = 0.51132813 * 2560; time = 0.0138s; samplesPerSecond = 186073.6
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[3771-3780, 86.10%]: CE.SM = 1.78061523 * 2560; Err = 0.51054687 * 2560; time = 0.0139s; samplesPerSecond = 184027.0
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[3781-3790, 86.33%]: CE.SM = 1.83051758 * 2560; Err = 0.52070313 * 2560; time = 0.0137s; samplesPerSecond = 186765.9
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[3791-3800, 86.56%]: CE.SM = 1.76044922 * 2560; Err = 0.50273437 * 2560; time = 0.0136s; samplesPerSecond = 188207.6
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[3801-3810, 86.79%]: CE.SM = 1.78642578 * 2560; Err = 0.52031250 * 2560; time = 0.0139s; samplesPerSecond = 184238.9
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[3811-3820, 87.02%]: CE.SM = 1.73989258 * 2560; Err = 0.50312500 * 2560; time = 0.0139s; samplesPerSecond = 184650.9
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[3821-3830, 87.24%]: CE.SM = 1.77290039 * 2560; Err = 0.51562500 * 2560; time = 0.0138s; samplesPerSecond = 185225.4
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[3831-3840, 87.47%]: CE.SM = 1.75117188 * 2560; Err = 0.50078125 * 2560; time = 0.0139s; samplesPerSecond = 183552.0
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[3841-3850, 87.70%]: CE.SM = 1.78813477 * 2560; Err = 0.50429687 * 2560; time = 0.0137s; samplesPerSecond = 186656.9
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[3851-3860, 87.93%]: CE.SM = 1.77622070 * 2560; Err = 0.51015625 * 2560; time = 0.0140s; samplesPerSecond = 183433.6
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[3861-3870, 88.15%]: CE.SM = 1.79946289 * 2560; Err = 0.52382812 * 2560; time = 0.0135s; samplesPerSecond = 189279.1
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[3871-3880, 88.38%]: CE.SM = 1.80185547 * 2560; Err = 0.50781250 * 2560; time = 0.0136s; samplesPerSecond = 187931.3
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[3881-3890, 88.61%]: CE.SM = 1.72548828 * 2560; Err = 0.48984375 * 2560; time = 0.0138s; samplesPerSecond = 185212.0
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[3891-3900, 88.84%]: CE.SM = 1.68803711 * 2560; Err = 0.47148438 * 2560; time = 0.0138s; samplesPerSecond = 184957.7
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[3901-3910, 89.07%]: CE.SM = 1.73535156 * 2560; Err = 0.50546875 * 2560; time = 0.0139s; samplesPerSecond = 184132.9
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[3911-3920, 89.29%]: CE.SM = 1.78852539 * 2560; Err = 0.50312500 * 2560; time = 0.0140s; samplesPerSecond = 182700.5
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[3921-3930, 89.52%]: CE.SM = 1.75307617 * 2560; Err = 0.49882813 * 2560; time = 0.0138s; samplesPerSecond = 186154.7
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[3931-3940, 89.75%]: CE.SM = 1.76098633 * 2560; Err = 0.50507813 * 2560; time = 0.0146s; samplesPerSecond = 175751.8
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[3941-3950, 89.98%]: CE.SM = 1.79526367 * 2560; Err = 0.51601562 * 2560; time = 0.0229s; samplesPerSecond = 111697.7
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[3951-3960, 90.21%]: CE.SM = 1.80815430 * 2560; Err = 0.51835937 * 2560; time = 0.0153s; samplesPerSecond = 167451.6
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[3961-3970, 90.43%]: CE.SM = 1.74394531 * 2560; Err = 0.49648437 * 2560; time = 0.0147s; samplesPerSecond = 173972.1
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[3971-3980, 90.66%]: CE.SM = 1.78457031 * 2560; Err = 0.50156250 * 2560; time = 0.0142s; samplesPerSecond = 180383.3
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[3981-3990, 90.89%]: CE.SM = 1.79887695 * 2560; Err = 0.50937500 * 2560; time = 0.0145s; samplesPerSecond = 176515.2
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[3991-4000, 91.12%]: CE.SM = 1.68554688 * 2560; Err = 0.48398438 * 2560; time = 0.0138s; samplesPerSecond = 185668.7
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4001-4010, 91.34%]: CE.SM = 1.74853516 * 2560; Err = 0.49492188 * 2560; time = 0.0140s; samplesPerSecond = 182844.1
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4011-4020, 91.57%]: CE.SM = 1.74360352 * 2560; Err = 0.49843750 * 2560; time = 0.0141s; samplesPerSecond = 182141.6
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4021-4030, 91.80%]: CE.SM = 1.72900391 * 2560; Err = 0.49609375 * 2560; time = 0.0141s; samplesPerSecond = 182038.0
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4031-4040, 92.03%]: CE.SM = 1.80351562 * 2560; Err = 0.52968750 * 2560; time = 0.0141s; samplesPerSecond = 181882.8
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4041-4050, 92.26%]: CE.SM = 1.75258789 * 2560; Err = 0.48632812 * 2560; time = 0.0158s; samplesPerSecond = 161861.4
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4051-4060, 92.48%]: CE.SM = 1.76665039 * 2560; Err = 0.49296875 * 2560; time = 0.0138s; samplesPerSecond = 185776.5
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4061-4070, 92.71%]: CE.SM = 1.74140625 * 2560; Err = 0.49960938 * 2560; time = 0.0139s; samplesPerSecond = 183631.0
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4071-4080, 92.94%]: CE.SM = 1.78110352 * 2560; Err = 0.51757812 * 2560; time = 0.0137s; samplesPerSecond = 186616.1
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4081-4090, 93.17%]: CE.SM = 1.75410156 * 2560; Err = 0.50195312 * 2560; time = 0.0142s; samplesPerSecond = 180434.2
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4091-4100, 93.39%]: CE.SM = 1.77402344 * 2560; Err = 0.51250000 * 2560; time = 0.0139s; samplesPerSecond = 184411.5
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4101-4110, 93.62%]: CE.SM = 1.73330078 * 2560; Err = 0.49882813 * 2560; time = 0.0138s; samplesPerSecond = 185655.2
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4111-4120, 93.85%]: CE.SM = 1.74477539 * 2560; Err = 0.49843750 * 2560; time = 0.0138s; samplesPerSecond = 185466.9
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4121-4130, 94.08%]: CE.SM = 1.77509766 * 2560; Err = 0.51914063 * 2560; time = 0.0140s; samplesPerSecond = 182752.7
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4131-4140, 94.31%]: CE.SM = 1.74487305 * 2560; Err = 0.51835937 * 2560; time = 0.0141s; samplesPerSecond = 181586.0
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4141-4150, 94.53%]: CE.SM = 1.76674805 * 2560; Err = 0.49804688 * 2560; time = 0.0140s; samplesPerSecond = 183499.4
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4151-4160, 94.76%]: CE.SM = 1.77138672 * 2560; Err = 0.51289063 * 2560; time = 0.0138s; samplesPerSecond = 185292.4
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4161-4170, 94.99%]: CE.SM = 1.71616211 * 2560; Err = 0.49492188 * 2560; time = 0.0138s; samplesPerSecond = 186060.0
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4171-4180, 95.22%]: CE.SM = 1.73305664 * 2560; Err = 0.50234375 * 2560; time = 0.0139s; samplesPerSecond = 184690.9
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4181-4190, 95.44%]: CE.SM = 1.75844727 * 2560; Err = 0.50468750 * 2560; time = 0.0139s; samplesPerSecond = 184185.9
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4191-4200, 95.67%]: CE.SM = 1.80429687 * 2560; Err = 0.50742188 * 2560; time = 0.0138s; samplesPerSecond = 185722.6
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4201-4210, 95.90%]: CE.SM = 1.72094727 * 2560; Err = 0.50703125 * 2560; time = 0.0139s; samplesPerSecond = 183538.9
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4211-4220, 96.13%]: CE.SM = 1.81860352 * 2560; Err = 0.51757812 * 2560; time = 0.0139s; samplesPerSecond = 184584.3
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4221-4230, 96.36%]: CE.SM = 1.73212891 * 2560; Err = 0.50351563 * 2560; time = 0.0138s; samplesPerSecond = 185426.6
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4231-4240, 96.58%]: CE.SM = 1.78437500 * 2560; Err = 0.50546875 * 2560; time = 0.0139s; samplesPerSecond = 184066.7
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4241-4250, 96.81%]: CE.SM = 1.77490234 * 2560; Err = 0.51250000 * 2560; time = 0.0138s; samplesPerSecond = 185198.6
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4251-4260, 97.04%]: CE.SM = 1.74863281 * 2560; Err = 0.50351563 * 2560; time = 0.0134s; samplesPerSecond = 191258.9
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4261-4270, 97.27%]: CE.SM = 1.74746094 * 2560; Err = 0.49921875 * 2560; time = 0.0136s; samplesPerSecond = 187931.3
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4271-4280, 97.49%]: CE.SM = 1.75229492 * 2560; Err = 0.50898438 * 2560; time = 0.0154s; samplesPerSecond = 166169.0
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4281-4290, 97.72%]: CE.SM = 1.76733398 * 2560; Err = 0.50546875 * 2560; time = 0.0135s; samplesPerSecond = 189953.3
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4291-4300, 97.95%]: CE.SM = 1.77675781 * 2560; Err = 0.50625000 * 2560; time = 0.0136s; samplesPerSecond = 188512.5
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4301-4310, 98.18%]: CE.SM = 1.80390625 * 2560; Err = 0.51640625 * 2560; time = 0.0138s; samplesPerSecond = 186168.3
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4311-4320, 98.41%]: CE.SM = 1.78803711 * 2560; Err = 0.51601562 * 2560; time = 0.0137s; samplesPerSecond = 187134.5
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4321-4330, 98.63%]: CE.SM = 1.78017578 * 2560; Err = 0.51210937 * 2560; time = 0.0136s; samplesPerSecond = 188124.6
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4331-4340, 98.86%]: CE.SM = 1.73378906 * 2560; Err = 0.48398438 * 2560; time = 0.0138s; samplesPerSecond = 185965.4
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4341-4350, 99.09%]: CE.SM = 1.80576172 * 2560; Err = 0.51484375 * 2560; time = 0.0138s; samplesPerSecond = 185145.0
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4351-4360, 99.32%]: CE.SM = 1.82202148 * 2560; Err = 0.51718750 * 2560; time = 0.0138s; samplesPerSecond = 185682.2
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4361-4370, 99.54%]: CE.SM = 1.76108398 * 2560; Err = 0.50078125 * 2560; time = 0.0140s; samplesPerSecond = 183262.9
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4371-4380, 99.77%]: CE.SM = 1.69843750 * 2560; Err = 0.48789063 * 2560; time = 0.0139s; samplesPerSecond = 183538.9
12/20/2016 15:27:10:  Epoch[ 2 of 2]-Minibatch[4381-4390, 100.00%]: CE.SM = 1.72407227 * 2560; Err = 0.49375000 * 2560; time = 0.0140s; samplesPerSecond = 182974.8
12/20/2016 15:27:10: Finished Epoch[ 2 of 2]: [Training] CE.SM = 1.80690173 * 1124823; Err = 0.51638258 * 1124823; totalSamplesSeen = 2249646; learningRatePerSample = 0.00039062501; epochTime=6.33634s
12/20/2016 15:27:10: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/dptmodel2/cntkSpeech.dnn'

12/20/2016 15:27:11: Action "train" complete.


12/20/2016 15:27:11: ##############################################################################
12/20/2016 15:27:11: #                                                                            #
12/20/2016 15:27:11: # TIMIT_AddLayer3 command (edit action)                                      #
12/20/2016 15:27:11: #                                                                            #
12/20/2016 15:27:11: ##############################################################################


12/20/2016 15:27:11: Action "edit" complete.


12/20/2016 15:27:11: ##############################################################################
12/20/2016 15:27:11: #                                                                            #
12/20/2016 15:27:11: # TIMIT_Train3 command (train action)                                        #
12/20/2016 15:27:11: #                                                                            #
12/20/2016 15:27:11: ##############################################################################

12/20/2016 15:27:11: 
Starting from checkpoint. Loading network from '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn.0'.
NDLBuilder Using GPU 0
Reading script file /home/philly/data/CNTKTestData/Speech/ASR/TIMIT.train.scp.fbank.fullpath.rnn ... 3696 entries
HTKDataDeserializer::HTKDataDeserializer: selected 3696 utterances grouped into 13 chunks, average chunk size: 284.3 utterances, 86524.8 frames (for I/O: 284.3 utterances, 86524.8 frames)
HTKDataDeserializer::HTKDataDeserializer: determined feature kind as 72-dimensional 'FBANK_D_A_Z' with frame shift 10.0 ms
total 183 state names in state list /home/philly/data/CNTKTestData/Speech/ASR/TIMIT.statelist
htkmlfreader: reading MLF file /home/philly/data/CNTKTestData/Speech/ASR/TIMIT.train.align_cistate.mlf.cntk ... total 3696 entries
MLFDataDeserializer::MLFDataDeserializer: 3696 utterances with 1124823 frames in 183 classes
12/20/2016 15:27:11: 
Model has 29 nodes. Using GPU 0.

12/20/2016 15:27:11: Training criterion:   CE.SM = CrossEntropyWithSoftmax
12/20/2016 15:27:11: Evaluation criterion: Err = ClassificationError

12/20/2016 15:27:11: Training 1025207 parameters in 8 out of 8 parameter tensors and 20 nodes with gradient:

12/20/2016 15:27:11: 	Node 'CE.BFF.B' (LearnableParameter operation) : [183]
12/20/2016 15:27:11: 	Node 'CE.BFF.W' (LearnableParameter operation) : [183 x 512]
12/20/2016 15:27:11: 	Node 'L1.BFF.B' (LearnableParameter operation) : [512]
12/20/2016 15:27:11: 	Node 'L1.BFF.W' (LearnableParameter operation) : [512 x 792]
12/20/2016 15:27:11: 	Node 'L2.BFF.B' (LearnableParameter operation) : [512]
12/20/2016 15:27:11: 	Node 'L2.BFF.W' (LearnableParameter operation) : [512 x 512]
12/20/2016 15:27:11: 	Node 'L3.BFF.B' (LearnableParameter operation) : [512]
12/20/2016 15:27:11: 	Node 'L3.BFF.W' (LearnableParameter operation) : [512 x 512]

12/20/2016 15:27:11: No PreCompute nodes found, or all already computed. Skipping pre-computation step.

12/20/2016 15:27:11: Starting Epoch 1: learning rate per sample = 0.003125  effective momentum = 0.900000  momentum as time constant = 2429.8 samples

12/20/2016 15:27:11: Starting minibatch loop.
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[   1-  10]: CE.SM = 5.11566734 * 2560; Err = 0.93046875 * 2560; time = 0.6622s; samplesPerSecond = 3866.0
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[  11-  20]: CE.SM = 3.51283913 * 2560; Err = 0.82265625 * 2560; time = 0.0174s; samplesPerSecond = 146999.7
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[  21-  30]: CE.SM = 2.89806061 * 2560; Err = 0.72382813 * 2560; time = 0.0175s; samplesPerSecond = 146085.4
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[  31-  40]: CE.SM = 2.66997070 * 2560; Err = 0.70742187 * 2560; time = 0.0174s; samplesPerSecond = 147448.5
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[  41-  50]: CE.SM = 2.44573975 * 2560; Err = 0.65664062 * 2560; time = 0.0175s; samplesPerSecond = 145968.8
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[  51-  60]: CE.SM = 2.36770935 * 2560; Err = 0.62890625 * 2560; time = 0.0176s; samplesPerSecond = 145827.4
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[  61-  70]: CE.SM = 2.30219574 * 2560; Err = 0.62656250 * 2560; time = 0.0174s; samplesPerSecond = 146831.1
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[  71-  80]: CE.SM = 2.23771820 * 2560; Err = 0.60976562 * 2560; time = 0.0173s; samplesPerSecond = 147669.6
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[  81-  90]: CE.SM = 2.18259735 * 2560; Err = 0.60781250 * 2560; time = 0.0174s; samplesPerSecond = 146713.3
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[  91- 100]: CE.SM = 2.19711304 * 2560; Err = 0.60117188 * 2560; time = 0.0175s; samplesPerSecond = 146578.9
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 101- 110]: CE.SM = 2.17466431 * 2560; Err = 0.60234375 * 2560; time = 0.0175s; samplesPerSecond = 146093.7
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 111- 120]: CE.SM = 2.16450195 * 2560; Err = 0.60781250 * 2560; time = 0.0175s; samplesPerSecond = 146612.5
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 121- 130]: CE.SM = 2.04068909 * 2560; Err = 0.58867187 * 2560; time = 0.0175s; samplesPerSecond = 146562.1
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 131- 140]: CE.SM = 2.06387024 * 2560; Err = 0.59492188 * 2560; time = 0.0172s; samplesPerSecond = 149192.8
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 141- 150]: CE.SM = 2.02385864 * 2560; Err = 0.56757813 * 2560; time = 0.0175s; samplesPerSecond = 145885.6
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 151- 160]: CE.SM = 2.10035706 * 2560; Err = 0.59023437 * 2560; time = 0.0175s; samplesPerSecond = 146060.4
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 161- 170]: CE.SM = 2.04062500 * 2560; Err = 0.58203125 * 2560; time = 0.0174s; samplesPerSecond = 147058.8
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 171- 180]: CE.SM = 2.02286377 * 2560; Err = 0.57578125 * 2560; time = 0.0173s; samplesPerSecond = 148302.6
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 181- 190]: CE.SM = 2.01557312 * 2560; Err = 0.56171875 * 2560; time = 0.0175s; samplesPerSecond = 146369.4
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 191- 200]: CE.SM = 1.99633789 * 2560; Err = 0.57187500 * 2560; time = 0.0172s; samplesPerSecond = 148405.8
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 201- 210]: CE.SM = 2.03856201 * 2560; Err = 0.58203125 * 2560; time = 0.0174s; samplesPerSecond = 147329.7
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 211- 220]: CE.SM = 1.95707703 * 2560; Err = 0.55781250 * 2560; time = 0.0175s; samplesPerSecond = 146093.7
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 221- 230]: CE.SM = 2.02566528 * 2560; Err = 0.58085937 * 2560; time = 0.0175s; samplesPerSecond = 146002.1
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 231- 240]: CE.SM = 2.03840332 * 2560; Err = 0.57656250 * 2560; time = 0.0175s; samplesPerSecond = 146679.7
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 241- 250]: CE.SM = 1.92643433 * 2560; Err = 0.56718750 * 2560; time = 0.0171s; samplesPerSecond = 149323.4
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 251- 260]: CE.SM = 1.97182007 * 2560; Err = 0.56953125 * 2560; time = 0.0173s; samplesPerSecond = 147763.3
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 261- 270]: CE.SM = 1.99577637 * 2560; Err = 0.55781250 * 2560; time = 0.0173s; samplesPerSecond = 148302.6
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 271- 280]: CE.SM = 2.02746582 * 2560; Err = 0.56835938 * 2560; time = 0.0173s; samplesPerSecond = 147882.8
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 281- 290]: CE.SM = 1.99648438 * 2560; Err = 0.57109375 * 2560; time = 0.0175s; samplesPerSecond = 145877.3
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 291- 300]: CE.SM = 1.96296997 * 2560; Err = 0.56640625 * 2560; time = 0.0175s; samplesPerSecond = 146662.8
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 301- 310]: CE.SM = 1.96181641 * 2560; Err = 0.55781250 * 2560; time = 0.0174s; samplesPerSecond = 147490.9
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 311- 320]: CE.SM = 1.91018066 * 2560; Err = 0.55585938 * 2560; time = 0.0173s; samplesPerSecond = 148173.9
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 321- 330]: CE.SM = 1.97628784 * 2560; Err = 0.54687500 * 2560; time = 0.0175s; samplesPerSecond = 146629.2
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 331- 340]: CE.SM = 1.97587280 * 2560; Err = 0.56328125 * 2560; time = 0.0175s; samplesPerSecond = 146620.8
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 341- 350]: CE.SM = 1.93742676 * 2560; Err = 0.56289062 * 2560; time = 0.0175s; samplesPerSecond = 146595.7
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 351- 360]: CE.SM = 1.95067139 * 2560; Err = 0.56406250 * 2560; time = 0.0174s; samplesPerSecond = 146932.2
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 361- 370]: CE.SM = 1.90878296 * 2560; Err = 0.55468750 * 2560; time = 0.0172s; samplesPerSecond = 148716.2
12/20/2016 15:27:12:  Epoch[ 1 of 25]-Minibatch[ 371- 380]: CE.SM = 1.94089355 * 2560; Err = 0.55195313 * 2560; time = 0.0193s; samplesPerSecond = 132601.3
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 381- 390]: CE.SM = 1.91099854 * 2560; Err = 0.54218750 * 2560; time = 0.0321s; samplesPerSecond = 79661.4
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 391- 400]: CE.SM = 1.97081909 * 2560; Err = 0.55625000 * 2560; time = 0.0172s; samplesPerSecond = 148889.1
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 401- 410]: CE.SM = 1.96477661 * 2560; Err = 0.56445312 * 2560; time = 0.0174s; samplesPerSecond = 147194.1
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 411- 420]: CE.SM = 1.92213135 * 2560; Err = 0.54648438 * 2560; time = 0.0172s; samplesPerSecond = 148742.1
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 421- 430]: CE.SM = 1.92429810 * 2560; Err = 0.54101562 * 2560; time = 0.0175s; samplesPerSecond = 146010.4
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 431- 440]: CE.SM = 1.89721069 * 2560; Err = 0.53867188 * 2560; time = 0.0174s; samplesPerSecond = 146932.2
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 441- 450]: CE.SM = 1.94199219 * 2560; Err = 0.55859375 * 2560; time = 0.0174s; samplesPerSecond = 147016.6
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 451- 460]: CE.SM = 1.87410889 * 2560; Err = 0.53554687 * 2560; time = 0.0174s; samplesPerSecond = 146721.7
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 461- 470]: CE.SM = 1.90876465 * 2560; Err = 0.53750000 * 2560; time = 0.0173s; samplesPerSecond = 148019.7
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 471- 480]: CE.SM = 1.93787231 * 2560; Err = 0.55937500 * 2560; time = 0.0173s; samplesPerSecond = 148251.1
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 481- 490]: CE.SM = 1.89516602 * 2560; Err = 0.54335937 * 2560; time = 0.0174s; samplesPerSecond = 147423.0
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 491- 500]: CE.SM = 1.88304443 * 2560; Err = 0.54375000 * 2560; time = 0.0174s; samplesPerSecond = 146999.7
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 501- 510]: CE.SM = 1.88367920 * 2560; Err = 0.54492188 * 2560; time = 0.0173s; samplesPerSecond = 147754.8
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 511- 520]: CE.SM = 1.87449951 * 2560; Err = 0.53007812 * 2560; time = 0.0175s; samplesPerSecond = 146394.5
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 521- 530]: CE.SM = 1.89022217 * 2560; Err = 0.53867188 * 2560; time = 0.0172s; samplesPerSecond = 149175.5
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 531- 540]: CE.SM = 1.90467529 * 2560; Err = 0.55312500 * 2560; time = 0.0174s; samplesPerSecond = 146923.8
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 541- 550]: CE.SM = 1.92485352 * 2560; Err = 0.55312500 * 2560; time = 0.0173s; samplesPerSecond = 148251.1
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 551- 560]: CE.SM = 1.93820801 * 2560; Err = 0.55937500 * 2560; time = 0.0175s; samplesPerSecond = 146570.5
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 561- 570]: CE.SM = 1.92316895 * 2560; Err = 0.55742187 * 2560; time = 0.0175s; samplesPerSecond = 145943.8
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 571- 580]: CE.SM = 1.87982178 * 2560; Err = 0.54296875 * 2560; time = 0.0173s; samplesPerSecond = 148028.2
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 581- 590]: CE.SM = 1.86378174 * 2560; Err = 0.53203125 * 2560; time = 0.0173s; samplesPerSecond = 148071.0
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 591- 600]: CE.SM = 1.85062256 * 2560; Err = 0.54453125 * 2560; time = 0.0174s; samplesPerSecond = 147414.5
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 601- 610]: CE.SM = 1.92567139 * 2560; Err = 0.55234375 * 2560; time = 0.0175s; samplesPerSecond = 146428.0
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 611- 620]: CE.SM = 1.86652832 * 2560; Err = 0.54257813 * 2560; time = 0.0175s; samplesPerSecond = 146077.0
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 621- 630]: CE.SM = 1.86790771 * 2560; Err = 0.53476563 * 2560; time = 0.0172s; samplesPerSecond = 148733.4
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 631- 640]: CE.SM = 1.84099121 * 2560; Err = 0.54335937 * 2560; time = 0.0173s; samplesPerSecond = 148165.3
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 641- 650]: CE.SM = 1.90380859 * 2560; Err = 0.54843750 * 2560; time = 0.0174s; samplesPerSecond = 147346.6
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 651- 660]: CE.SM = 1.86322021 * 2560; Err = 0.53359375 * 2560; time = 0.0172s; samplesPerSecond = 148612.6
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 661- 670]: CE.SM = 1.88048096 * 2560; Err = 0.55625000 * 2560; time = 0.0174s; samplesPerSecond = 146890.1
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 671- 680]: CE.SM = 1.85605469 * 2560; Err = 0.54570312 * 2560; time = 0.0172s; samplesPerSecond = 148517.7
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 681- 690]: CE.SM = 1.88332520 * 2560; Err = 0.54375000 * 2560; time = 0.0241s; samplesPerSecond = 106153.6
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 691- 700]: CE.SM = 1.83688965 * 2560; Err = 0.52695313 * 2560; time = 0.0173s; samplesPerSecond = 147644.0
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 701- 710]: CE.SM = 1.92294922 * 2560; Err = 0.55898437 * 2560; time = 0.0185s; samplesPerSecond = 138460.7
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 711- 720]: CE.SM = 1.87282715 * 2560; Err = 0.54492188 * 2560; time = 0.0175s; samplesPerSecond = 146085.4
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 721- 730]: CE.SM = 1.83375244 * 2560; Err = 0.52812500 * 2560; time = 0.0173s; samplesPerSecond = 147686.6
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 731- 740]: CE.SM = 1.90773926 * 2560; Err = 0.54648438 * 2560; time = 0.0173s; samplesPerSecond = 147737.8
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 741- 750]: CE.SM = 1.83405762 * 2560; Err = 0.53242188 * 2560; time = 0.0172s; samplesPerSecond = 148811.3
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 751- 760]: CE.SM = 1.81676025 * 2560; Err = 0.51914063 * 2560; time = 0.0169s; samplesPerSecond = 151345.0
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 761- 770]: CE.SM = 1.80989990 * 2560; Err = 0.51484375 * 2560; time = 0.0170s; samplesPerSecond = 150703.5
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 771- 780]: CE.SM = 1.85001221 * 2560; Err = 0.53593750 * 2560; time = 0.0170s; samplesPerSecond = 150473.2
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 781- 790]: CE.SM = 1.86616211 * 2560; Err = 0.53476563 * 2560; time = 0.0171s; samplesPerSecond = 149690.1
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 791- 800]: CE.SM = 1.87690430 * 2560; Err = 0.53125000 * 2560; time = 0.0170s; samplesPerSecond = 150181.9
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 801- 810]: CE.SM = 1.85938721 * 2560; Err = 0.53710938 * 2560; time = 0.0173s; samplesPerSecond = 147763.3
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 811- 820]: CE.SM = 1.85238037 * 2560; Err = 0.53398437 * 2560; time = 0.0173s; samplesPerSecond = 148036.8
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 821- 830]: CE.SM = 1.84261475 * 2560; Err = 0.52343750 * 2560; time = 0.0174s; samplesPerSecond = 147423.0
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 831- 840]: CE.SM = 1.80770264 * 2560; Err = 0.52187500 * 2560; time = 0.0175s; samplesPerSecond = 146587.3
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 841- 850]: CE.SM = 1.86235352 * 2560; Err = 0.53085938 * 2560; time = 0.0168s; samplesPerSecond = 152199.8
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 851- 860]: CE.SM = 1.82082520 * 2560; Err = 0.53242188 * 2560; time = 0.0169s; samplesPerSecond = 151775.7
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 861- 870]: CE.SM = 1.87440186 * 2560; Err = 0.53710938 * 2560; time = 0.0171s; samplesPerSecond = 149997.1
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 871- 880]: CE.SM = 1.86011963 * 2560; Err = 0.53671875 * 2560; time = 0.0170s; samplesPerSecond = 150499.7
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 881- 890]: CE.SM = 1.86102295 * 2560; Err = 0.53125000 * 2560; time = 0.0171s; samplesPerSecond = 149497.8
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 891- 900]: CE.SM = 1.82933350 * 2560; Err = 0.52265625 * 2560; time = 0.0171s; samplesPerSecond = 149812.7
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 901- 910]: CE.SM = 1.91491699 * 2560; Err = 0.54726562 * 2560; time = 0.0171s; samplesPerSecond = 149655.1
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 911- 920]: CE.SM = 1.82818604 * 2560; Err = 0.52890625 * 2560; time = 0.0170s; samplesPerSecond = 150155.4
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 921- 930]: CE.SM = 1.86141357 * 2560; Err = 0.54609375 * 2560; time = 0.0171s; samplesPerSecond = 149786.4
12/20/2016 15:27:13:  Epoch[ 1 of 25]-Minibatch[ 931- 940]: CE.SM = 1.80335693 * 2560; Err = 0.52773437 * 2560; time = 0.0169s; samplesPerSecond = 151157.3
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[ 941- 950]: CE.SM = 1.80914307 * 2560; Err = 0.52226562 * 2560; time = 0.0172s; samplesPerSecond = 148984.5
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[ 951- 960]: CE.SM = 1.86401367 * 2560; Err = 0.53046875 * 2560; time = 0.0173s; samplesPerSecond = 147882.8
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[ 961- 970]: CE.SM = 1.84821777 * 2560; Err = 0.54531250 * 2560; time = 0.0187s; samplesPerSecond = 136649.9
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[ 971- 980]: CE.SM = 1.80655518 * 2560; Err = 0.53125000 * 2560; time = 0.0177s; samplesPerSecond = 144502.1
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[ 981- 990]: CE.SM = 1.79488525 * 2560; Err = 0.49843750 * 2560; time = 0.0174s; samplesPerSecond = 147118.0
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[ 991-1000]: CE.SM = 1.81378174 * 2560; Err = 0.52734375 * 2560; time = 0.0175s; samplesPerSecond = 146386.1
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1001-1010]: CE.SM = 1.82830811 * 2560; Err = 0.52890625 * 2560; time = 0.0176s; samplesPerSecond = 145355.4
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1011-1020]: CE.SM = 1.81676025 * 2560; Err = 0.51679688 * 2560; time = 0.0175s; samplesPerSecond = 146386.1
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1021-1030]: CE.SM = 1.80413818 * 2560; Err = 0.52500000 * 2560; time = 0.0175s; samplesPerSecond = 146302.4
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1031-1040]: CE.SM = 1.75466309 * 2560; Err = 0.51601562 * 2560; time = 0.0173s; samplesPerSecond = 147669.6
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1041-1050]: CE.SM = 1.82426758 * 2560; Err = 0.53281250 * 2560; time = 0.0173s; samplesPerSecond = 148088.2
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1051-1060]: CE.SM = 1.83310547 * 2560; Err = 0.52929688 * 2560; time = 0.0173s; samplesPerSecond = 147627.0
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1061-1070]: CE.SM = 1.80776367 * 2560; Err = 0.53789062 * 2560; time = 0.0174s; samplesPerSecond = 147278.8
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1071-1080]: CE.SM = 1.83149414 * 2560; Err = 0.54296875 * 2560; time = 0.0175s; samplesPerSecond = 146419.6
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1081-1090]: CE.SM = 1.80876465 * 2560; Err = 0.52500000 * 2560; time = 0.0174s; samplesPerSecond = 147126.4
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1091-1100]: CE.SM = 1.84860840 * 2560; Err = 0.53710938 * 2560; time = 0.0174s; samplesPerSecond = 147253.4
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1101-1110]: CE.SM = 1.79311523 * 2560; Err = 0.51875000 * 2560; time = 0.0173s; samplesPerSecond = 148028.2
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1111-1120]: CE.SM = 1.83002930 * 2560; Err = 0.51562500 * 2560; time = 0.0175s; samplesPerSecond = 146520.1
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1121-1130]: CE.SM = 1.78715820 * 2560; Err = 0.51796875 * 2560; time = 0.0174s; samplesPerSecond = 147177.2
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1131-1140]: CE.SM = 1.83859863 * 2560; Err = 0.52812500 * 2560; time = 0.0175s; samplesPerSecond = 146302.4
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1141-1150]: CE.SM = 1.77058105 * 2560; Err = 0.51210937 * 2560; time = 0.0175s; samplesPerSecond = 146654.4
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1151-1160]: CE.SM = 1.75178223 * 2560; Err = 0.52187500 * 2560; time = 0.0172s; samplesPerSecond = 148932.5
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1161-1170]: CE.SM = 1.77368164 * 2560; Err = 0.50078125 * 2560; time = 0.0175s; samplesPerSecond = 146595.7
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1171-1180]: CE.SM = 1.76914062 * 2560; Err = 0.51015625 * 2560; time = 0.0174s; samplesPerSecond = 147406.0
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1181-1190]: CE.SM = 1.84548340 * 2560; Err = 0.52460938 * 2560; time = 0.0174s; samplesPerSecond = 147253.4
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1191-1200]: CE.SM = 1.84387207 * 2560; Err = 0.52617187 * 2560; time = 0.0174s; samplesPerSecond = 147101.1
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1201-1210]: CE.SM = 1.82309570 * 2560; Err = 0.53007812 * 2560; time = 0.0175s; samplesPerSecond = 146110.4
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1211-1220]: CE.SM = 1.81940918 * 2560; Err = 0.51367188 * 2560; time = 0.0172s; samplesPerSecond = 148794.0
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1221-1230]: CE.SM = 1.81203613 * 2560; Err = 0.53750000 * 2560; time = 0.0190s; samplesPerSecond = 134623.5
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1231-1240]: CE.SM = 1.80017090 * 2560; Err = 0.52187500 * 2560; time = 0.0171s; samplesPerSecond = 149891.7
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1241-1250]: CE.SM = 1.77661133 * 2560; Err = 0.51171875 * 2560; time = 0.0171s; samplesPerSecond = 149550.2
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1251-1260]: CE.SM = 1.77321777 * 2560; Err = 0.51523438 * 2560; time = 0.0173s; samplesPerSecond = 147848.7
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1261-1270]: CE.SM = 1.78930664 * 2560; Err = 0.50937500 * 2560; time = 0.0172s; samplesPerSecond = 148897.8
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1271-1280]: CE.SM = 1.80104980 * 2560; Err = 0.52343750 * 2560; time = 0.0170s; samplesPerSecond = 150721.2
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1281-1290]: CE.SM = 1.81958008 * 2560; Err = 0.53242188 * 2560; time = 0.0172s; samplesPerSecond = 148612.6
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1291-1300]: CE.SM = 1.81730957 * 2560; Err = 0.51796875 * 2560; time = 0.0174s; samplesPerSecond = 147126.4
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1301-1310]: CE.SM = 1.77548828 * 2560; Err = 0.52968750 * 2560; time = 0.0173s; samplesPerSecond = 147917.0
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1311-1320]: CE.SM = 1.79035645 * 2560; Err = 0.52070313 * 2560; time = 0.0173s; samplesPerSecond = 148268.3
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1321-1330]: CE.SM = 1.81232910 * 2560; Err = 0.52460938 * 2560; time = 0.0173s; samplesPerSecond = 148122.4
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1331-1340]: CE.SM = 1.81640625 * 2560; Err = 0.53125000 * 2560; time = 0.0171s; samplesPerSecond = 149909.2
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1341-1350]: CE.SM = 1.73366699 * 2560; Err = 0.51523438 * 2560; time = 0.0173s; samplesPerSecond = 148319.8
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1351-1360]: CE.SM = 1.77707520 * 2560; Err = 0.51210937 * 2560; time = 0.0170s; samplesPerSecond = 150482.0
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1361-1370]: CE.SM = 1.79316406 * 2560; Err = 0.52265625 * 2560; time = 0.0170s; samplesPerSecond = 150703.5
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1371-1380]: CE.SM = 1.80944824 * 2560; Err = 0.53476563 * 2560; time = 0.0171s; samplesPerSecond = 149602.6
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1381-1390]: CE.SM = 1.83227539 * 2560; Err = 0.53007812 * 2560; time = 0.0169s; samplesPerSecond = 151273.4
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1391-1400]: CE.SM = 1.72832031 * 2560; Err = 0.51289063 * 2560; time = 0.0167s; samplesPerSecond = 152991.1
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1401-1410]: CE.SM = 1.83105469 * 2560; Err = 0.52421875 * 2560; time = 0.0169s; samplesPerSecond = 151228.7
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1411-1420]: CE.SM = 1.81799316 * 2560; Err = 0.53164062 * 2560; time = 0.0173s; samplesPerSecond = 147610.0
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1421-1430]: CE.SM = 1.77565918 * 2560; Err = 0.52734375 * 2560; time = 0.0174s; samplesPerSecond = 147278.8
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1431-1440]: CE.SM = 1.81552734 * 2560; Err = 0.52773437 * 2560; time = 0.0174s; samplesPerSecond = 146780.6
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1441-1450]: CE.SM = 1.75275879 * 2560; Err = 0.52539062 * 2560; time = 0.0173s; samplesPerSecond = 148062.5
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1451-1460]: CE.SM = 1.78322754 * 2560; Err = 0.52695313 * 2560; time = 0.0175s; samplesPerSecond = 146302.4
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1461-1470]: CE.SM = 1.80546875 * 2560; Err = 0.52500000 * 2560; time = 0.0174s; samplesPerSecond = 147075.7
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1471-1480]: CE.SM = 1.73540039 * 2560; Err = 0.49570313 * 2560; time = 0.0175s; samplesPerSecond = 146696.5
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1481-1490]: CE.SM = 1.77341309 * 2560; Err = 0.50625000 * 2560; time = 0.0174s; samplesPerSecond = 147168.7
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1491-1500]: CE.SM = 1.76943359 * 2560; Err = 0.51679688 * 2560; time = 0.0194s; samplesPerSecond = 131633.1
12/20/2016 15:27:14:  Epoch[ 1 of 25]-Minibatch[1501-1510]: CE.SM = 1.78674316 * 2560; Err = 0.53007812 * 2560; time = 0.0171s; samplesPerSecond = 149297.3
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1511-1520]: CE.SM = 1.75324707 * 2560; Err = 0.50625000 * 2560; time = 0.0174s; samplesPerSecond = 147101.1
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1521-1530]: CE.SM = 1.81137695 * 2560; Err = 0.51289063 * 2560; time = 0.0175s; samplesPerSecond = 146218.9
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1531-1540]: CE.SM = 1.79890137 * 2560; Err = 0.52421875 * 2560; time = 0.0171s; samplesPerSecond = 149366.9
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1541-1550]: CE.SM = 1.87280273 * 2560; Err = 0.54843750 * 2560; time = 0.0175s; samplesPerSecond = 146478.2
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1551-1560]: CE.SM = 1.77636719 * 2560; Err = 0.52656250 * 2560; time = 0.0176s; samplesPerSecond = 145545.5
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1561-1570]: CE.SM = 1.75302734 * 2560; Err = 0.51718750 * 2560; time = 0.0174s; samplesPerSecond = 147304.2
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1571-1580]: CE.SM = 1.73691406 * 2560; Err = 0.50078125 * 2560; time = 0.0184s; samplesPerSecond = 138866.3
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1581-1590]: CE.SM = 1.78579102 * 2560; Err = 0.51523438 * 2560; time = 0.0176s; samplesPerSecond = 145124.7
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1591-1600]: CE.SM = 1.77629395 * 2560; Err = 0.51054687 * 2560; time = 0.0172s; samplesPerSecond = 148819.9
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1601-1610]: CE.SM = 1.74787598 * 2560; Err = 0.51445312 * 2560; time = 0.0175s; samplesPerSecond = 145885.6
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1611-1620]: CE.SM = 1.77971191 * 2560; Err = 0.51718750 * 2560; time = 0.0172s; samplesPerSecond = 148655.7
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1621-1630]: CE.SM = 1.76335449 * 2560; Err = 0.52578125 * 2560; time = 0.0173s; samplesPerSecond = 147917.0
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1631-1640]: CE.SM = 1.74431152 * 2560; Err = 0.50195312 * 2560; time = 0.0175s; samplesPerSecond = 146654.4
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1641-1650]: CE.SM = 1.76313477 * 2560; Err = 0.51953125 * 2560; time = 0.0175s; samplesPerSecond = 146269.0
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1651-1660]: CE.SM = 1.76520996 * 2560; Err = 0.52421875 * 2560; time = 0.0173s; samplesPerSecond = 147934.1
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1661-1670]: CE.SM = 1.75966797 * 2560; Err = 0.51640625 * 2560; time = 0.0174s; samplesPerSecond = 147414.5
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1671-1680]: CE.SM = 1.78120117 * 2560; Err = 0.51757812 * 2560; time = 0.0174s; samplesPerSecond = 147295.7
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1681-1690]: CE.SM = 1.78066406 * 2560; Err = 0.53671875 * 2560; time = 0.0173s; samplesPerSecond = 147618.5
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1691-1700]: CE.SM = 1.73959961 * 2560; Err = 0.52343750 * 2560; time = 0.0175s; samplesPerSecond = 145985.4
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1701-1710]: CE.SM = 1.68537598 * 2560; Err = 0.47773437 * 2560; time = 0.0177s; samplesPerSecond = 144960.4
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1711-1720]: CE.SM = 1.77946777 * 2560; Err = 0.51679688 * 2560; time = 0.0175s; samplesPerSecond = 146243.9
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1721-1730]: CE.SM = 1.78002930 * 2560; Err = 0.50703125 * 2560; time = 0.0177s; samplesPerSecond = 144976.8
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1731-1740]: CE.SM = 1.76159668 * 2560; Err = 0.51562500 * 2560; time = 0.0174s; samplesPerSecond = 147456.9
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1741-1750]: CE.SM = 1.76020508 * 2560; Err = 0.51796875 * 2560; time = 0.0173s; samplesPerSecond = 147644.0
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1751-1760]: CE.SM = 1.74714355 * 2560; Err = 0.51171875 * 2560; time = 0.0175s; samplesPerSecond = 146578.9
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1761-1770]: CE.SM = 1.69831543 * 2560; Err = 0.49804688 * 2560; time = 0.0175s; samplesPerSecond = 146210.5
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1771-1780]: CE.SM = 1.77695312 * 2560; Err = 0.51484375 * 2560; time = 0.0174s; samplesPerSecond = 147482.4
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1781-1790]: CE.SM = 1.72814941 * 2560; Err = 0.50585938 * 2560; time = 0.0174s; samplesPerSecond = 146839.5
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1791-1800]: CE.SM = 1.77219238 * 2560; Err = 0.51328125 * 2560; time = 0.0173s; samplesPerSecond = 148268.3
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1801-1810]: CE.SM = 1.76748047 * 2560; Err = 0.49960938 * 2560; time = 0.0177s; samplesPerSecond = 144347.3
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1811-1820]: CE.SM = 1.72573242 * 2560; Err = 0.51640625 * 2560; time = 0.0175s; samplesPerSecond = 146361.0
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1821-1830]: CE.SM = 1.70124512 * 2560; Err = 0.50000000 * 2560; time = 0.0174s; samplesPerSecond = 146772.2
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1831-1840]: CE.SM = 1.75100098 * 2560; Err = 0.51953125 * 2560; time = 0.0173s; samplesPerSecond = 148311.2
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1841-1850]: CE.SM = 1.77243652 * 2560; Err = 0.52148438 * 2560; time = 0.0173s; samplesPerSecond = 147771.9
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1851-1860]: CE.SM = 1.72551270 * 2560; Err = 0.50625000 * 2560; time = 0.0173s; samplesPerSecond = 147874.3
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1861-1870]: CE.SM = 1.74306641 * 2560; Err = 0.49062500 * 2560; time = 0.0176s; samplesPerSecond = 145553.8
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1871-1880]: CE.SM = 1.78432617 * 2560; Err = 0.51718750 * 2560; time = 0.0176s; samplesPerSecond = 145711.2
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1881-1890]: CE.SM = 1.80900879 * 2560; Err = 0.52890625 * 2560; time = 0.0174s; samplesPerSecond = 147025.0
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1891-1900]: CE.SM = 1.73415527 * 2560; Err = 0.50898438 * 2560; time = 0.0175s; samplesPerSecond = 146361.0
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1901-1910]: CE.SM = 1.71313477 * 2560; Err = 0.51953125 * 2560; time = 0.0172s; samplesPerSecond = 148681.6
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1911-1920]: CE.SM = 1.75603027 * 2560; Err = 0.50273437 * 2560; time = 0.0174s; samplesPerSecond = 146789.0
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1921-1930]: CE.SM = 1.74335938 * 2560; Err = 0.51562500 * 2560; time = 0.0175s; samplesPerSecond = 145877.3
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1931-1940]: CE.SM = 1.75151367 * 2560; Err = 0.50937500 * 2560; time = 0.0174s; samplesPerSecond = 147507.9
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1941-1950]: CE.SM = 1.75214844 * 2560; Err = 0.51132813 * 2560; time = 0.0174s; samplesPerSecond = 146898.5
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1951-1960]: CE.SM = 1.73332520 * 2560; Err = 0.51093750 * 2560; time = 0.0175s; samplesPerSecond = 146118.7
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1961-1970]: CE.SM = 1.77368164 * 2560; Err = 0.51054687 * 2560; time = 0.0173s; samplesPerSecond = 148113.9
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1971-1980]: CE.SM = 1.71748047 * 2560; Err = 0.50546875 * 2560; time = 0.0172s; samplesPerSecond = 148733.4
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1981-1990]: CE.SM = 1.73740234 * 2560; Err = 0.50429687 * 2560; time = 0.0173s; samplesPerSecond = 147823.1
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[1991-2000]: CE.SM = 1.72890625 * 2560; Err = 0.49492188 * 2560; time = 0.0173s; samplesPerSecond = 148113.9
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[2001-2010]: CE.SM = 1.73110352 * 2560; Err = 0.49609375 * 2560; time = 0.0174s; samplesPerSecond = 146831.1
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[2011-2020]: CE.SM = 1.76789551 * 2560; Err = 0.51171875 * 2560; time = 0.0174s; samplesPerSecond = 147075.7
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[2021-2030]: CE.SM = 1.74030762 * 2560; Err = 0.51210937 * 2560; time = 0.0172s; samplesPerSecond = 148975.8
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[2031-2040]: CE.SM = 1.72604980 * 2560; Err = 0.50507813 * 2560; time = 0.0174s; samplesPerSecond = 146957.5
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[2041-2050]: CE.SM = 1.79108887 * 2560; Err = 0.53437500 * 2560; time = 0.0175s; samplesPerSecond = 146637.6
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[2051-2060]: CE.SM = 1.74147949 * 2560; Err = 0.51171875 * 2560; time = 0.0176s; samplesPerSecond = 145595.2
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[2061-2070]: CE.SM = 1.70571289 * 2560; Err = 0.50195312 * 2560; time = 0.0174s; samplesPerSecond = 147211.0
12/20/2016 15:27:15:  Epoch[ 1 of 25]-Minibatch[2071-2080]: CE.SM = 1.73632812 * 2560; Err = 0.51718750 * 2560; time = 0.0171s; samplesPerSecond = 149393.1
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2081-2090]: CE.SM = 1.74313965 * 2560; Err = 0.51015625 * 2560; time = 0.0178s; samplesPerSecond = 143529.9
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2091-2100]: CE.SM = 1.71164551 * 2560; Err = 0.49804688 * 2560; time = 0.0176s; samplesPerSecond = 145289.4
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2101-2110]: CE.SM = 1.68383789 * 2560; Err = 0.48867187 * 2560; time = 0.0174s; samplesPerSecond = 146831.1
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2111-2120]: CE.SM = 1.73923340 * 2560; Err = 0.51914063 * 2560; time = 0.0173s; samplesPerSecond = 148002.5
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2121-2130]: CE.SM = 1.68823242 * 2560; Err = 0.51289063 * 2560; time = 0.0174s; samplesPerSecond = 147372.1
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2131-2140]: CE.SM = 1.71218262 * 2560; Err = 0.49570313 * 2560; time = 0.0172s; samplesPerSecond = 148440.2
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2141-2150]: CE.SM = 1.71628418 * 2560; Err = 0.51289063 * 2560; time = 0.0176s; samplesPerSecond = 145595.2
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2151-2160]: CE.SM = 1.68269043 * 2560; Err = 0.48828125 * 2560; time = 0.0175s; samplesPerSecond = 146612.5
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2161-2170]: CE.SM = 1.70478516 * 2560; Err = 0.49960938 * 2560; time = 0.0175s; samplesPerSecond = 145918.8
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2171-2180]: CE.SM = 1.71713867 * 2560; Err = 0.51914063 * 2560; time = 0.0175s; samplesPerSecond = 146688.1
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2181-2190]: CE.SM = 1.71101074 * 2560; Err = 0.51250000 * 2560; time = 0.0174s; samplesPerSecond = 146822.7
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2191-2200]: CE.SM = 1.66967773 * 2560; Err = 0.49335937 * 2560; time = 0.0171s; samplesPerSecond = 149358.2
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2201-2210]: CE.SM = 1.69418945 * 2560; Err = 0.49531250 * 2560; time = 0.0172s; samplesPerSecond = 148932.5
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2211-2220]: CE.SM = 1.67172852 * 2560; Err = 0.49101563 * 2560; time = 0.0170s; samplesPerSecond = 150411.3
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2221-2230]: CE.SM = 1.71679688 * 2560; Err = 0.49101563 * 2560; time = 0.0172s; samplesPerSecond = 148560.8
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2231-2240]: CE.SM = 1.68652344 * 2560; Err = 0.49335937 * 2560; time = 0.0173s; samplesPerSecond = 147584.5
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2241-2250]: CE.SM = 1.69594727 * 2560; Err = 0.50195312 * 2560; time = 0.0174s; samplesPerSecond = 147202.6
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2251-2260]: CE.SM = 1.78232422 * 2560; Err = 0.51523438 * 2560; time = 0.0174s; samplesPerSecond = 147312.7
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2261-2270]: CE.SM = 1.65332031 * 2560; Err = 0.49257812 * 2560; time = 0.0173s; samplesPerSecond = 148405.8
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2271-2280]: CE.SM = 1.67636719 * 2560; Err = 0.50468750 * 2560; time = 0.0175s; samplesPerSecond = 146260.6
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2281-2290]: CE.SM = 1.69033203 * 2560; Err = 0.49218750 * 2560; time = 0.0192s; samplesPerSecond = 133416.7
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2291-2300]: CE.SM = 1.68276367 * 2560; Err = 0.48945312 * 2560; time = 0.0174s; samplesPerSecond = 147473.9
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2301-2310]: CE.SM = 1.70795898 * 2560; Err = 0.49296875 * 2560; time = 0.0175s; samplesPerSecond = 146620.8
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2311-2320]: CE.SM = 1.71411133 * 2560; Err = 0.51132813 * 2560; time = 0.0174s; samplesPerSecond = 147092.6
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2321-2330]: CE.SM = 1.70336914 * 2560; Err = 0.50703125 * 2560; time = 0.0175s; samplesPerSecond = 145902.2
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2331-2340]: CE.SM = 1.66347656 * 2560; Err = 0.49804688 * 2560; time = 0.0175s; samplesPerSecond = 146193.8
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2341-2350]: CE.SM = 1.63515625 * 2560; Err = 0.49882813 * 2560; time = 0.0175s; samplesPerSecond = 146118.7
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2351-2360]: CE.SM = 1.68837891 * 2560; Err = 0.49609375 * 2560; time = 0.0174s; samplesPerSecond = 147550.4
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2361-2370]: CE.SM = 1.74882812 * 2560; Err = 0.51132813 * 2560; time = 0.0172s; samplesPerSecond = 149001.8
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2371-2380]: CE.SM = 1.68291016 * 2560; Err = 0.50195312 * 2560; time = 0.0175s; samplesPerSecond = 146696.5
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2381-2390]: CE.SM = 1.70029297 * 2560; Err = 0.50000000 * 2560; time = 0.0174s; samplesPerSecond = 146949.1
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2391-2400]: CE.SM = 1.71958008 * 2560; Err = 0.50820312 * 2560; time = 0.0174s; samplesPerSecond = 146982.8
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2401-2410]: CE.SM = 1.70219727 * 2560; Err = 0.49804688 * 2560; time = 0.0175s; samplesPerSecond = 146578.9
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2411-2420]: CE.SM = 1.71264648 * 2560; Err = 0.52109375 * 2560; time = 0.0174s; samplesPerSecond = 147448.5
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2421-2430]: CE.SM = 1.71928711 * 2560; Err = 0.51015625 * 2560; time = 0.0173s; samplesPerSecond = 148311.2
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2431-2440]: CE.SM = 1.66538086 * 2560; Err = 0.48242188 * 2560; time = 0.0175s; samplesPerSecond = 146604.1
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2441-2450]: CE.SM = 1.70864258 * 2560; Err = 0.49531250 * 2560; time = 0.0175s; samplesPerSecond = 146085.4
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2451-2460]: CE.SM = 1.73774414 * 2560; Err = 0.51367188 * 2560; time = 0.0174s; samplesPerSecond = 147423.0
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2461-2470]: CE.SM = 1.72441406 * 2560; Err = 0.50234375 * 2560; time = 0.0176s; samplesPerSecond = 145727.8
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2471-2480]: CE.SM = 1.72246094 * 2560; Err = 0.50742188 * 2560; time = 0.0173s; samplesPerSecond = 148285.4
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2481-2490]: CE.SM = 1.68798828 * 2560; Err = 0.49960938 * 2560; time = 0.0174s; samplesPerSecond = 147092.6
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2491-2500]: CE.SM = 1.68251953 * 2560; Err = 0.50507813 * 2560; time = 0.0175s; samplesPerSecond = 146252.3
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2501-2510]: CE.SM = 1.70556641 * 2560; Err = 0.51093750 * 2560; time = 0.0173s; samplesPerSecond = 148242.5
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2511-2520]: CE.SM = 1.65791016 * 2560; Err = 0.48398438 * 2560; time = 0.0176s; samplesPerSecond = 145794.2
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2521-2530]: CE.SM = 1.68974609 * 2560; Err = 0.49570313 * 2560; time = 0.0174s; samplesPerSecond = 146991.3
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2531-2540]: CE.SM = 1.66166992 * 2560; Err = 0.48671875 * 2560; time = 0.0173s; samplesPerSecond = 147823.1
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2541-2550]: CE.SM = 1.69223633 * 2560; Err = 0.50507813 * 2560; time = 0.0176s; samplesPerSecond = 145165.9
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2551-2560]: CE.SM = 1.71547852 * 2560; Err = 0.50156250 * 2560; time = 0.0174s; samplesPerSecond = 146856.4
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2561-2570]: CE.SM = 1.72509766 * 2560; Err = 0.51796875 * 2560; time = 0.0174s; samplesPerSecond = 147109.5
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2571-2580]: CE.SM = 1.72861328 * 2560; Err = 0.50234375 * 2560; time = 0.0173s; samplesPerSecond = 147874.3
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2581-2590]: CE.SM = 1.64541016 * 2560; Err = 0.48046875 * 2560; time = 0.0243s; samplesPerSecond = 105436.6
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2591-2600]: CE.SM = 1.74594727 * 2560; Err = 0.50898438 * 2560; time = 0.0174s; samplesPerSecond = 147321.2
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2601-2610]: CE.SM = 1.71752930 * 2560; Err = 0.49843750 * 2560; time = 0.0171s; samplesPerSecond = 149550.2
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2611-2620]: CE.SM = 1.67221680 * 2560; Err = 0.49062500 * 2560; time = 0.0172s; samplesPerSecond = 148664.3
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2621-2630]: CE.SM = 1.66000977 * 2560; Err = 0.49296875 * 2560; time = 0.0173s; samplesPerSecond = 148242.5
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2631-2640]: CE.SM = 1.68349609 * 2560; Err = 0.50078125 * 2560; time = 0.0175s; samplesPerSecond = 146143.7
12/20/2016 15:27:16:  Epoch[ 1 of 25]-Minibatch[2641-2650]: CE.SM = 1.70434570 * 2560; Err = 0.50468750 * 2560; time = 0.0171s; samplesPerSecond = 149550.2
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2651-2660]: CE.SM = 1.70966797 * 2560; Err = 0.50546875 * 2560; time = 0.0170s; samplesPerSecond = 150208.3
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2661-2670]: CE.SM = 1.68330078 * 2560; Err = 0.49531250 * 2560; time = 0.0173s; samplesPerSecond = 147627.0
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2671-2680]: CE.SM = 1.64716797 * 2560; Err = 0.49179688 * 2560; time = 0.0173s; samplesPerSecond = 148328.4
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2681-2690]: CE.SM = 1.59965820 * 2560; Err = 0.48007813 * 2560; time = 0.0171s; samplesPerSecond = 149497.8
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2691-2700]: CE.SM = 1.68081055 * 2560; Err = 0.50664062 * 2560; time = 0.0174s; samplesPerSecond = 147516.4
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2701-2710]: CE.SM = 1.72231445 * 2560; Err = 0.50195312 * 2560; time = 0.0169s; samplesPerSecond = 151578.0
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2711-2720]: CE.SM = 1.64697266 * 2560; Err = 0.48046875 * 2560; time = 0.0172s; samplesPerSecond = 148768.0
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2721-2730]: CE.SM = 1.68569336 * 2560; Err = 0.49531250 * 2560; time = 0.0172s; samplesPerSecond = 149175.5
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2731-2740]: CE.SM = 1.70351563 * 2560; Err = 0.49804688 * 2560; time = 0.0170s; samplesPerSecond = 150552.8
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2741-2750]: CE.SM = 1.66284180 * 2560; Err = 0.49023438 * 2560; time = 0.0171s; samplesPerSecond = 150085.0
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2751-2760]: CE.SM = 1.75429688 * 2560; Err = 0.51914063 * 2560; time = 0.0172s; samplesPerSecond = 148423.0
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2761-2770]: CE.SM = 1.65463867 * 2560; Err = 0.49765625 * 2560; time = 0.0168s; samplesPerSecond = 152245.0
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2771-2780]: CE.SM = 1.66689453 * 2560; Err = 0.49960938 * 2560; time = 0.0171s; samplesPerSecond = 149812.7
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2781-2790]: CE.SM = 1.70239258 * 2560; Err = 0.49921875 * 2560; time = 0.0173s; samplesPerSecond = 147951.2
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2791-2800]: CE.SM = 1.67812500 * 2560; Err = 0.49062500 * 2560; time = 0.0171s; samplesPerSecond = 149804.0
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2801-2810]: CE.SM = 1.69887695 * 2560; Err = 0.49257812 * 2560; time = 0.0171s; samplesPerSecond = 149471.6
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2811-2820]: CE.SM = 1.63637695 * 2560; Err = 0.48750000 * 2560; time = 0.0173s; samplesPerSecond = 148113.9
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2821-2830]: CE.SM = 1.71430664 * 2560; Err = 0.49609375 * 2560; time = 0.0174s; samplesPerSecond = 147084.2
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2831-2840]: CE.SM = 1.69736328 * 2560; Err = 0.48945312 * 2560; time = 0.0175s; samplesPerSecond = 146152.1
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2841-2850]: CE.SM = 1.67944336 * 2560; Err = 0.48789063 * 2560; time = 0.0173s; samplesPerSecond = 148294.0
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2851-2860]: CE.SM = 1.69106445 * 2560; Err = 0.50234375 * 2560; time = 0.0174s; samplesPerSecond = 147287.3
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2861-2870]: CE.SM = 1.70390625 * 2560; Err = 0.50000000 * 2560; time = 0.0174s; samplesPerSecond = 147431.5
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2871-2880]: CE.SM = 1.74125977 * 2560; Err = 0.51132813 * 2560; time = 0.0173s; samplesPerSecond = 147814.5
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2881-2890]: CE.SM = 1.63417969 * 2560; Err = 0.48085937 * 2560; time = 0.0174s; samplesPerSecond = 147550.4
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2891-2900]: CE.SM = 1.61210937 * 2560; Err = 0.48359375 * 2560; time = 0.0173s; samplesPerSecond = 148268.3
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2901-2910]: CE.SM = 1.65786133 * 2560; Err = 0.49375000 * 2560; time = 0.0195s; samplesPerSecond = 131356.1
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2911-2920]: CE.SM = 1.65039062 * 2560; Err = 0.48828125 * 2560; time = 0.0171s; samplesPerSecond = 149663.8
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2921-2930]: CE.SM = 1.63251953 * 2560; Err = 0.49414062 * 2560; time = 0.0171s; samplesPerSecond = 149349.5
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2931-2940]: CE.SM = 1.69736328 * 2560; Err = 0.50156250 * 2560; time = 0.0169s; samplesPerSecond = 151175.2
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2941-2950]: CE.SM = 1.67597656 * 2560; Err = 0.50312500 * 2560; time = 0.0172s; samplesPerSecond = 148854.5
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2951-2960]: CE.SM = 1.68232422 * 2560; Err = 0.50976562 * 2560; time = 0.0173s; samplesPerSecond = 147575.9
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2961-2970]: CE.SM = 1.67758789 * 2560; Err = 0.49921875 * 2560; time = 0.0173s; samplesPerSecond = 148354.2
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2971-2980]: CE.SM = 1.67041016 * 2560; Err = 0.49609375 * 2560; time = 0.0172s; samplesPerSecond = 148915.1
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2981-2990]: CE.SM = 1.70424805 * 2560; Err = 0.50781250 * 2560; time = 0.0173s; samplesPerSecond = 148096.7
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[2991-3000]: CE.SM = 1.66674805 * 2560; Err = 0.49296875 * 2560; time = 0.0171s; samplesPerSecond = 149489.1
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[3001-3010]: CE.SM = 1.68623047 * 2560; Err = 0.50195312 * 2560; time = 0.0171s; samplesPerSecond = 149576.4
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[3011-3020]: CE.SM = 1.68857422 * 2560; Err = 0.48828125 * 2560; time = 0.0172s; samplesPerSecond = 148958.5
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[3021-3030]: CE.SM = 1.66440430 * 2560; Err = 0.49960938 * 2560; time = 0.0172s; samplesPerSecond = 148759.4
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[3031-3040]: CE.SM = 1.64418945 * 2560; Err = 0.48671875 * 2560; time = 0.0172s; samplesPerSecond = 148768.0
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[3041-3050]: CE.SM = 1.67675781 * 2560; Err = 0.49804688 * 2560; time = 0.0171s; samplesPerSecond = 149874.1
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[3051-3060]: CE.SM = 1.65981445 * 2560; Err = 0.49453125 * 2560; time = 0.0170s; samplesPerSecond = 150261.2
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[3061-3070]: CE.SM = 1.67866211 * 2560; Err = 0.49882813 * 2560; time = 0.0172s; samplesPerSecond = 148612.6
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[3071-3080]: CE.SM = 1.69179688 * 2560; Err = 0.50703125 * 2560; time = 0.0172s; samplesPerSecond = 149166.8
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[3081-3090]: CE.SM = 1.59462891 * 2560; Err = 0.46601562 * 2560; time = 0.0171s; samplesPerSecond = 149593.9
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[3091-3100]: CE.SM = 1.66210938 * 2560; Err = 0.49843750 * 2560; time = 0.0173s; samplesPerSecond = 148388.6
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[3101-3110]: CE.SM = 1.67495117 * 2560; Err = 0.50195312 * 2560; time = 0.0172s; samplesPerSecond = 148889.1
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[3111-3120]: CE.SM = 1.66572266 * 2560; Err = 0.49843750 * 2560; time = 0.0171s; samplesPerSecond = 149804.0
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[3121-3130]: CE.SM = 1.70297852 * 2560; Err = 0.48984375 * 2560; time = 0.0171s; samplesPerSecond = 149428.0
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[3131-3140]: CE.SM = 1.67387695 * 2560; Err = 0.49296875 * 2560; time = 0.0172s; samplesPerSecond = 148742.1
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[3141-3150]: CE.SM = 1.69785156 * 2560; Err = 0.49414062 * 2560; time = 0.0171s; samplesPerSecond = 149733.9
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[3151-3160]: CE.SM = 1.69301758 * 2560; Err = 0.49179688 * 2560; time = 0.0171s; samplesPerSecond = 149393.1
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[3161-3170]: CE.SM = 1.63193359 * 2560; Err = 0.47890625 * 2560; time = 0.0171s; samplesPerSecond = 149462.9
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[3171-3180]: CE.SM = 1.57001953 * 2560; Err = 0.46875000 * 2560; time = 0.0172s; samplesPerSecond = 148915.1
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[3181-3190]: CE.SM = 1.65849609 * 2560; Err = 0.48476562 * 2560; time = 0.0186s; samplesPerSecond = 137990.5
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[3191-3200]: CE.SM = 1.63452148 * 2560; Err = 0.48789063 * 2560; time = 0.0172s; samplesPerSecond = 148440.2
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[3201-3210]: CE.SM = 1.58774414 * 2560; Err = 0.48164062 * 2560; time = 0.0175s; samplesPerSecond = 146553.7
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[3211-3220]: CE.SM = 1.65766602 * 2560; Err = 0.50000000 * 2560; time = 0.0172s; samplesPerSecond = 149088.6
12/20/2016 15:27:17:  Epoch[ 1 of 25]-Minibatch[3221-3230]: CE.SM = 1.64492188 * 2560; Err = 0.47968750 * 2560; time = 0.0170s; samplesPerSecond = 150437.8
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3231-3240]: CE.SM = 1.64199219 * 2560; Err = 0.48242188 * 2560; time = 0.0172s; samplesPerSecond = 148845.9
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3241-3250]: CE.SM = 1.69384766 * 2560; Err = 0.50351563 * 2560; time = 0.0173s; samplesPerSecond = 147976.9
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3251-3260]: CE.SM = 1.64702148 * 2560; Err = 0.49335937 * 2560; time = 0.0173s; samplesPerSecond = 148208.2
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3261-3270]: CE.SM = 1.69658203 * 2560; Err = 0.50156250 * 2560; time = 0.0173s; samplesPerSecond = 148028.2
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3271-3280]: CE.SM = 1.62885742 * 2560; Err = 0.49296875 * 2560; time = 0.0173s; samplesPerSecond = 147754.8
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3281-3290]: CE.SM = 1.66733398 * 2560; Err = 0.49179688 * 2560; time = 0.0170s; samplesPerSecond = 150676.9
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3291-3300]: CE.SM = 1.68193359 * 2560; Err = 0.50039062 * 2560; time = 0.0171s; samplesPerSecond = 149900.5
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3301-3310]: CE.SM = 1.69155273 * 2560; Err = 0.49453125 * 2560; time = 0.0172s; samplesPerSecond = 148603.9
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3311-3320]: CE.SM = 1.65400391 * 2560; Err = 0.48476562 * 2560; time = 0.0173s; samplesPerSecond = 147865.8
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3321-3330]: CE.SM = 1.67524414 * 2560; Err = 0.50312500 * 2560; time = 0.0172s; samplesPerSecond = 149114.6
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3331-3340]: CE.SM = 1.64780273 * 2560; Err = 0.48671875 * 2560; time = 0.0172s; samplesPerSecond = 148967.1
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3341-3350]: CE.SM = 1.63325195 * 2560; Err = 0.49101563 * 2560; time = 0.0170s; samplesPerSecond = 150490.9
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3351-3360]: CE.SM = 1.65039062 * 2560; Err = 0.48710938 * 2560; time = 0.0173s; samplesPerSecond = 147994.0
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3361-3370]: CE.SM = 1.67656250 * 2560; Err = 0.48398438 * 2560; time = 0.0172s; samplesPerSecond = 148543.6
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3371-3380]: CE.SM = 1.61293945 * 2560; Err = 0.48398438 * 2560; time = 0.0173s; samplesPerSecond = 147925.6
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3381-3390]: CE.SM = 1.64741211 * 2560; Err = 0.49101563 * 2560; time = 0.0172s; samplesPerSecond = 148431.6
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3391-3400]: CE.SM = 1.68369141 * 2560; Err = 0.50390625 * 2560; time = 0.0173s; samplesPerSecond = 148131.0
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3401-3410]: CE.SM = 1.56987305 * 2560; Err = 0.48007813 * 2560; time = 0.0171s; samplesPerSecond = 149830.3
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3411-3420]: CE.SM = 1.67573242 * 2560; Err = 0.49804688 * 2560; time = 0.0172s; samplesPerSecond = 148906.5
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3421-3430]: CE.SM = 1.61850586 * 2560; Err = 0.48046875 * 2560; time = 0.0173s; samplesPerSecond = 148105.3
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3431-3440]: CE.SM = 1.63847656 * 2560; Err = 0.47812500 * 2560; time = 0.0171s; samplesPerSecond = 149909.2
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3441-3450]: CE.SM = 1.65278320 * 2560; Err = 0.49804688 * 2560; time = 0.0172s; samplesPerSecond = 149210.2
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3451-3460]: CE.SM = 1.62382812 * 2560; Err = 0.48242188 * 2560; time = 0.0173s; samplesPerSecond = 148173.9
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3461-3470]: CE.SM = 1.63559570 * 2560; Err = 0.48750000 * 2560; time = 0.0170s; samplesPerSecond = 150384.8
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3471-3480]: CE.SM = 1.61313477 * 2560; Err = 0.48046875 * 2560; time = 0.0172s; samplesPerSecond = 149001.8
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3481-3490]: CE.SM = 1.66943359 * 2560; Err = 0.50234375 * 2560; time = 0.0173s; samplesPerSecond = 148036.8
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3491-3500]: CE.SM = 1.62207031 * 2560; Err = 0.48320313 * 2560; time = 0.0172s; samplesPerSecond = 148629.8
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3501-3510]: CE.SM = 1.67387695 * 2560; Err = 0.49414062 * 2560; time = 0.0174s; samplesPerSecond = 146966.0
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3511-3520]: CE.SM = 1.66225586 * 2560; Err = 0.48945312 * 2560; time = 0.0171s; samplesPerSecond = 150058.6
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3521-3530]: CE.SM = 1.67006836 * 2560; Err = 0.49140625 * 2560; time = 0.0170s; samplesPerSecond = 150517.4
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3531-3540]: CE.SM = 1.58291016 * 2560; Err = 0.46328125 * 2560; time = 0.0171s; samplesPerSecond = 150005.9
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3541-3550]: CE.SM = 1.68452148 * 2560; Err = 0.50703125 * 2560; time = 0.0170s; samplesPerSecond = 150694.6
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3551-3560]: CE.SM = 1.65468750 * 2560; Err = 0.49023438 * 2560; time = 0.0171s; samplesPerSecond = 149795.2
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3561-3570]: CE.SM = 1.58291016 * 2560; Err = 0.47539063 * 2560; time = 0.0171s; samplesPerSecond = 150023.4
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3571-3580]: CE.SM = 1.59101563 * 2560; Err = 0.48085937 * 2560; time = 0.0170s; samplesPerSecond = 150827.8
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3581-3590]: CE.SM = 1.65751953 * 2560; Err = 0.49023438 * 2560; time = 0.0169s; samplesPerSecond = 151201.9
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3591-3600]: CE.SM = 1.60180664 * 2560; Err = 0.46562500 * 2560; time = 0.0172s; samplesPerSecond = 148819.9
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3601-3610]: CE.SM = 1.69399414 * 2560; Err = 0.49921875 * 2560; time = 0.0171s; samplesPerSecond = 149375.7
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3611-3620]: CE.SM = 1.59213867 * 2560; Err = 0.46679688 * 2560; time = 0.0171s; samplesPerSecond = 149681.3
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3621-3630]: CE.SM = 1.65019531 * 2560; Err = 0.47812500 * 2560; time = 0.0170s; samplesPerSecond = 150437.8
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3631-3640]: CE.SM = 1.64223633 * 2560; Err = 0.48750000 * 2560; time = 0.0171s; samplesPerSecond = 149812.7
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3641-3650]: CE.SM = 1.61098633 * 2560; Err = 0.48046875 * 2560; time = 0.0172s; samplesPerSecond = 149071.2
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3651-3660]: CE.SM = 1.57094727 * 2560; Err = 0.46835938 * 2560; time = 0.0172s; samplesPerSecond = 149053.9
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3661-3670]: CE.SM = 1.60541992 * 2560; Err = 0.48085937 * 2560; time = 0.0172s; samplesPerSecond = 148586.7
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3671-3680]: CE.SM = 1.62265625 * 2560; Err = 0.48710938 * 2560; time = 0.0172s; samplesPerSecond = 149253.7
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3681-3690]: CE.SM = 1.63476562 * 2560; Err = 0.49960938 * 2560; time = 0.0172s; samplesPerSecond = 149079.9
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3691-3700]: CE.SM = 1.56557617 * 2560; Err = 0.47851562 * 2560; time = 0.0171s; samplesPerSecond = 149821.5
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3701-3710]: CE.SM = 1.71240234 * 2560; Err = 0.51093750 * 2560; time = 0.0171s; samplesPerSecond = 149332.1
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3711-3720]: CE.SM = 1.67480469 * 2560; Err = 0.49218750 * 2560; time = 0.0191s; samplesPerSecond = 133765.3
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3721-3730]: CE.SM = 1.63310547 * 2560; Err = 0.49062500 * 2560; time = 0.0247s; samplesPerSecond = 103534.7
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3731-3740]: CE.SM = 1.61679687 * 2560; Err = 0.48554687 * 2560; time = 0.0171s; samplesPerSecond = 149445.4
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3741-3750]: CE.SM = 1.56147461 * 2560; Err = 0.46718750 * 2560; time = 0.0170s; samplesPerSecond = 150358.3
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3751-3760]: CE.SM = 1.62753906 * 2560; Err = 0.48437500 * 2560; time = 0.0171s; samplesPerSecond = 150014.6
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3761-3770]: CE.SM = 1.64340820 * 2560; Err = 0.49453125 * 2560; time = 0.0172s; samplesPerSecond = 148500.5
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3771-3780]: CE.SM = 1.61562500 * 2560; Err = 0.47851562 * 2560; time = 0.0172s; samplesPerSecond = 149132.0
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3781-3790]: CE.SM = 1.66206055 * 2560; Err = 0.50234375 * 2560; time = 0.0173s; samplesPerSecond = 148036.8
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3791-3800]: CE.SM = 1.69536133 * 2560; Err = 0.50585938 * 2560; time = 0.0171s; samplesPerSecond = 150067.4
12/20/2016 15:27:18:  Epoch[ 1 of 25]-Minibatch[3801-3810]: CE.SM = 1.65429688 * 2560; Err = 0.50351563 * 2560; time = 0.0170s; samplesPerSecond = 150810.0
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[3811-3820]: CE.SM = 1.64780273 * 2560; Err = 0.49296875 * 2560; time = 0.0171s; samplesPerSecond = 149515.2
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[3821-3830]: CE.SM = 1.61713867 * 2560; Err = 0.47773437 * 2560; time = 0.0171s; samplesPerSecond = 149297.3
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[3831-3840]: CE.SM = 1.59174805 * 2560; Err = 0.47382812 * 2560; time = 0.0173s; samplesPerSecond = 148156.7
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[3841-3850]: CE.SM = 1.61904297 * 2560; Err = 0.48046875 * 2560; time = 0.0171s; samplesPerSecond = 149655.1
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[3851-3860]: CE.SM = 1.57558594 * 2560; Err = 0.47148438 * 2560; time = 0.0172s; samplesPerSecond = 149001.8
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[3861-3870]: CE.SM = 1.60795898 * 2560; Err = 0.47773437 * 2560; time = 0.0170s; samplesPerSecond = 150730.1
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[3871-3880]: CE.SM = 1.59658203 * 2560; Err = 0.49023438 * 2560; time = 0.0174s; samplesPerSecond = 147372.1
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[3881-3890]: CE.SM = 1.63417969 * 2560; Err = 0.47890625 * 2560; time = 0.0172s; samplesPerSecond = 149227.6
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[3891-3900]: CE.SM = 1.63017578 * 2560; Err = 0.47656250 * 2560; time = 0.0171s; samplesPerSecond = 149384.4
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[3901-3910]: CE.SM = 1.66450195 * 2560; Err = 0.49179688 * 2560; time = 0.0171s; samplesPerSecond = 149830.3
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[3911-3920]: CE.SM = 1.62250977 * 2560; Err = 0.47617188 * 2560; time = 0.0172s; samplesPerSecond = 149053.9
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[3921-3930]: CE.SM = 1.62841797 * 2560; Err = 0.48593750 * 2560; time = 0.0170s; samplesPerSecond = 150987.9
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[3931-3940]: CE.SM = 1.60654297 * 2560; Err = 0.47734375 * 2560; time = 0.0171s; samplesPerSecond = 149681.3
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[3941-3950]: CE.SM = 1.60253906 * 2560; Err = 0.47460938 * 2560; time = 0.0172s; samplesPerSecond = 149210.2
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[3951-3960]: CE.SM = 1.61328125 * 2560; Err = 0.47109375 * 2560; time = 0.0172s; samplesPerSecond = 148629.8
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[3961-3970]: CE.SM = 1.62993164 * 2560; Err = 0.48125000 * 2560; time = 0.0171s; samplesPerSecond = 149637.6
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[3971-3980]: CE.SM = 1.64707031 * 2560; Err = 0.48710938 * 2560; time = 0.0171s; samplesPerSecond = 150067.4
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[3981-3990]: CE.SM = 1.60268555 * 2560; Err = 0.47929688 * 2560; time = 0.0172s; samplesPerSecond = 149097.3
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[3991-4000]: CE.SM = 1.63803711 * 2560; Err = 0.47578125 * 2560; time = 0.0171s; samplesPerSecond = 149979.5
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4001-4010]: CE.SM = 1.65136719 * 2560; Err = 0.49609375 * 2560; time = 0.0232s; samplesPerSecond = 110378.1
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4011-4020]: CE.SM = 1.57236328 * 2560; Err = 0.46445313 * 2560; time = 0.0172s; samplesPerSecond = 148733.4
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4021-4030]: CE.SM = 1.59272461 * 2560; Err = 0.47734375 * 2560; time = 0.0172s; samplesPerSecond = 148724.8
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4031-4040]: CE.SM = 1.58505859 * 2560; Err = 0.47460938 * 2560; time = 0.0170s; samplesPerSecond = 150756.7
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4041-4050]: CE.SM = 1.61147461 * 2560; Err = 0.46484375 * 2560; time = 0.0171s; samplesPerSecond = 149471.6
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4051-4060]: CE.SM = 1.61977539 * 2560; Err = 0.48085937 * 2560; time = 0.0172s; samplesPerSecond = 149001.8
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4061-4070]: CE.SM = 1.65058594 * 2560; Err = 0.47734375 * 2560; time = 0.0173s; samplesPerSecond = 147737.8
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4071-4080]: CE.SM = 1.57202148 * 2560; Err = 0.48320313 * 2560; time = 0.0173s; samplesPerSecond = 147746.3
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4081-4090]: CE.SM = 1.55371094 * 2560; Err = 0.46289062 * 2560; time = 0.0173s; samplesPerSecond = 147925.6
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4091-4100]: CE.SM = 1.56542969 * 2560; Err = 0.46992187 * 2560; time = 0.0174s; samplesPerSecond = 146915.4
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4101-4110]: CE.SM = 1.59208984 * 2560; Err = 0.48593750 * 2560; time = 0.0173s; samplesPerSecond = 147635.5
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4111-4120]: CE.SM = 1.59995117 * 2560; Err = 0.47382812 * 2560; time = 0.0174s; samplesPerSecond = 147312.7
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4121-4130]: CE.SM = 1.61328125 * 2560; Err = 0.48164062 * 2560; time = 0.0174s; samplesPerSecond = 146805.8
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4131-4140]: CE.SM = 1.60175781 * 2560; Err = 0.47851562 * 2560; time = 0.0175s; samplesPerSecond = 146478.2
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4141-4150]: CE.SM = 1.60991211 * 2560; Err = 0.48710938 * 2560; time = 0.0173s; samplesPerSecond = 147942.7
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4151-4160]: CE.SM = 1.66992188 * 2560; Err = 0.49531250 * 2560; time = 0.0174s; samplesPerSecond = 147516.4
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4161-4170]: CE.SM = 1.62148438 * 2560; Err = 0.50039062 * 2560; time = 0.0174s; samplesPerSecond = 147194.1
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4171-4180]: CE.SM = 1.56928711 * 2560; Err = 0.47265625 * 2560; time = 0.0174s; samplesPerSecond = 146864.8
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4181-4190]: CE.SM = 1.61347656 * 2560; Err = 0.47617188 * 2560; time = 0.0173s; samplesPerSecond = 147593.0
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4191-4200]: CE.SM = 1.61342773 * 2560; Err = 0.48203125 * 2560; time = 0.0174s; samplesPerSecond = 147084.2
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4201-4210]: CE.SM = 1.58525391 * 2560; Err = 0.46796875 * 2560; time = 0.0175s; samplesPerSecond = 146210.5
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4211-4220]: CE.SM = 1.63666992 * 2560; Err = 0.50195312 * 2560; time = 0.0174s; samplesPerSecond = 147516.4
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4221-4230]: CE.SM = 1.57905273 * 2560; Err = 0.47773437 * 2560; time = 0.0170s; samplesPerSecond = 150181.9
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4231-4240]: CE.SM = 1.57563477 * 2560; Err = 0.45859375 * 2560; time = 0.0173s; samplesPerSecond = 147584.5
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4241-4250]: CE.SM = 1.63398438 * 2560; Err = 0.46562500 * 2560; time = 0.0171s; samplesPerSecond = 149821.5
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4251-4260]: CE.SM = 1.61523438 * 2560; Err = 0.48085937 * 2560; time = 0.0173s; samplesPerSecond = 147865.8
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4261-4270]: CE.SM = 1.64301758 * 2560; Err = 0.48984375 * 2560; time = 0.0169s; samplesPerSecond = 151130.5
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4271-4280]: CE.SM = 1.59770508 * 2560; Err = 0.47539063 * 2560; time = 0.0172s; samplesPerSecond = 149140.7
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4281-4290]: CE.SM = 1.64135742 * 2560; Err = 0.49687500 * 2560; time = 0.0176s; samplesPerSecond = 145289.4
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4291-4300]: CE.SM = 1.59082031 * 2560; Err = 0.48476562 * 2560; time = 0.0174s; samplesPerSecond = 147329.7
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4301-4310]: CE.SM = 1.61875000 * 2560; Err = 0.48085937 * 2560; time = 0.0174s; samplesPerSecond = 146864.8
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4311-4320]: CE.SM = 1.60073242 * 2560; Err = 0.48164062 * 2560; time = 0.0172s; samplesPerSecond = 148750.7
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4321-4330]: CE.SM = 1.57656250 * 2560; Err = 0.47382812 * 2560; time = 0.0173s; samplesPerSecond = 148225.3
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4331-4340]: CE.SM = 1.59448242 * 2560; Err = 0.48554687 * 2560; time = 0.0173s; samplesPerSecond = 148148.1
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4341-4350]: CE.SM = 1.62197266 * 2560; Err = 0.48242188 * 2560; time = 0.0175s; samplesPerSecond = 146461.5
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4351-4360]: CE.SM = 1.58061523 * 2560; Err = 0.47578125 * 2560; time = 0.0173s; samplesPerSecond = 147857.2
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4361-4370]: CE.SM = 1.57373047 * 2560; Err = 0.46601562 * 2560; time = 0.0172s; samplesPerSecond = 148431.6
12/20/2016 15:27:19:  Epoch[ 1 of 25]-Minibatch[4371-4380]: CE.SM = 1.56254883 * 2560; Err = 0.46640625 * 2560; time = 0.0173s; samplesPerSecond = 147763.3
12/20/2016 15:27:20:  Epoch[ 1 of 25]-Minibatch[4381-4390]: CE.SM = 1.57812500 * 2560; Err = 0.47109375 * 2560; time = 0.0174s; samplesPerSecond = 146873.2
12/20/2016 15:27:20: Finished Epoch[ 1 of 25]: [Training] CE.SM = 1.76313384 * 1124823; Err = 0.51372171 * 1124823; totalSamplesSeen = 1124823; learningRatePerSample = 0.003125; epochTime=8.50722s
12/20/2016 15:27:20: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn.1'

12/20/2016 15:27:20: Starting Epoch 2: learning rate per sample = 0.003125  effective momentum = 0.900000  momentum as time constant = 9719.0 samples

12/20/2016 15:27:20: Starting minibatch loop.
12/20/2016 15:27:20:  Epoch[ 2 of 25]-Minibatch[   1-  10, 0.91%]: CE.SM = 1.58246536 * 10240; Err = 0.47812500 * 10240; time = 0.0590s; samplesPerSecond = 173456.4
12/20/2016 15:27:20:  Epoch[ 2 of 25]-Minibatch[  11-  20, 1.82%]: CE.SM = 1.58246078 * 10240; Err = 0.47265625 * 10240; time = 0.0421s; samplesPerSecond = 243305.5
12/20/2016 15:27:20:  Epoch[ 2 of 25]-Minibatch[  21-  30, 2.73%]: CE.SM = 1.56565666 * 10240; Err = 0.47871094 * 10240; time = 0.0470s; samplesPerSecond = 217853.8
12/20/2016 15:27:20:  Epoch[ 2 of 25]-Minibatch[  31-  40, 3.64%]: CE.SM = 1.56890907 * 10240; Err = 0.46835938 * 10240; time = 0.0422s; samplesPerSecond = 242809.4
12/20/2016 15:27:20:  Epoch[ 2 of 25]-Minibatch[  41-  50, 4.56%]: CE.SM = 1.60285339 * 10240; Err = 0.47968750 * 10240; time = 0.0410s; samplesPerSecond = 249859.7
12/20/2016 15:27:20:  Epoch[ 2 of 25]-Minibatch[  51-  60, 5.47%]: CE.SM = 1.58846970 * 10240; Err = 0.47099609 * 10240; time = 0.0412s; samplesPerSecond = 248314.7
12/20/2016 15:27:20:  Epoch[ 2 of 25]-Minibatch[  61-  70, 6.38%]: CE.SM = 1.57636185 * 10240; Err = 0.47607422 * 10240; time = 0.0412s; samplesPerSecond = 248592.0
12/20/2016 15:27:20:  Epoch[ 2 of 25]-Minibatch[  71-  80, 7.29%]: CE.SM = 1.59845276 * 10240; Err = 0.48281250 * 10240; time = 0.0417s; samplesPerSecond = 245805.2
12/20/2016 15:27:20:  Epoch[ 2 of 25]-Minibatch[  81-  90, 8.20%]: CE.SM = 1.59133606 * 10240; Err = 0.47304687 * 10240; time = 0.0399s; samplesPerSecond = 256924.9
12/20/2016 15:27:20:  Epoch[ 2 of 25]-Minibatch[  91- 100, 9.11%]: CE.SM = 1.56974182 * 10240; Err = 0.47216797 * 10240; time = 0.0411s; samplesPerSecond = 249093.9
12/20/2016 15:27:20:  Epoch[ 2 of 25]-Minibatch[ 101- 110, 10.02%]: CE.SM = 1.60553894 * 10240; Err = 0.47177734 * 10240; time = 0.0485s; samplesPerSecond = 211312.7
12/20/2016 15:27:20:  Epoch[ 2 of 25]-Minibatch[ 111- 120, 10.93%]: CE.SM = 1.53762970 * 10240; Err = 0.46513672 * 10240; time = 0.0428s; samplesPerSecond = 239090.3
12/20/2016 15:27:20:  Epoch[ 2 of 25]-Minibatch[ 121- 130, 11.85%]: CE.SM = 1.56923065 * 10240; Err = 0.46416016 * 10240; time = 0.0463s; samplesPerSecond = 221309.7
12/20/2016 15:27:20:  Epoch[ 2 of 25]-Minibatch[ 131- 140, 12.76%]: CE.SM = 1.54934387 * 10240; Err = 0.46416016 * 10240; time = 0.0417s; samplesPerSecond = 245334.1
12/20/2016 15:27:20:  Epoch[ 2 of 25]-Minibatch[ 141- 150, 13.67%]: CE.SM = 1.56606445 * 10240; Err = 0.47314453 * 10240; time = 0.0404s; samplesPerSecond = 253327.4
12/20/2016 15:27:20:  Epoch[ 2 of 25]-Minibatch[ 151- 160, 14.58%]: CE.SM = 1.57536621 * 10240; Err = 0.47021484 * 10240; time = 0.0416s; samplesPerSecond = 246307.8
12/20/2016 15:27:21:  Epoch[ 2 of 25]-Minibatch[ 161- 170, 15.49%]: CE.SM = 1.56776276 * 10240; Err = 0.47255859 * 10240; time = 0.0399s; samplesPerSecond = 256924.9
12/20/2016 15:27:21:  Epoch[ 2 of 25]-Minibatch[ 171- 180, 16.40%]: CE.SM = 1.55851746 * 10240; Err = 0.47148438 * 10240; time = 0.0419s; samplesPerSecond = 244467.3
12/20/2016 15:27:21:  Epoch[ 2 of 25]-Minibatch[ 181- 190, 17.31%]: CE.SM = 1.54776306 * 10240; Err = 0.46611328 * 10240; time = 0.0449s; samplesPerSecond = 228108.1
12/20/2016 15:27:21:  Epoch[ 2 of 25]-Minibatch[ 191- 200, 18.22%]: CE.SM = 1.56339111 * 10240; Err = 0.47031250 * 10240; time = 0.0449s; samplesPerSecond = 228021.7
12/20/2016 15:27:21:  Epoch[ 2 of 25]-Minibatch[ 201- 210, 19.13%]: CE.SM = 1.56317749 * 10240; Err = 0.46083984 * 10240; time = 0.0450s; samplesPerSecond = 227530.3
12/20/2016 15:27:21:  Epoch[ 2 of 25]-Minibatch[ 211- 220, 20.05%]: CE.SM = 1.53389282 * 10240; Err = 0.46044922 * 10240; time = 0.0409s; samplesPerSecond = 250587.3
12/20/2016 15:27:21:  Epoch[ 2 of 25]-Minibatch[ 221- 230, 20.96%]: CE.SM = 1.56185303 * 10240; Err = 0.47187500 * 10240; time = 0.0415s; samplesPerSecond = 246681.6
12/20/2016 15:27:21:  Epoch[ 2 of 25]-Minibatch[ 231- 240, 21.87%]: CE.SM = 1.55430603 * 10240; Err = 0.46884766 * 10240; time = 0.0410s; samplesPerSecond = 249969.5
12/20/2016 15:27:21:  Epoch[ 2 of 25]-Minibatch[ 241- 250, 22.78%]: CE.SM = 1.54589539 * 10240; Err = 0.46689453 * 10240; time = 0.0420s; samplesPerSecond = 243972.2
12/20/2016 15:27:21:  Epoch[ 2 of 25]-Minibatch[ 251- 260, 23.69%]: CE.SM = 1.54432373 * 10240; Err = 0.46699219 * 10240; time = 0.0415s; samplesPerSecond = 246699.4
12/20/2016 15:27:21:  Epoch[ 2 of 25]-Minibatch[ 261- 270, 24.60%]: CE.SM = 1.57476807 * 10240; Err = 0.47529297 * 10240; time = 0.0419s; samplesPerSecond = 244671.7
12/20/2016 15:27:21:  Epoch[ 2 of 25]-Minibatch[ 271- 280, 25.51%]: CE.SM = 1.55689087 * 10240; Err = 0.46552734 * 10240; time = 0.0425s; samplesPerSecond = 241054.6
12/20/2016 15:27:21:  Epoch[ 2 of 25]-Minibatch[ 281- 290, 26.42%]: CE.SM = 1.56448059 * 10240; Err = 0.46884766 * 10240; time = 0.0402s; samplesPerSecond = 254428.9
12/20/2016 15:27:21:  Epoch[ 2 of 25]-Minibatch[ 291- 300, 27.33%]: CE.SM = 1.52126770 * 10240; Err = 0.45820312 * 10240; time = 0.0411s; samplesPerSecond = 249445.8
12/20/2016 15:27:21:  Epoch[ 2 of 25]-Minibatch[ 301- 310, 28.25%]: CE.SM = 1.55354004 * 10240; Err = 0.46943359 * 10240; time = 0.0415s; samplesPerSecond = 247020.8
12/20/2016 15:27:21:  Epoch[ 2 of 25]-Minibatch[ 311- 320, 29.16%]: CE.SM = 1.55346375 * 10240; Err = 0.47167969 * 10240; time = 0.0400s; samplesPerSecond = 255699.6
12/20/2016 15:27:21:  Epoch[ 2 of 25]-Minibatch[ 321- 330, 30.07%]: CE.SM = 1.56849060 * 10240; Err = 0.46660156 * 10240; time = 0.0399s; samplesPerSecond = 256448.8
12/20/2016 15:27:21:  Epoch[ 2 of 25]-Minibatch[ 331- 340, 30.98%]: CE.SM = 1.54410400 * 10240; Err = 0.46181641 * 10240; time = 0.0411s; samplesPerSecond = 249445.8
12/20/2016 15:27:21:  Epoch[ 2 of 25]-Minibatch[ 341- 350, 31.89%]: CE.SM = 1.54474487 * 10240; Err = 0.46494141 * 10240; time = 0.0416s; samplesPerSecond = 246065.1
12/20/2016 15:27:21:  Epoch[ 2 of 25]-Minibatch[ 351- 360, 32.80%]: CE.SM = 1.57144775 * 10240; Err = 0.47363281 * 10240; time = 0.0412s; samplesPerSecond = 248567.8
12/20/2016 15:27:21:  Epoch[ 2 of 25]-Minibatch[ 361- 370, 33.71%]: CE.SM = 1.55744629 * 10240; Err = 0.46816406 * 10240; time = 0.0497s; samplesPerSecond = 205949.2
12/20/2016 15:27:21:  Epoch[ 2 of 25]-Minibatch[ 371- 380, 34.62%]: CE.SM = 1.54799194 * 10240; Err = 0.46835938 * 10240; time = 0.0438s; samplesPerSecond = 233811.3
12/20/2016 15:27:21:  Epoch[ 2 of 25]-Minibatch[ 381- 390, 35.54%]: CE.SM = 1.52231445 * 10240; Err = 0.45908203 * 10240; time = 0.0444s; samplesPerSecond = 230698.2
12/20/2016 15:27:21:  Epoch[ 2 of 25]-Minibatch[ 391- 400, 36.45%]: CE.SM = 1.55355835 * 10240; Err = 0.46152344 * 10240; time = 0.0414s; samplesPerSecond = 247259.4
12/20/2016 15:27:22:  Epoch[ 2 of 25]-Minibatch[ 401- 410, 37.36%]: CE.SM = 1.54296265 * 10240; Err = 0.46708984 * 10240; time = 0.0410s; samplesPerSecond = 249756.1
12/20/2016 15:27:22:  Epoch[ 2 of 25]-Minibatch[ 411- 420, 38.27%]: CE.SM = 1.54056396 * 10240; Err = 0.45791016 * 10240; time = 0.0418s; samplesPerSecond = 245187.2
12/20/2016 15:27:22:  Epoch[ 2 of 25]-Minibatch[ 421- 430, 39.18%]: CE.SM = 1.53822021 * 10240; Err = 0.46640625 * 10240; time = 0.0410s; samplesPerSecond = 249939.0
12/20/2016 15:27:22:  Epoch[ 2 of 25]-Minibatch[ 431- 440, 40.09%]: CE.SM = 1.54398804 * 10240; Err = 0.46767578 * 10240; time = 0.0403s; samplesPerSecond = 253987.2
12/20/2016 15:27:22:  Epoch[ 2 of 25]-Minibatch[ 441- 450, 41.00%]: CE.SM = 1.52210693 * 10240; Err = 0.45673828 * 10240; time = 0.0406s; samplesPerSecond = 251912.7
12/20/2016 15:27:22:  Epoch[ 2 of 25]-Minibatch[ 451- 460, 41.91%]: CE.SM = 1.51351929 * 10240; Err = 0.45781250 * 10240; time = 0.0401s; samplesPerSecond = 255565.5
12/20/2016 15:27:22:  Epoch[ 2 of 25]-Minibatch[ 461- 470, 42.82%]: CE.SM = 1.51707764 * 10240; Err = 0.46445313 * 10240; time = 0.0444s; samplesPerSecond = 230506.0
12/20/2016 15:27:22:  Epoch[ 2 of 25]-Minibatch[ 471- 480, 43.74%]: CE.SM = 1.50535889 * 10240; Err = 0.45996094 * 10240; time = 0.0414s; samplesPerSecond = 247152.0
12/20/2016 15:27:22:  Epoch[ 2 of 25]-Minibatch[ 481- 490, 44.65%]: CE.SM = 1.51300659 * 10240; Err = 0.46113281 * 10240; time = 0.0411s; samplesPerSecond = 248888.0
12/20/2016 15:27:22:  Epoch[ 2 of 25]-Minibatch[ 491- 500, 45.56%]: CE.SM = 1.55207520 * 10240; Err = 0.46875000 * 10240; time = 0.0400s; samplesPerSecond = 256076.8
12/20/2016 15:27:22:  Epoch[ 2 of 25]-Minibatch[ 501- 510, 46.47%]: CE.SM = 1.50676880 * 10240; Err = 0.45351562 * 10240; time = 0.0423s; samplesPerSecond = 242366.9
12/20/2016 15:27:22:  Epoch[ 2 of 25]-Minibatch[ 511- 520, 47.38%]: CE.SM = 1.55104980 * 10240; Err = 0.47412109 * 10240; time = 0.0472s; samplesPerSecond = 217156.2
12/20/2016 15:27:22:  Epoch[ 2 of 25]-Minibatch[ 521- 530, 48.29%]: CE.SM = 1.50769043 * 10240; Err = 0.46113281 * 10240; time = 0.0413s; samplesPerSecond = 247953.9
12/20/2016 15:27:22:  Epoch[ 2 of 25]-Minibatch[ 531- 540, 49.20%]: CE.SM = 1.50725708 * 10240; Err = 0.45986328 * 10240; time = 0.0416s; samplesPerSecond = 246065.1
12/20/2016 15:27:22:  Epoch[ 2 of 25]-Minibatch[ 541- 550, 50.11%]: CE.SM = 1.50804443 * 10240; Err = 0.44902344 * 10240; time = 0.0411s; samplesPerSecond = 249269.7
12/20/2016 15:27:22:  Epoch[ 2 of 25]-Minibatch[ 551- 560, 51.03%]: CE.SM = 1.51472778 * 10240; Err = 0.46181641 * 10240; time = 0.0408s; samplesPerSecond = 250753.0
12/20/2016 15:27:22:  Epoch[ 2 of 25]-Minibatch[ 561- 570, 51.94%]: CE.SM = 1.54004517 * 10240; Err = 0.46679688 * 10240; time = 0.0445s; samplesPerSecond = 230262.4
12/20/2016 15:27:22:  Epoch[ 2 of 25]-Minibatch[ 571- 580, 52.85%]: CE.SM = 1.50709229 * 10240; Err = 0.45937500 * 10240; time = 0.0406s; samplesPerSecond = 252465.5
12/20/2016 15:27:22:  Epoch[ 2 of 25]-Minibatch[ 581- 590, 53.76%]: CE.SM = 1.50729980 * 10240; Err = 0.46220703 * 10240; time = 0.0433s; samplesPerSecond = 236549.7
12/20/2016 15:27:22:  Epoch[ 2 of 25]-Minibatch[ 591- 600, 54.67%]: CE.SM = 1.52664795 * 10240; Err = 0.46367188 * 10240; time = 0.0444s; samplesPerSecond = 230485.3
12/20/2016 15:27:22:  Epoch[ 2 of 25]-Minibatch[ 601- 610, 55.58%]: CE.SM = 1.49658813 * 10240; Err = 0.45546875 * 10240; time = 0.0489s; samplesPerSecond = 209578.4
12/20/2016 15:27:22:  Epoch[ 2 of 25]-Minibatch[ 611- 620, 56.49%]: CE.SM = 1.49868164 * 10240; Err = 0.44970703 * 10240; time = 0.0417s; samplesPerSecond = 245428.2
12/20/2016 15:27:22:  Epoch[ 2 of 25]-Minibatch[ 621- 630, 57.40%]: CE.SM = 1.49858398 * 10240; Err = 0.45126953 * 10240; time = 0.0421s; samplesPerSecond = 243334.4
12/20/2016 15:27:23:  Epoch[ 2 of 25]-Minibatch[ 631- 640, 58.31%]: CE.SM = 1.51031494 * 10240; Err = 0.45615234 * 10240; time = 0.0479s; samplesPerSecond = 213595.9
12/20/2016 15:27:23:  Epoch[ 2 of 25]-Minibatch[ 641- 650, 59.23%]: CE.SM = 1.51919556 * 10240; Err = 0.46250000 * 10240; time = 0.0412s; samplesPerSecond = 248477.3
12/20/2016 15:27:23:  Epoch[ 2 of 25]-Minibatch[ 651- 660, 60.14%]: CE.SM = 1.52016602 * 10240; Err = 0.46103516 * 10240; time = 0.0421s; samplesPerSecond = 243374.9
12/20/2016 15:27:23:  Epoch[ 2 of 25]-Minibatch[ 661- 670, 61.05%]: CE.SM = 1.52784424 * 10240; Err = 0.46406250 * 10240; time = 0.0444s; samplesPerSecond = 230677.4
12/20/2016 15:27:23:  Epoch[ 2 of 25]-Minibatch[ 671- 680, 61.96%]: CE.SM = 1.49887695 * 10240; Err = 0.45615234 * 10240; time = 0.0409s; samplesPerSecond = 250630.2
12/20/2016 15:27:23:  Epoch[ 2 of 25]-Minibatch[ 681- 690, 62.87%]: CE.SM = 1.50270996 * 10240; Err = 0.45683594 * 10240; time = 0.0414s; samplesPerSecond = 247396.8
12/20/2016 15:27:23:  Epoch[ 2 of 25]-Minibatch[ 691- 700, 63.78%]: CE.SM = 1.52094727 * 10240; Err = 0.46181641 * 10240; time = 0.0409s; samplesPerSecond = 250409.6
12/20/2016 15:27:23:  Epoch[ 2 of 25]-Minibatch[ 701- 710, 64.69%]: CE.SM = 1.52066650 * 10240; Err = 0.46259766 * 10240; time = 0.0433s; samplesPerSecond = 236369.5
12/20/2016 15:27:23:  Epoch[ 2 of 25]-Minibatch[ 711- 720, 65.60%]: CE.SM = 1.49411621 * 10240; Err = 0.46220703 * 10240; time = 0.0414s; samplesPerSecond = 247295.2
12/20/2016 15:27:23:  Epoch[ 2 of 25]-Minibatch[ 721- 730, 66.51%]: CE.SM = 1.49122314 * 10240; Err = 0.45517578 * 10240; time = 0.0396s; samplesPerSecond = 258670.8
12/20/2016 15:27:23:  Epoch[ 2 of 25]-Minibatch[ 731- 740, 67.43%]: CE.SM = 1.47027588 * 10240; Err = 0.44853516 * 10240; time = 0.0405s; samplesPerSecond = 253102.0
12/20/2016 15:27:23:  Epoch[ 2 of 25]-Minibatch[ 741- 750, 68.34%]: CE.SM = 1.47718506 * 10240; Err = 0.44433594 * 10240; time = 0.0403s; samplesPerSecond = 253804.6
12/20/2016 15:27:23:  Epoch[ 2 of 25]-Minibatch[ 751- 760, 69.25%]: CE.SM = 1.48763428 * 10240; Err = 0.44775391 * 10240; time = 0.0406s; samplesPerSecond = 252415.7
12/20/2016 15:27:23:  Epoch[ 2 of 25]-Minibatch[ 761- 770, 70.16%]: CE.SM = 1.50186768 * 10240; Err = 0.45322266 * 10240; time = 0.0418s; samplesPerSecond = 244858.9
12/20/2016 15:27:23:  Epoch[ 2 of 25]-Minibatch[ 771- 780, 71.07%]: CE.SM = 1.50264893 * 10240; Err = 0.45458984 * 10240; time = 0.0406s; samplesPerSecond = 252509.1
12/20/2016 15:27:23:  Epoch[ 2 of 25]-Minibatch[ 781- 790, 71.98%]: CE.SM = 1.49951172 * 10240; Err = 0.45966797 * 10240; time = 0.0409s; samplesPerSecond = 250587.3
12/20/2016 15:27:23:  Epoch[ 2 of 25]-Minibatch[ 791- 800, 72.89%]: CE.SM = 1.52775879 * 10240; Err = 0.46230469 * 10240; time = 0.0406s; samplesPerSecond = 252142.2
12/20/2016 15:27:23:  Epoch[ 2 of 25]-Minibatch[ 801- 810, 73.80%]: CE.SM = 1.49316406 * 10240; Err = 0.45957031 * 10240; time = 0.0447s; samplesPerSecond = 229169.9
12/20/2016 15:27:23:  Epoch[ 2 of 25]-Minibatch[ 811- 820, 74.72%]: CE.SM = 1.46741943 * 10240; Err = 0.44882813 * 10240; time = 0.0396s; samplesPerSecond = 258742.7
12/20/2016 15:27:23:  Epoch[ 2 of 25]-Minibatch[ 821- 830, 75.63%]: CE.SM = 1.48822021 * 10240; Err = 0.45615234 * 10240; time = 0.0396s; samplesPerSecond = 258834.2
12/20/2016 15:27:23:  Epoch[ 2 of 25]-Minibatch[ 831- 840, 76.54%]: CE.SM = 1.48833008 * 10240; Err = 0.45214844 * 10240; time = 0.0417s; samplesPerSecond = 245469.4
12/20/2016 15:27:23:  Epoch[ 2 of 25]-Minibatch[ 841- 850, 77.45%]: CE.SM = 1.47918701 * 10240; Err = 0.45136719 * 10240; time = 0.0470s; samplesPerSecond = 217751.9
12/20/2016 15:27:23:  Epoch[ 2 of 25]-Minibatch[ 851- 860, 78.36%]: CE.SM = 1.51467285 * 10240; Err = 0.45673828 * 10240; time = 0.0422s; samplesPerSecond = 242803.6
12/20/2016 15:27:23:  Epoch[ 2 of 25]-Minibatch[ 861- 870, 79.27%]: CE.SM = 1.47113037 * 10240; Err = 0.44550781 * 10240; time = 0.0428s; samplesPerSecond = 239129.4
12/20/2016 15:27:24:  Epoch[ 2 of 25]-Minibatch[ 871- 880, 80.18%]: CE.SM = 1.49293213 * 10240; Err = 0.45429687 * 10240; time = 0.0417s; samplesPerSecond = 245705.0
12/20/2016 15:27:24:  Epoch[ 2 of 25]-Minibatch[ 881- 890, 81.09%]: CE.SM = 1.51804199 * 10240; Err = 0.45839844 * 10240; time = 0.0430s; samplesPerSecond = 238200.5
12/20/2016 15:27:24:  Epoch[ 2 of 25]-Minibatch[ 891- 900, 82.00%]: CE.SM = 1.48527832 * 10240; Err = 0.45585938 * 10240; time = 0.0451s; samplesPerSecond = 226965.4
12/20/2016 15:27:24:  Epoch[ 2 of 25]-Minibatch[ 901- 910, 82.92%]: CE.SM = 1.49360352 * 10240; Err = 0.45234375 * 10240; time = 0.0419s; samplesPerSecond = 244123.4
12/20/2016 15:27:24:  Epoch[ 2 of 25]-Minibatch[ 911- 920, 83.83%]: CE.SM = 1.46337891 * 10240; Err = 0.44648437 * 10240; time = 0.0428s; samplesPerSecond = 239185.3
12/20/2016 15:27:24:  Epoch[ 2 of 25]-Minibatch[ 921- 930, 84.74%]: CE.SM = 1.44265137 * 10240; Err = 0.44814453 * 10240; time = 0.0440s; samplesPerSecond = 232939.0
12/20/2016 15:27:24:  Epoch[ 2 of 25]-Minibatch[ 931- 940, 85.65%]: CE.SM = 1.47514648 * 10240; Err = 0.44804688 * 10240; time = 0.0428s; samplesPerSecond = 239280.3
12/20/2016 15:27:24:  Epoch[ 2 of 25]-Minibatch[ 941- 950, 86.56%]: CE.SM = 1.49561768 * 10240; Err = 0.45625000 * 10240; time = 0.0456s; samplesPerSecond = 224389.2
12/20/2016 15:27:24:  Epoch[ 2 of 25]-Minibatch[ 951- 960, 87.47%]: CE.SM = 1.46875000 * 10240; Err = 0.45195313 * 10240; time = 0.0456s; samplesPerSecond = 224659.9
12/20/2016 15:27:24:  Epoch[ 2 of 25]-Minibatch[ 961- 970, 88.38%]: CE.SM = 1.48920898 * 10240; Err = 0.45576172 * 10240; time = 0.0453s; samplesPerSecond = 226228.3
12/20/2016 15:27:24:  Epoch[ 2 of 25]-Minibatch[ 971- 980, 89.29%]: CE.SM = 1.44343262 * 10240; Err = 0.43662109 * 10240; time = 0.0423s; samplesPerSecond = 241948.8
12/20/2016 15:27:24:  Epoch[ 2 of 25]-Minibatch[ 981- 990, 90.21%]: CE.SM = 1.47719727 * 10240; Err = 0.45097656 * 10240; time = 0.0408s; samplesPerSecond = 251097.3
12/20/2016 15:27:24:  Epoch[ 2 of 25]-Minibatch[ 991-1000, 91.12%]: CE.SM = 1.45489502 * 10240; Err = 0.44287109 * 10240; time = 0.0403s; samplesPerSecond = 254271.0
12/20/2016 15:27:24:  Epoch[ 2 of 25]-Minibatch[1001-1010, 92.03%]: CE.SM = 1.46821289 * 10240; Err = 0.44140625 * 10240; time = 0.0409s; samplesPerSecond = 250226.0
12/20/2016 15:27:24:  Epoch[ 2 of 25]-Minibatch[1011-1020, 92.94%]: CE.SM = 1.46142578 * 10240; Err = 0.44316406 * 10240; time = 0.0412s; samplesPerSecond = 248803.4
12/20/2016 15:27:24:  Epoch[ 2 of 25]-Minibatch[1021-1030, 93.85%]: CE.SM = 1.47268066 * 10240; Err = 0.45302734 * 10240; time = 0.0418s; samplesPerSecond = 245005.4
12/20/2016 15:27:24:  Epoch[ 2 of 25]-Minibatch[1031-1040, 94.76%]: CE.SM = 1.48309326 * 10240; Err = 0.45761719 * 10240; time = 0.0396s; samplesPerSecond = 258298.9
12/20/2016 15:27:24:  Epoch[ 2 of 25]-Minibatch[1041-1050, 95.67%]: CE.SM = 1.46152344 * 10240; Err = 0.44541016 * 10240; time = 0.0409s; samplesPerSecond = 250268.8
12/20/2016 15:27:24:  Epoch[ 2 of 25]-Minibatch[1051-1060, 96.58%]: CE.SM = 1.47220459 * 10240; Err = 0.44863281 * 10240; time = 0.0402s; samplesPerSecond = 254587.0
12/20/2016 15:27:24:  Epoch[ 2 of 25]-Minibatch[1061-1070, 97.49%]: CE.SM = 1.45733643 * 10240; Err = 0.44941406 * 10240; time = 0.0397s; samplesPerSecond = 257726.8
12/20/2016 15:27:24:  Epoch[ 2 of 25]-Minibatch[1071-1080, 98.41%]: CE.SM = 1.49428711 * 10240; Err = 0.45419922 * 10240; time = 0.0466s; samplesPerSecond = 219596.4
12/20/2016 15:27:24:  Epoch[ 2 of 25]-Minibatch[1081-1090, 99.32%]: CE.SM = 1.48432617 * 10240; Err = 0.45126953 * 10240; time = 0.0440s; samplesPerSecond = 232552.9
12/20/2016 15:27:24: Finished Epoch[ 2 of 25]: [Training] CE.SM = 1.52137158 * 1124823; Err = 0.46046089 * 1124823; totalSamplesSeen = 2249646; learningRatePerSample = 0.003125; epochTime=4.85236s
12/20/2016 15:27:24: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn.2'

12/20/2016 15:27:25: Starting Epoch 3: learning rate per sample = 0.003125  effective momentum = 0.900000  momentum as time constant = 9719.0 samples

12/20/2016 15:27:25: Starting minibatch loop.
12/20/2016 15:27:25:  Epoch[ 3 of 25]-Minibatch[   1-  10, 0.91%]: CE.SM = 1.43683815 * 10240; Err = 0.44082031 * 10240; time = 0.0509s; samplesPerSecond = 201198.5
12/20/2016 15:27:25:  Epoch[ 3 of 25]-Minibatch[  11-  20, 1.82%]: CE.SM = 1.44095058 * 10240; Err = 0.43964844 * 10240; time = 0.0428s; samplesPerSecond = 239436.9
12/20/2016 15:27:25:  Epoch[ 3 of 25]-Minibatch[  21-  30, 2.73%]: CE.SM = 1.42724476 * 10240; Err = 0.44091797 * 10240; time = 0.0401s; samplesPerSecond = 255648.5
12/20/2016 15:27:25:  Epoch[ 3 of 25]-Minibatch[  31-  40, 3.64%]: CE.SM = 1.44737129 * 10240; Err = 0.44423828 * 10240; time = 0.0398s; samplesPerSecond = 257267.0
12/20/2016 15:27:25:  Epoch[ 3 of 25]-Minibatch[  41-  50, 4.56%]: CE.SM = 1.45663681 * 10240; Err = 0.44306641 * 10240; time = 0.0410s; samplesPerSecond = 249579.6
12/20/2016 15:27:25:  Epoch[ 3 of 25]-Minibatch[  51-  60, 5.47%]: CE.SM = 1.43562164 * 10240; Err = 0.43994141 * 10240; time = 0.0412s; samplesPerSecond = 248501.5
12/20/2016 15:27:25:  Epoch[ 3 of 25]-Minibatch[  61-  70, 6.38%]: CE.SM = 1.42691727 * 10240; Err = 0.44042969 * 10240; time = 0.0410s; samplesPerSecond = 249914.6
12/20/2016 15:27:25:  Epoch[ 3 of 25]-Minibatch[  71-  80, 7.29%]: CE.SM = 1.43590698 * 10240; Err = 0.43935547 * 10240; time = 0.0402s; samplesPerSecond = 254840.5
12/20/2016 15:27:25:  Epoch[ 3 of 25]-Minibatch[  81-  90, 8.20%]: CE.SM = 1.44843445 * 10240; Err = 0.44150391 * 10240; time = 0.0401s; samplesPerSecond = 255253.4
12/20/2016 15:27:25:  Epoch[ 3 of 25]-Minibatch[  91- 100, 9.11%]: CE.SM = 1.45344086 * 10240; Err = 0.44296875 * 10240; time = 0.0431s; samplesPerSecond = 237724.9
12/20/2016 15:27:25:  Epoch[ 3 of 25]-Minibatch[ 101- 110, 10.02%]: CE.SM = 1.46259460 * 10240; Err = 0.45058594 * 10240; time = 0.0475s; samplesPerSecond = 215633.4
12/20/2016 15:27:25:  Epoch[ 3 of 25]-Minibatch[ 111- 120, 10.93%]: CE.SM = 1.46175385 * 10240; Err = 0.44658203 * 10240; time = 0.0423s; samplesPerSecond = 242028.9
12/20/2016 15:27:25:  Epoch[ 3 of 25]-Minibatch[ 121- 130, 11.85%]: CE.SM = 1.44047699 * 10240; Err = 0.44306641 * 10240; time = 0.0445s; samplesPerSecond = 230236.5
12/20/2016 15:27:25:  Epoch[ 3 of 25]-Minibatch[ 131- 140, 12.76%]: CE.SM = 1.45618896 * 10240; Err = 0.45234375 * 10240; time = 0.0439s; samplesPerSecond = 233023.8
12/20/2016 15:27:25:  Epoch[ 3 of 25]-Minibatch[ 141- 150, 13.67%]: CE.SM = 1.39315948 * 10240; Err = 0.43144531 * 10240; time = 0.0442s; samplesPerSecond = 231852.6
12/20/2016 15:27:25:  Epoch[ 3 of 25]-Minibatch[ 151- 160, 14.58%]: CE.SM = 1.44313507 * 10240; Err = 0.43740234 * 10240; time = 0.0415s; samplesPerSecond = 246913.6
12/20/2016 15:27:25:  Epoch[ 3 of 25]-Minibatch[ 161- 170, 15.49%]: CE.SM = 1.43794556 * 10240; Err = 0.44472656 * 10240; time = 0.0442s; samplesPerSecond = 231721.4
12/20/2016 15:27:25:  Epoch[ 3 of 25]-Minibatch[ 171- 180, 16.40%]: CE.SM = 1.45457764 * 10240; Err = 0.45048828 * 10240; time = 0.0475s; samplesPerSecond = 215728.8
12/20/2016 15:27:26:  Epoch[ 3 of 25]-Minibatch[ 181- 190, 17.31%]: CE.SM = 1.41578979 * 10240; Err = 0.43466797 * 10240; time = 0.0416s; samplesPerSecond = 245923.3
12/20/2016 15:27:26:  Epoch[ 3 of 25]-Minibatch[ 191- 200, 18.22%]: CE.SM = 1.39819641 * 10240; Err = 0.43017578 * 10240; time = 0.0411s; samplesPerSecond = 249372.9
12/20/2016 15:27:26:  Epoch[ 3 of 25]-Minibatch[ 201- 210, 19.13%]: CE.SM = 1.44436340 * 10240; Err = 0.44296875 * 10240; time = 0.0475s; samplesPerSecond = 215737.9
12/20/2016 15:27:26:  Epoch[ 3 of 25]-Minibatch[ 211- 220, 20.05%]: CE.SM = 1.46193542 * 10240; Err = 0.44882813 * 10240; time = 0.0454s; samplesPerSecond = 225704.8
12/20/2016 15:27:26:  Epoch[ 3 of 25]-Minibatch[ 221- 230, 20.96%]: CE.SM = 1.40391541 * 10240; Err = 0.43847656 * 10240; time = 0.0417s; samplesPerSecond = 245516.4
12/20/2016 15:27:26:  Epoch[ 3 of 25]-Minibatch[ 231- 240, 21.87%]: CE.SM = 1.42628479 * 10240; Err = 0.44082031 * 10240; time = 0.0416s; samplesPerSecond = 246313.7
12/20/2016 15:27:26:  Epoch[ 3 of 25]-Minibatch[ 241- 250, 22.78%]: CE.SM = 1.41159668 * 10240; Err = 0.43623047 * 10240; time = 0.0409s; samplesPerSecond = 250122.1
12/20/2016 15:27:26:  Epoch[ 3 of 25]-Minibatch[ 251- 260, 23.69%]: CE.SM = 1.41656799 * 10240; Err = 0.43359375 * 10240; time = 0.0420s; samplesPerSecond = 243931.5
12/20/2016 15:27:26:  Epoch[ 3 of 25]-Minibatch[ 261- 270, 24.60%]: CE.SM = 1.42076111 * 10240; Err = 0.44208984 * 10240; time = 0.0411s; samplesPerSecond = 249300.1
12/20/2016 15:27:26:  Epoch[ 3 of 25]-Minibatch[ 271- 280, 25.51%]: CE.SM = 1.42387390 * 10240; Err = 0.43554688 * 10240; time = 0.0418s; samplesPerSecond = 244981.9
12/20/2016 15:27:26:  Epoch[ 3 of 25]-Minibatch[ 281- 290, 26.42%]: CE.SM = 1.40029602 * 10240; Err = 0.42587891 * 10240; time = 0.0410s; samplesPerSecond = 249524.8
12/20/2016 15:27:26:  Epoch[ 3 of 25]-Minibatch[ 291- 300, 27.33%]: CE.SM = 1.40063782 * 10240; Err = 0.43027344 * 10240; time = 0.0427s; samplesPerSecond = 239717.2
12/20/2016 15:27:26:  Epoch[ 3 of 25]-Minibatch[ 301- 310, 28.25%]: CE.SM = 1.41532898 * 10240; Err = 0.44082031 * 10240; time = 0.0404s; samplesPerSecond = 253339.9
12/20/2016 15:27:26:  Epoch[ 3 of 25]-Minibatch[ 311- 320, 29.16%]: CE.SM = 1.40893250 * 10240; Err = 0.43759766 * 10240; time = 0.0445s; samplesPerSecond = 230334.9
12/20/2016 15:27:26:  Epoch[ 3 of 25]-Minibatch[ 321- 330, 30.07%]: CE.SM = 1.39343262 * 10240; Err = 0.43125000 * 10240; time = 0.0407s; samplesPerSecond = 251838.4
12/20/2016 15:27:26:  Epoch[ 3 of 25]-Minibatch[ 331- 340, 30.98%]: CE.SM = 1.42695618 * 10240; Err = 0.44101563 * 10240; time = 0.0421s; samplesPerSecond = 243322.9
12/20/2016 15:27:26:  Epoch[ 3 of 25]-Minibatch[ 341- 350, 31.89%]: CE.SM = 1.43694153 * 10240; Err = 0.44111328 * 10240; time = 0.0443s; samplesPerSecond = 230911.5
12/20/2016 15:27:26:  Epoch[ 3 of 25]-Minibatch[ 351- 360, 32.80%]: CE.SM = 1.42338562 * 10240; Err = 0.44228516 * 10240; time = 0.0417s; samplesPerSecond = 245551.8
12/20/2016 15:27:26:  Epoch[ 3 of 25]-Minibatch[ 361- 370, 33.71%]: CE.SM = 1.44453125 * 10240; Err = 0.44189453 * 10240; time = 0.0421s; samplesPerSecond = 243247.7
12/20/2016 15:27:26:  Epoch[ 3 of 25]-Minibatch[ 371- 380, 34.62%]: CE.SM = 1.43007813 * 10240; Err = 0.43623047 * 10240; time = 0.0425s; samplesPerSecond = 240714.6
12/20/2016 15:27:26:  Epoch[ 3 of 25]-Minibatch[ 381- 390, 35.54%]: CE.SM = 1.45139771 * 10240; Err = 0.43964844 * 10240; time = 0.0410s; samplesPerSecond = 249725.6
12/20/2016 15:27:26:  Epoch[ 3 of 25]-Minibatch[ 391- 400, 36.45%]: CE.SM = 1.41055298 * 10240; Err = 0.43505859 * 10240; time = 0.0417s; samplesPerSecond = 245469.4
12/20/2016 15:27:26:  Epoch[ 3 of 25]-Minibatch[ 401- 410, 37.36%]: CE.SM = 1.43289795 * 10240; Err = 0.43554688 * 10240; time = 0.0410s; samplesPerSecond = 249524.8
12/20/2016 15:27:26:  Epoch[ 3 of 25]-Minibatch[ 411- 420, 38.27%]: CE.SM = 1.42119141 * 10240; Err = 0.43447266 * 10240; time = 0.0414s; samplesPerSecond = 247050.6
12/20/2016 15:27:27:  Epoch[ 3 of 25]-Minibatch[ 421- 430, 39.18%]: CE.SM = 1.42234497 * 10240; Err = 0.43691406 * 10240; time = 0.0397s; samplesPerSecond = 258097.0
12/20/2016 15:27:27:  Epoch[ 3 of 25]-Minibatch[ 431- 440, 40.09%]: CE.SM = 1.42329712 * 10240; Err = 0.44130859 * 10240; time = 0.0412s; samplesPerSecond = 248531.6
12/20/2016 15:27:27:  Epoch[ 3 of 25]-Minibatch[ 441- 450, 41.00%]: CE.SM = 1.43639526 * 10240; Err = 0.44384766 * 10240; time = 0.0417s; samplesPerSecond = 245740.3
12/20/2016 15:27:27:  Epoch[ 3 of 25]-Minibatch[ 451- 460, 41.91%]: CE.SM = 1.40765991 * 10240; Err = 0.42548828 * 10240; time = 0.0440s; samplesPerSecond = 232563.4
12/20/2016 15:27:27:  Epoch[ 3 of 25]-Minibatch[ 461- 470, 42.82%]: CE.SM = 1.40850830 * 10240; Err = 0.43212891 * 10240; time = 0.0467s; samplesPerSecond = 219342.4
12/20/2016 15:27:27:  Epoch[ 3 of 25]-Minibatch[ 471- 480, 43.74%]: CE.SM = 1.42274170 * 10240; Err = 0.43378906 * 10240; time = 0.0453s; samplesPerSecond = 225948.8
12/20/2016 15:27:27:  Epoch[ 3 of 25]-Minibatch[ 481- 490, 44.65%]: CE.SM = 1.39224243 * 10240; Err = 0.42714844 * 10240; time = 0.0421s; samplesPerSecond = 243328.7
12/20/2016 15:27:27:  Epoch[ 3 of 25]-Minibatch[ 491- 500, 45.56%]: CE.SM = 1.42710571 * 10240; Err = 0.43935547 * 10240; time = 0.0433s; samplesPerSecond = 236713.7
12/20/2016 15:27:27:  Epoch[ 3 of 25]-Minibatch[ 501- 510, 46.47%]: CE.SM = 1.41005859 * 10240; Err = 0.43378906 * 10240; time = 0.0418s; samplesPerSecond = 245140.3
12/20/2016 15:27:27:  Epoch[ 3 of 25]-Minibatch[ 511- 520, 47.38%]: CE.SM = 1.38491211 * 10240; Err = 0.42822266 * 10240; time = 0.0418s; samplesPerSecond = 244917.5
12/20/2016 15:27:27:  Epoch[ 3 of 25]-Minibatch[ 521- 530, 48.29%]: CE.SM = 1.41060181 * 10240; Err = 0.43564453 * 10240; time = 0.0415s; samplesPerSecond = 247002.9
12/20/2016 15:27:27:  Epoch[ 3 of 25]-Minibatch[ 531- 540, 49.20%]: CE.SM = 1.40348511 * 10240; Err = 0.43281250 * 10240; time = 0.0413s; samplesPerSecond = 248176.2
12/20/2016 15:27:27:  Epoch[ 3 of 25]-Minibatch[ 541- 550, 50.11%]: CE.SM = 1.40844116 * 10240; Err = 0.43681641 * 10240; time = 0.0415s; samplesPerSecond = 247044.6
12/20/2016 15:27:27:  Epoch[ 3 of 25]-Minibatch[ 551- 560, 51.03%]: CE.SM = 1.41932983 * 10240; Err = 0.43779297 * 10240; time = 0.0415s; samplesPerSecond = 246883.8
12/20/2016 15:27:27:  Epoch[ 3 of 25]-Minibatch[ 561- 570, 51.94%]: CE.SM = 1.40725708 * 10240; Err = 0.43261719 * 10240; time = 0.0416s; samplesPerSecond = 246290.0
12/20/2016 15:27:27:  Epoch[ 3 of 25]-Minibatch[ 571- 580, 52.85%]: CE.SM = 1.38831787 * 10240; Err = 0.43164062 * 10240; time = 0.0406s; samplesPerSecond = 252235.4
12/20/2016 15:27:27:  Epoch[ 3 of 25]-Minibatch[ 581- 590, 53.76%]: CE.SM = 1.40946045 * 10240; Err = 0.43232422 * 10240; time = 0.0411s; samplesPerSecond = 249391.1
12/20/2016 15:27:27:  Epoch[ 3 of 25]-Minibatch[ 591- 600, 54.67%]: CE.SM = 1.40541382 * 10240; Err = 0.43242188 * 10240; time = 0.0420s; samplesPerSecond = 243978.0
12/20/2016 15:27:27:  Epoch[ 3 of 25]-Minibatch[ 601- 610, 55.58%]: CE.SM = 1.41571655 * 10240; Err = 0.44199219 * 10240; time = 0.0420s; samplesPerSecond = 243728.3
12/20/2016 15:27:27:  Epoch[ 3 of 25]-Minibatch[ 611- 620, 56.49%]: CE.SM = 1.44235229 * 10240; Err = 0.44433594 * 10240; time = 0.0407s; samplesPerSecond = 251535.2
12/20/2016 15:27:27:  Epoch[ 3 of 25]-Minibatch[ 621- 630, 57.40%]: CE.SM = 1.40316162 * 10240; Err = 0.43535156 * 10240; time = 0.0420s; samplesPerSecond = 243792.1
12/20/2016 15:27:27:  Epoch[ 3 of 25]-Minibatch[ 631- 640, 58.31%]: CE.SM = 1.42893066 * 10240; Err = 0.43212891 * 10240; time = 0.0410s; samplesPerSecond = 249579.6
12/20/2016 15:27:27:  Epoch[ 3 of 25]-Minibatch[ 641- 650, 59.23%]: CE.SM = 1.40125732 * 10240; Err = 0.43603516 * 10240; time = 0.0417s; samplesPerSecond = 245722.7
12/20/2016 15:27:28:  Epoch[ 3 of 25]-Minibatch[ 651- 660, 60.14%]: CE.SM = 1.40809937 * 10240; Err = 0.44023438 * 10240; time = 0.0403s; samplesPerSecond = 254050.2
12/20/2016 15:27:28:  Epoch[ 3 of 25]-Minibatch[ 661- 670, 61.05%]: CE.SM = 1.41096802 * 10240; Err = 0.43183594 * 10240; time = 0.0412s; samplesPerSecond = 248797.3
12/20/2016 15:27:28:  Epoch[ 3 of 25]-Minibatch[ 671- 680, 61.96%]: CE.SM = 1.41934204 * 10240; Err = 0.43505859 * 10240; time = 0.0409s; samplesPerSecond = 250409.6
12/20/2016 15:27:28:  Epoch[ 3 of 25]-Minibatch[ 681- 690, 62.87%]: CE.SM = 1.42999878 * 10240; Err = 0.43652344 * 10240; time = 0.0400s; samplesPerSecond = 256179.3
12/20/2016 15:27:28:  Epoch[ 3 of 25]-Minibatch[ 691- 700, 63.78%]: CE.SM = 1.40348511 * 10240; Err = 0.43349609 * 10240; time = 0.0405s; samplesPerSecond = 252939.4
12/20/2016 15:27:28:  Epoch[ 3 of 25]-Minibatch[ 701- 710, 64.69%]: CE.SM = 1.41678467 * 10240; Err = 0.43525391 * 10240; time = 0.0418s; samplesPerSecond = 244747.7
12/20/2016 15:27:28:  Epoch[ 3 of 25]-Minibatch[ 711- 720, 65.60%]: CE.SM = 1.40165405 * 10240; Err = 0.42910156 * 10240; time = 0.0413s; samplesPerSecond = 247923.9
12/20/2016 15:27:28:  Epoch[ 3 of 25]-Minibatch[ 721- 730, 66.51%]: CE.SM = 1.40172119 * 10240; Err = 0.43554688 * 10240; time = 0.0432s; samplesPerSecond = 237300.7
12/20/2016 15:27:28:  Epoch[ 3 of 25]-Minibatch[ 731- 740, 67.43%]: CE.SM = 1.41429443 * 10240; Err = 0.42910156 * 10240; time = 0.0427s; samplesPerSecond = 239621.8
12/20/2016 15:27:28:  Epoch[ 3 of 25]-Minibatch[ 741- 750, 68.34%]: CE.SM = 1.41450195 * 10240; Err = 0.43662109 * 10240; time = 0.0429s; samplesPerSecond = 238633.5
12/20/2016 15:27:28:  Epoch[ 3 of 25]-Minibatch[ 751- 760, 69.25%]: CE.SM = 1.39815674 * 10240; Err = 0.42587891 * 10240; time = 0.0462s; samplesPerSecond = 221525.1
12/20/2016 15:27:28:  Epoch[ 3 of 25]-Minibatch[ 761- 770, 70.16%]: CE.SM = 1.41639404 * 10240; Err = 0.43583984 * 10240; time = 0.0445s; samplesPerSecond = 230329.7
12/20/2016 15:27:28:  Epoch[ 3 of 25]-Minibatch[ 771- 780, 71.07%]: CE.SM = 1.37977295 * 10240; Err = 0.42695312 * 10240; time = 0.0450s; samplesPerSecond = 227318.1
12/20/2016 15:27:28:  Epoch[ 3 of 25]-Minibatch[ 781- 790, 71.98%]: CE.SM = 1.40030518 * 10240; Err = 0.43291016 * 10240; time = 0.0458s; samplesPerSecond = 223756.7
12/20/2016 15:27:28:  Epoch[ 3 of 25]-Minibatch[ 791- 800, 72.89%]: CE.SM = 1.39813232 * 10240; Err = 0.42783203 * 10240; time = 0.0415s; samplesPerSecond = 246782.7
12/20/2016 15:27:28:  Epoch[ 3 of 25]-Minibatch[ 801- 810, 73.80%]: CE.SM = 1.38756104 * 10240; Err = 0.42880859 * 10240; time = 0.0410s; samplesPerSecond = 250036.6
12/20/2016 15:27:28:  Epoch[ 3 of 25]-Minibatch[ 811- 820, 74.72%]: CE.SM = 1.40822754 * 10240; Err = 0.43740234 * 10240; time = 0.0410s; samplesPerSecond = 249616.1
12/20/2016 15:27:28:  Epoch[ 3 of 25]-Minibatch[ 821- 830, 75.63%]: CE.SM = 1.37075195 * 10240; Err = 0.42382812 * 10240; time = 0.0415s; samplesPerSecond = 246473.8
12/20/2016 15:27:28:  Epoch[ 3 of 25]-Minibatch[ 831- 840, 76.54%]: CE.SM = 1.40799561 * 10240; Err = 0.43046875 * 10240; time = 0.0417s; samplesPerSecond = 245834.7
12/20/2016 15:27:28:  Epoch[ 3 of 25]-Minibatch[ 841- 850, 77.45%]: CE.SM = 1.40078125 * 10240; Err = 0.43437500 * 10240; time = 0.0397s; samplesPerSecond = 257785.2
12/20/2016 15:27:28:  Epoch[ 3 of 25]-Minibatch[ 851- 860, 78.36%]: CE.SM = 1.38524170 * 10240; Err = 0.42998047 * 10240; time = 0.0420s; samplesPerSecond = 243797.9
12/20/2016 15:27:28:  Epoch[ 3 of 25]-Minibatch[ 861- 870, 79.27%]: CE.SM = 1.38395996 * 10240; Err = 0.42871094 * 10240; time = 0.0448s; samplesPerSecond = 228494.9
12/20/2016 15:27:28:  Epoch[ 3 of 25]-Minibatch[ 871- 880, 80.18%]: CE.SM = 1.36800537 * 10240; Err = 0.42441406 * 10240; time = 0.0459s; samplesPerSecond = 223147.2
12/20/2016 15:27:28:  Epoch[ 3 of 25]-Minibatch[ 881- 890, 81.09%]: CE.SM = 1.37120361 * 10240; Err = 0.42734375 * 10240; time = 0.0441s; samplesPerSecond = 232236.4
12/20/2016 15:27:29:  Epoch[ 3 of 25]-Minibatch[ 891- 900, 82.00%]: CE.SM = 1.37761230 * 10240; Err = 0.42421875 * 10240; time = 0.0471s; samplesPerSecond = 217613.1
12/20/2016 15:27:29:  Epoch[ 3 of 25]-Minibatch[ 901- 910, 82.92%]: CE.SM = 1.38947754 * 10240; Err = 0.42880859 * 10240; time = 0.0411s; samplesPerSecond = 249318.3
12/20/2016 15:27:29:  Epoch[ 3 of 25]-Minibatch[ 911- 920, 83.83%]: CE.SM = 1.37745361 * 10240; Err = 0.42519531 * 10240; time = 0.0475s; samplesPerSecond = 215397.6
12/20/2016 15:27:29:  Epoch[ 3 of 25]-Minibatch[ 921- 930, 84.74%]: CE.SM = 1.40581055 * 10240; Err = 0.43388672 * 10240; time = 0.0402s; samplesPerSecond = 254625.0
12/20/2016 15:27:29:  Epoch[ 3 of 25]-Minibatch[ 931- 940, 85.65%]: CE.SM = 1.39715576 * 10240; Err = 0.43251953 * 10240; time = 0.0459s; samplesPerSecond = 223225.0
12/20/2016 15:27:29:  Epoch[ 3 of 25]-Minibatch[ 941- 950, 86.56%]: CE.SM = 1.40847168 * 10240; Err = 0.43281250 * 10240; time = 0.0438s; samplesPerSecond = 233944.8
12/20/2016 15:27:29:  Epoch[ 3 of 25]-Minibatch[ 951- 960, 87.47%]: CE.SM = 1.41075439 * 10240; Err = 0.43710938 * 10240; time = 0.0413s; samplesPerSecond = 248044.0
12/20/2016 15:27:29:  Epoch[ 3 of 25]-Minibatch[ 961- 970, 88.38%]: CE.SM = 1.40965576 * 10240; Err = 0.43427734 * 10240; time = 0.0425s; samplesPerSecond = 241122.7
12/20/2016 15:27:29:  Epoch[ 3 of 25]-Minibatch[ 971- 980, 89.29%]: CE.SM = 1.42269287 * 10240; Err = 0.43876953 * 10240; time = 0.0523s; samplesPerSecond = 195696.2
12/20/2016 15:27:29:  Epoch[ 3 of 25]-Minibatch[ 981- 990, 90.21%]: CE.SM = 1.39174805 * 10240; Err = 0.42363281 * 10240; time = 0.0426s; samplesPerSecond = 240172.6
12/20/2016 15:27:29:  Epoch[ 3 of 25]-Minibatch[ 991-1000, 91.12%]: CE.SM = 1.40589600 * 10240; Err = 0.42802734 * 10240; time = 0.0415s; samplesPerSecond = 246628.1
12/20/2016 15:27:29:  Epoch[ 3 of 25]-Minibatch[1001-1010, 92.03%]: CE.SM = 1.38231201 * 10240; Err = 0.42089844 * 10240; time = 0.0404s; samplesPerSecond = 253415.2
12/20/2016 15:27:29:  Epoch[ 3 of 25]-Minibatch[1011-1020, 92.94%]: CE.SM = 1.39470215 * 10240; Err = 0.42226562 * 10240; time = 0.0408s; samplesPerSecond = 251251.3
12/20/2016 15:27:29:  Epoch[ 3 of 25]-Minibatch[1021-1030, 93.85%]: CE.SM = 1.37530518 * 10240; Err = 0.42812500 * 10240; time = 0.0406s; samplesPerSecond = 251968.5
12/20/2016 15:27:29:  Epoch[ 3 of 25]-Minibatch[1031-1040, 94.76%]: CE.SM = 1.38004150 * 10240; Err = 0.42275391 * 10240; time = 0.0406s; samplesPerSecond = 252223.0
12/20/2016 15:27:29:  Epoch[ 3 of 25]-Minibatch[1041-1050, 95.67%]: CE.SM = 1.39342041 * 10240; Err = 0.43095703 * 10240; time = 0.0410s; samplesPerSecond = 249603.9
12/20/2016 15:27:29:  Epoch[ 3 of 25]-Minibatch[1051-1060, 96.58%]: CE.SM = 1.39744873 * 10240; Err = 0.42763672 * 10240; time = 0.0404s; samplesPerSecond = 253741.7
12/20/2016 15:27:29:  Epoch[ 3 of 25]-Minibatch[1061-1070, 97.49%]: CE.SM = 1.37592773 * 10240; Err = 0.42812500 * 10240; time = 0.0389s; samplesPerSecond = 263333.8
12/20/2016 15:27:29:  Epoch[ 3 of 25]-Minibatch[1071-1080, 98.41%]: CE.SM = 1.37415771 * 10240; Err = 0.42021484 * 10240; time = 0.0398s; samplesPerSecond = 257260.6
12/20/2016 15:27:29:  Epoch[ 3 of 25]-Minibatch[1081-1090, 99.32%]: CE.SM = 1.34824219 * 10240; Err = 0.41894531 * 10240; time = 0.0398s; samplesPerSecond = 257002.3
12/20/2016 15:27:29: Finished Epoch[ 3 of 25]: [Training] CE.SM = 1.41227942 * 1124823; Err = 0.43463549 * 1124823; totalSamplesSeen = 3374469; learningRatePerSample = 0.003125; epochTime=4.84743s
12/20/2016 15:27:29: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn.3'

12/20/2016 15:27:29: Starting Epoch 4: learning rate per sample = 0.003125  effective momentum = 0.900000  momentum as time constant = 9719.0 samples

12/20/2016 15:27:30: Starting minibatch loop.
12/20/2016 15:27:30:  Epoch[ 4 of 25]-Minibatch[   1-  10, 0.91%]: CE.SM = 1.35817089 * 10240; Err = 0.41425781 * 10240; time = 0.0599s; samplesPerSecond = 170905.9
12/20/2016 15:27:30:  Epoch[ 4 of 25]-Minibatch[  11-  20, 1.82%]: CE.SM = 1.34110641 * 10240; Err = 0.41396484 * 10240; time = 0.0431s; samplesPerSecond = 237465.8
12/20/2016 15:27:30:  Epoch[ 4 of 25]-Minibatch[  21-  30, 2.73%]: CE.SM = 1.39832973 * 10240; Err = 0.43212891 * 10240; time = 0.0425s; samplesPerSecond = 240725.9
12/20/2016 15:27:30:  Epoch[ 4 of 25]-Minibatch[  31-  40, 3.64%]: CE.SM = 1.37360268 * 10240; Err = 0.43417969 * 10240; time = 0.0410s; samplesPerSecond = 249646.5
12/20/2016 15:27:30:  Epoch[ 4 of 25]-Minibatch[  41-  50, 4.56%]: CE.SM = 1.37409439 * 10240; Err = 0.42714844 * 10240; time = 0.0411s; samplesPerSecond = 249154.5
12/20/2016 15:27:30:  Epoch[ 4 of 25]-Minibatch[  51-  60, 5.47%]: CE.SM = 1.38064957 * 10240; Err = 0.42978516 * 10240; time = 0.0426s; samplesPerSecond = 240291.0
12/20/2016 15:27:30:  Epoch[ 4 of 25]-Minibatch[  61-  70, 6.38%]: CE.SM = 1.36758728 * 10240; Err = 0.42519531 * 10240; time = 0.0413s; samplesPerSecond = 248176.2
12/20/2016 15:27:30:  Epoch[ 4 of 25]-Minibatch[  71-  80, 7.29%]: CE.SM = 1.38228302 * 10240; Err = 0.42851563 * 10240; time = 0.0420s; samplesPerSecond = 243960.5
12/20/2016 15:27:30:  Epoch[ 4 of 25]-Minibatch[  81-  90, 8.20%]: CE.SM = 1.36353378 * 10240; Err = 0.42089844 * 10240; time = 0.0411s; samplesPerSecond = 249415.4
12/20/2016 15:27:30:  Epoch[ 4 of 25]-Minibatch[  91- 100, 9.11%]: CE.SM = 1.34400101 * 10240; Err = 0.42031250 * 10240; time = 0.0431s; samplesPerSecond = 237647.7
12/20/2016 15:27:30:  Epoch[ 4 of 25]-Minibatch[ 101- 110, 10.02%]: CE.SM = 1.34909973 * 10240; Err = 0.42255859 * 10240; time = 0.0419s; samplesPerSecond = 244473.1
12/20/2016 15:27:30:  Epoch[ 4 of 25]-Minibatch[ 111- 120, 10.93%]: CE.SM = 1.34623566 * 10240; Err = 0.41699219 * 10240; time = 0.0417s; samplesPerSecond = 245604.8
12/20/2016 15:27:30:  Epoch[ 4 of 25]-Minibatch[ 121- 130, 11.85%]: CE.SM = 1.33050385 * 10240; Err = 0.41796875 * 10240; time = 0.0399s; samplesPerSecond = 256570.9
12/20/2016 15:27:30:  Epoch[ 4 of 25]-Minibatch[ 131- 140, 12.76%]: CE.SM = 1.34740906 * 10240; Err = 0.42089844 * 10240; time = 0.0418s; samplesPerSecond = 245169.6
12/20/2016 15:27:30:  Epoch[ 4 of 25]-Minibatch[ 141- 150, 13.67%]: CE.SM = 1.36467896 * 10240; Err = 0.42236328 * 10240; time = 0.0418s; samplesPerSecond = 245087.5
12/20/2016 15:27:30:  Epoch[ 4 of 25]-Minibatch[ 151- 160, 14.58%]: CE.SM = 1.36355438 * 10240; Err = 0.41982422 * 10240; time = 0.0419s; samplesPerSecond = 244251.5
12/20/2016 15:27:30:  Epoch[ 4 of 25]-Minibatch[ 161- 170, 15.49%]: CE.SM = 1.37304535 * 10240; Err = 0.42373047 * 10240; time = 0.0401s; samplesPerSecond = 255107.1
12/20/2016 15:27:30:  Epoch[ 4 of 25]-Minibatch[ 171- 180, 16.40%]: CE.SM = 1.33528442 * 10240; Err = 0.41816406 * 10240; time = 0.0424s; samplesPerSecond = 241270.4
12/20/2016 15:27:30:  Epoch[ 4 of 25]-Minibatch[ 181- 190, 17.31%]: CE.SM = 1.37006531 * 10240; Err = 0.42148438 * 10240; time = 0.0396s; samplesPerSecond = 258579.3
12/20/2016 15:27:30:  Epoch[ 4 of 25]-Minibatch[ 191- 200, 18.22%]: CE.SM = 1.34842529 * 10240; Err = 0.42109375 * 10240; time = 0.0412s; samplesPerSecond = 248368.9
12/20/2016 15:27:31:  Epoch[ 4 of 25]-Minibatch[ 201- 210, 19.13%]: CE.SM = 1.37335815 * 10240; Err = 0.42509766 * 10240; time = 0.0398s; samplesPerSecond = 257247.7
12/20/2016 15:27:31:  Epoch[ 4 of 25]-Minibatch[ 211- 220, 20.05%]: CE.SM = 1.35396118 * 10240; Err = 0.41591797 * 10240; time = 0.0411s; samplesPerSecond = 249003.0
12/20/2016 15:27:31:  Epoch[ 4 of 25]-Minibatch[ 221- 230, 20.96%]: CE.SM = 1.34263000 * 10240; Err = 0.41748047 * 10240; time = 0.0405s; samplesPerSecond = 252902.0
12/20/2016 15:27:31:  Epoch[ 4 of 25]-Minibatch[ 231- 240, 21.87%]: CE.SM = 1.34555664 * 10240; Err = 0.41445312 * 10240; time = 0.0413s; samplesPerSecond = 247953.9
12/20/2016 15:27:31:  Epoch[ 4 of 25]-Minibatch[ 241- 250, 22.78%]: CE.SM = 1.36204529 * 10240; Err = 0.42451172 * 10240; time = 0.0419s; samplesPerSecond = 244630.8
12/20/2016 15:27:31:  Epoch[ 4 of 25]-Minibatch[ 251- 260, 23.69%]: CE.SM = 1.34776611 * 10240; Err = 0.41386719 * 10240; time = 0.0405s; samplesPerSecond = 252565.1
12/20/2016 15:27:31:  Epoch[ 4 of 25]-Minibatch[ 261- 270, 24.60%]: CE.SM = 1.35566101 * 10240; Err = 0.42294922 * 10240; time = 0.0435s; samplesPerSecond = 235570.2
12/20/2016 15:27:31:  Epoch[ 4 of 25]-Minibatch[ 271- 280, 25.51%]: CE.SM = 1.33554687 * 10240; Err = 0.41298828 * 10240; time = 0.0416s; samplesPerSecond = 245876.1
12/20/2016 15:27:31:  Epoch[ 4 of 25]-Minibatch[ 281- 290, 26.42%]: CE.SM = 1.36751404 * 10240; Err = 0.42275391 * 10240; time = 0.0412s; samplesPerSecond = 248405.0
12/20/2016 15:27:31:  Epoch[ 4 of 25]-Minibatch[ 291- 300, 27.33%]: CE.SM = 1.33270569 * 10240; Err = 0.42050781 * 10240; time = 0.0422s; samplesPerSecond = 242861.2
12/20/2016 15:27:31:  Epoch[ 4 of 25]-Minibatch[ 301- 310, 28.25%]: CE.SM = 1.36026306 * 10240; Err = 0.43134766 * 10240; time = 0.0514s; samplesPerSecond = 199322.6
12/20/2016 15:27:31:  Epoch[ 4 of 25]-Minibatch[ 311- 320, 29.16%]: CE.SM = 1.34549561 * 10240; Err = 0.41611328 * 10240; time = 0.0431s; samplesPerSecond = 237576.0
12/20/2016 15:27:31:  Epoch[ 4 of 25]-Minibatch[ 321- 330, 30.07%]: CE.SM = 1.32396851 * 10240; Err = 0.41064453 * 10240; time = 0.0410s; samplesPerSecond = 249597.8
12/20/2016 15:27:31:  Epoch[ 4 of 25]-Minibatch[ 331- 340, 30.98%]: CE.SM = 1.34854126 * 10240; Err = 0.42255859 * 10240; time = 0.0406s; samplesPerSecond = 252111.2
12/20/2016 15:27:31:  Epoch[ 4 of 25]-Minibatch[ 341- 350, 31.89%]: CE.SM = 1.35168152 * 10240; Err = 0.42285156 * 10240; time = 0.0413s; samplesPerSecond = 248092.1
12/20/2016 15:27:31:  Epoch[ 4 of 25]-Minibatch[ 351- 360, 32.80%]: CE.SM = 1.34613953 * 10240; Err = 0.42060547 * 10240; time = 0.0398s; samplesPerSecond = 257357.6
12/20/2016 15:27:31:  Epoch[ 4 of 25]-Minibatch[ 361- 370, 33.71%]: CE.SM = 1.33414001 * 10240; Err = 0.41406250 * 10240; time = 0.0420s; samplesPerSecond = 243902.4
12/20/2016 15:27:31:  Epoch[ 4 of 25]-Minibatch[ 371- 380, 34.62%]: CE.SM = 1.35834961 * 10240; Err = 0.42832031 * 10240; time = 0.0411s; samplesPerSecond = 249130.2
12/20/2016 15:27:31:  Epoch[ 4 of 25]-Minibatch[ 381- 390, 35.54%]: CE.SM = 1.33499756 * 10240; Err = 0.41025391 * 10240; time = 0.0416s; samplesPerSecond = 246450.1
12/20/2016 15:27:31:  Epoch[ 4 of 25]-Minibatch[ 391- 400, 36.45%]: CE.SM = 1.35923462 * 10240; Err = 0.42753906 * 10240; time = 0.0422s; samplesPerSecond = 242470.2
12/20/2016 15:27:31:  Epoch[ 4 of 25]-Minibatch[ 401- 410, 37.36%]: CE.SM = 1.35809326 * 10240; Err = 0.42568359 * 10240; time = 0.0419s; samplesPerSecond = 244338.9
12/20/2016 15:27:31:  Epoch[ 4 of 25]-Minibatch[ 411- 420, 38.27%]: CE.SM = 1.32155151 * 10240; Err = 0.40820312 * 10240; time = 0.0461s; samplesPerSecond = 221914.0
12/20/2016 15:27:31:  Epoch[ 4 of 25]-Minibatch[ 421- 430, 39.18%]: CE.SM = 1.31958618 * 10240; Err = 0.41513672 * 10240; time = 0.0426s; samplesPerSecond = 240652.4
12/20/2016 15:27:32:  Epoch[ 4 of 25]-Minibatch[ 431- 440, 40.09%]: CE.SM = 1.35014648 * 10240; Err = 0.42304687 * 10240; time = 0.0458s; samplesPerSecond = 223434.4
12/20/2016 15:27:32:  Epoch[ 4 of 25]-Minibatch[ 441- 450, 41.00%]: CE.SM = 1.32628784 * 10240; Err = 0.41425781 * 10240; time = 0.0407s; samplesPerSecond = 251714.6
12/20/2016 15:27:32:  Epoch[ 4 of 25]-Minibatch[ 451- 460, 41.91%]: CE.SM = 1.34207764 * 10240; Err = 0.42138672 * 10240; time = 0.0414s; samplesPerSecond = 247516.4
12/20/2016 15:27:32:  Epoch[ 4 of 25]-Minibatch[ 461- 470, 42.82%]: CE.SM = 1.34669800 * 10240; Err = 0.42099609 * 10240; time = 0.0428s; samplesPerSecond = 239280.3
12/20/2016 15:27:32:  Epoch[ 4 of 25]-Minibatch[ 471- 480, 43.74%]: CE.SM = 1.37875977 * 10240; Err = 0.42597656 * 10240; time = 0.0424s; samplesPerSecond = 241441.1
12/20/2016 15:27:32:  Epoch[ 4 of 25]-Minibatch[ 481- 490, 44.65%]: CE.SM = 1.34729614 * 10240; Err = 0.41943359 * 10240; time = 0.0417s; samplesPerSecond = 245404.7
12/20/2016 15:27:32:  Epoch[ 4 of 25]-Minibatch[ 491- 500, 45.56%]: CE.SM = 1.31108398 * 10240; Err = 0.40849609 * 10240; time = 0.0405s; samplesPerSecond = 252839.5
12/20/2016 15:27:32:  Epoch[ 4 of 25]-Minibatch[ 501- 510, 46.47%]: CE.SM = 1.32390747 * 10240; Err = 0.41582031 * 10240; time = 0.0406s; samplesPerSecond = 252067.7
12/20/2016 15:27:32:  Epoch[ 4 of 25]-Minibatch[ 511- 520, 47.38%]: CE.SM = 1.34372559 * 10240; Err = 0.41904297 * 10240; time = 0.0407s; samplesPerSecond = 251782.6
12/20/2016 15:27:32:  Epoch[ 4 of 25]-Minibatch[ 521- 530, 48.29%]: CE.SM = 1.33041382 * 10240; Err = 0.41591797 * 10240; time = 0.0411s; samplesPerSecond = 249385.1
12/20/2016 15:27:32:  Epoch[ 4 of 25]-Minibatch[ 531- 540, 49.20%]: CE.SM = 1.33610229 * 10240; Err = 0.41152344 * 10240; time = 0.0420s; samplesPerSecond = 243885.0
12/20/2016 15:27:32:  Epoch[ 4 of 25]-Minibatch[ 541- 550, 50.11%]: CE.SM = 1.32331543 * 10240; Err = 0.41230469 * 10240; time = 0.0407s; samplesPerSecond = 251510.5
12/20/2016 15:27:32:  Epoch[ 4 of 25]-Minibatch[ 551- 560, 51.03%]: CE.SM = 1.32390137 * 10240; Err = 0.40878906 * 10240; time = 0.0409s; samplesPerSecond = 250250.5
12/20/2016 15:27:32:  Epoch[ 4 of 25]-Minibatch[ 561- 570, 51.94%]: CE.SM = 1.34798584 * 10240; Err = 0.41826172 * 10240; time = 0.0440s; samplesPerSecond = 232960.2
12/20/2016 15:27:32:  Epoch[ 4 of 25]-Minibatch[ 571- 580, 52.85%]: CE.SM = 1.34954834 * 10240; Err = 0.42011719 * 10240; time = 0.0425s; samplesPerSecond = 240748.6
12/20/2016 15:27:32:  Epoch[ 4 of 25]-Minibatch[ 581- 590, 53.76%]: CE.SM = 1.33047485 * 10240; Err = 0.41298828 * 10240; time = 0.0409s; samplesPerSecond = 250605.7
12/20/2016 15:27:32:  Epoch[ 4 of 25]-Minibatch[ 591- 600, 54.67%]: CE.SM = 1.31434937 * 10240; Err = 0.41318359 * 10240; time = 0.0407s; samplesPerSecond = 251813.6
12/20/2016 15:27:32:  Epoch[ 4 of 25]-Minibatch[ 601- 610, 55.58%]: CE.SM = 1.33697510 * 10240; Err = 0.42031250 * 10240; time = 0.0406s; samplesPerSecond = 252502.8
12/20/2016 15:27:32:  Epoch[ 4 of 25]-Minibatch[ 611- 620, 56.49%]: CE.SM = 1.34573364 * 10240; Err = 0.41748047 * 10240; time = 0.0415s; samplesPerSecond = 247014.8
12/20/2016 15:27:32:  Epoch[ 4 of 25]-Minibatch[ 621- 630, 57.40%]: CE.SM = 1.32885742 * 10240; Err = 0.40830078 * 10240; time = 0.0411s; samplesPerSecond = 249391.1
12/20/2016 15:27:32:  Epoch[ 4 of 25]-Minibatch[ 631- 640, 58.31%]: CE.SM = 1.33634033 * 10240; Err = 0.41484375 * 10240; time = 0.0413s; samplesPerSecond = 247887.9
12/20/2016 15:27:32:  Epoch[ 4 of 25]-Minibatch[ 641- 650, 59.23%]: CE.SM = 1.32958984 * 10240; Err = 0.41513672 * 10240; time = 0.0412s; samplesPerSecond = 248573.9
12/20/2016 15:27:32:  Epoch[ 4 of 25]-Minibatch[ 651- 660, 60.14%]: CE.SM = 1.35247192 * 10240; Err = 0.42402344 * 10240; time = 0.0418s; samplesPerSecond = 244771.1
12/20/2016 15:27:32:  Epoch[ 4 of 25]-Minibatch[ 661- 670, 61.05%]: CE.SM = 1.35823975 * 10240; Err = 0.42167969 * 10240; time = 0.0414s; samplesPerSecond = 247372.9
12/20/2016 15:27:33:  Epoch[ 4 of 25]-Minibatch[ 671- 680, 61.96%]: CE.SM = 1.34785156 * 10240; Err = 0.41142578 * 10240; time = 0.0444s; samplesPerSecond = 230480.1
12/20/2016 15:27:33:  Epoch[ 4 of 25]-Minibatch[ 681- 690, 62.87%]: CE.SM = 1.35758057 * 10240; Err = 0.41845703 * 10240; time = 0.0421s; samplesPerSecond = 243224.6
12/20/2016 15:27:33:  Epoch[ 4 of 25]-Minibatch[ 691- 700, 63.78%]: CE.SM = 1.33103638 * 10240; Err = 0.42099609 * 10240; time = 0.0494s; samplesPerSecond = 207262.3
12/20/2016 15:27:33:  Epoch[ 4 of 25]-Minibatch[ 701- 710, 64.69%]: CE.SM = 1.33524170 * 10240; Err = 0.41845703 * 10240; time = 0.0508s; samplesPerSecond = 201503.4
12/20/2016 15:27:33:  Epoch[ 4 of 25]-Minibatch[ 711- 720, 65.60%]: CE.SM = 1.35054321 * 10240; Err = 0.41728516 * 10240; time = 0.0408s; samplesPerSecond = 250703.9
12/20/2016 15:27:33:  Epoch[ 4 of 25]-Minibatch[ 721- 730, 66.51%]: CE.SM = 1.32147827 * 10240; Err = 0.41347656 * 10240; time = 0.0403s; samplesPerSecond = 254245.7
12/20/2016 15:27:33:  Epoch[ 4 of 25]-Minibatch[ 731- 740, 67.43%]: CE.SM = 1.32918701 * 10240; Err = 0.41220703 * 10240; time = 0.0407s; samplesPerSecond = 251838.4
12/20/2016 15:27:33:  Epoch[ 4 of 25]-Minibatch[ 741- 750, 68.34%]: CE.SM = 1.33402100 * 10240; Err = 0.40878906 * 10240; time = 0.0408s; samplesPerSecond = 251060.4
12/20/2016 15:27:33:  Epoch[ 4 of 25]-Minibatch[ 751- 760, 69.25%]: CE.SM = 1.33899536 * 10240; Err = 0.41542969 * 10240; time = 0.0440s; samplesPerSecond = 232473.7
12/20/2016 15:27:33:  Epoch[ 4 of 25]-Minibatch[ 761- 770, 70.16%]: CE.SM = 1.34428101 * 10240; Err = 0.41923828 * 10240; time = 0.0400s; samplesPerSecond = 255916.8
12/20/2016 15:27:33:  Epoch[ 4 of 25]-Minibatch[ 771- 780, 71.07%]: CE.SM = 1.35068359 * 10240; Err = 0.41757813 * 10240; time = 0.0454s; samplesPerSecond = 225739.6
12/20/2016 15:27:33:  Epoch[ 4 of 25]-Minibatch[ 781- 790, 71.98%]: CE.SM = 1.32767334 * 10240; Err = 0.41015625 * 10240; time = 0.0421s; samplesPerSecond = 243421.2
12/20/2016 15:27:33:  Epoch[ 4 of 25]-Minibatch[ 791- 800, 72.89%]: CE.SM = 1.32790527 * 10240; Err = 0.41689453 * 10240; time = 0.0414s; samplesPerSecond = 247295.2
12/20/2016 15:27:33:  Epoch[ 4 of 25]-Minibatch[ 801- 810, 73.80%]: CE.SM = 1.33123779 * 10240; Err = 0.41445312 * 10240; time = 0.0433s; samplesPerSecond = 236244.1
12/20/2016 15:27:33:  Epoch[ 4 of 25]-Minibatch[ 811- 820, 74.72%]: CE.SM = 1.32718506 * 10240; Err = 0.41328125 * 10240; time = 0.0391s; samplesPerSecond = 261691.8
12/20/2016 15:27:33:  Epoch[ 4 of 25]-Minibatch[ 821- 830, 75.63%]: CE.SM = 1.32331543 * 10240; Err = 0.41220703 * 10240; time = 0.0402s; samplesPerSecond = 254625.0
12/20/2016 15:27:33:  Epoch[ 4 of 25]-Minibatch[ 831- 840, 76.54%]: CE.SM = 1.32615967 * 10240; Err = 0.41542969 * 10240; time = 0.0416s; samplesPerSecond = 245870.1
12/20/2016 15:27:33:  Epoch[ 4 of 25]-Minibatch[ 841- 850, 77.45%]: CE.SM = 1.34902344 * 10240; Err = 0.42138672 * 10240; time = 0.0402s; samplesPerSecond = 254504.8
12/20/2016 15:27:33:  Epoch[ 4 of 25]-Minibatch[ 851- 860, 78.36%]: CE.SM = 1.37004395 * 10240; Err = 0.42197266 * 10240; time = 0.0406s; samplesPerSecond = 251980.9
12/20/2016 15:27:33:  Epoch[ 4 of 25]-Minibatch[ 861- 870, 79.27%]: CE.SM = 1.30803223 * 10240; Err = 0.40644531 * 10240; time = 0.0405s; samplesPerSecond = 252895.7
12/20/2016 15:27:33:  Epoch[ 4 of 25]-Minibatch[ 871- 880, 80.18%]: CE.SM = 1.32404785 * 10240; Err = 0.42021484 * 10240; time = 0.0394s; samplesPerSecond = 259872.1
12/20/2016 15:27:33:  Epoch[ 4 of 25]-Minibatch[ 881- 890, 81.09%]: CE.SM = 1.32064209 * 10240; Err = 0.41875000 * 10240; time = 0.0409s; samplesPerSecond = 250183.2
12/20/2016 15:27:33:  Epoch[ 4 of 25]-Minibatch[ 891- 900, 82.00%]: CE.SM = 1.33375244 * 10240; Err = 0.41943359 * 10240; time = 0.0411s; samplesPerSecond = 248960.6
12/20/2016 15:27:33:  Epoch[ 4 of 25]-Minibatch[ 901- 910, 82.92%]: CE.SM = 1.30600586 * 10240; Err = 0.41513672 * 10240; time = 0.0417s; samplesPerSecond = 245516.4
12/20/2016 15:27:34:  Epoch[ 4 of 25]-Minibatch[ 911- 920, 83.83%]: CE.SM = 1.30021973 * 10240; Err = 0.41220703 * 10240; time = 0.0412s; samplesPerSecond = 248592.0
12/20/2016 15:27:34:  Epoch[ 4 of 25]-Minibatch[ 921- 930, 84.74%]: CE.SM = 1.31831055 * 10240; Err = 0.41210938 * 10240; time = 0.0412s; samplesPerSecond = 248779.2
12/20/2016 15:27:34:  Epoch[ 4 of 25]-Minibatch[ 931- 940, 85.65%]: CE.SM = 1.33393555 * 10240; Err = 0.41044922 * 10240; time = 0.0398s; samplesPerSecond = 257112.0
12/20/2016 15:27:34:  Epoch[ 4 of 25]-Minibatch[ 941- 950, 86.56%]: CE.SM = 1.30236816 * 10240; Err = 0.40214844 * 10240; time = 0.0405s; samplesPerSecond = 253070.7
12/20/2016 15:27:34:  Epoch[ 4 of 25]-Minibatch[ 951- 960, 87.47%]: CE.SM = 1.31190186 * 10240; Err = 0.40966797 * 10240; time = 0.0398s; samplesPerSecond = 257041.0
12/20/2016 15:27:34:  Epoch[ 4 of 25]-Minibatch[ 961- 970, 88.38%]: CE.SM = 1.32177734 * 10240; Err = 0.40839844 * 10240; time = 0.0412s; samplesPerSecond = 248821.5
12/20/2016 15:27:34:  Epoch[ 4 of 25]-Minibatch[ 971- 980, 89.29%]: CE.SM = 1.32911377 * 10240; Err = 0.41464844 * 10240; time = 0.0404s; samplesPerSecond = 253246.0
12/20/2016 15:27:34:  Epoch[ 4 of 25]-Minibatch[ 981- 990, 90.21%]: CE.SM = 1.32486572 * 10240; Err = 0.41240234 * 10240; time = 0.0406s; samplesPerSecond = 252434.4
12/20/2016 15:27:34:  Epoch[ 4 of 25]-Minibatch[ 991-1000, 91.12%]: CE.SM = 1.33121338 * 10240; Err = 0.41132812 * 10240; time = 0.0410s; samplesPerSecond = 249865.8
12/20/2016 15:27:34:  Epoch[ 4 of 25]-Minibatch[1001-1010, 92.03%]: CE.SM = 1.32386475 * 10240; Err = 0.40859375 * 10240; time = 0.0402s; samplesPerSecond = 254542.7
12/20/2016 15:27:34:  Epoch[ 4 of 25]-Minibatch[1011-1020, 92.94%]: CE.SM = 1.32733154 * 10240; Err = 0.41923828 * 10240; time = 0.0446s; samplesPerSecond = 229524.4
12/20/2016 15:27:34:  Epoch[ 4 of 25]-Minibatch[1021-1030, 93.85%]: CE.SM = 1.33489990 * 10240; Err = 0.41171875 * 10240; time = 0.0463s; samplesPerSecond = 221333.6
12/20/2016 15:27:34:  Epoch[ 4 of 25]-Minibatch[1031-1040, 94.76%]: CE.SM = 1.30596924 * 10240; Err = 0.40830078 * 10240; time = 0.0527s; samplesPerSecond = 194333.2
12/20/2016 15:27:34:  Epoch[ 4 of 25]-Minibatch[1041-1050, 95.67%]: CE.SM = 1.32220459 * 10240; Err = 0.41259766 * 10240; time = 0.0425s; samplesPerSecond = 241020.6
12/20/2016 15:27:34:  Epoch[ 4 of 25]-Minibatch[1051-1060, 96.58%]: CE.SM = 1.31068115 * 10240; Err = 0.41347656 * 10240; time = 0.0405s; samplesPerSecond = 252945.7
12/20/2016 15:27:34:  Epoch[ 4 of 25]-Minibatch[1061-1070, 97.49%]: CE.SM = 1.31147461 * 10240; Err = 0.41035156 * 10240; time = 0.0405s; samplesPerSecond = 252970.7
12/20/2016 15:27:34:  Epoch[ 4 of 25]-Minibatch[1071-1080, 98.41%]: CE.SM = 1.30133057 * 10240; Err = 0.41289063 * 10240; time = 0.0413s; samplesPerSecond = 247749.9
12/20/2016 15:27:34:  Epoch[ 4 of 25]-Minibatch[1081-1090, 99.32%]: CE.SM = 1.30510254 * 10240; Err = 0.41250000 * 10240; time = 0.0415s; samplesPerSecond = 246830.3
12/20/2016 15:27:34: Finished Epoch[ 4 of 25]: [Training] CE.SM = 1.33966022 * 1124823; Err = 0.41725765 * 1124823; totalSamplesSeen = 4499292; learningRatePerSample = 0.003125; epochTime=4.81139s
12/20/2016 15:27:34: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn.4'

12/20/2016 15:27:34: Starting Epoch 5: learning rate per sample = 0.003125  effective momentum = 0.900000  momentum as time constant = 9719.0 samples

12/20/2016 15:27:35: Starting minibatch loop.
12/20/2016 15:27:35:  Epoch[ 5 of 25]-Minibatch[   1-  10, 0.91%]: CE.SM = 1.28021259 * 10240; Err = 0.40234375 * 10240; time = 0.0521s; samplesPerSecond = 196624.4
12/20/2016 15:27:35:  Epoch[ 5 of 25]-Minibatch[  11-  20, 1.82%]: CE.SM = 1.33386364 * 10240; Err = 0.41708984 * 10240; time = 0.0425s; samplesPerSecond = 240782.5
12/20/2016 15:27:35:  Epoch[ 5 of 25]-Minibatch[  21-  30, 2.73%]: CE.SM = 1.29482460 * 10240; Err = 0.40693359 * 10240; time = 0.0407s; samplesPerSecond = 251399.4
12/20/2016 15:27:35:  Epoch[ 5 of 25]-Minibatch[  31-  40, 3.64%]: CE.SM = 1.31196480 * 10240; Err = 0.40527344 * 10240; time = 0.0415s; samplesPerSecond = 246616.3
12/20/2016 15:27:35:  Epoch[ 5 of 25]-Minibatch[  41-  50, 4.56%]: CE.SM = 1.31090317 * 10240; Err = 0.40605469 * 10240; time = 0.0413s; samplesPerSecond = 247672.0
12/20/2016 15:27:35:  Epoch[ 5 of 25]-Minibatch[  51-  60, 5.47%]: CE.SM = 1.28200836 * 10240; Err = 0.40380859 * 10240; time = 0.0411s; samplesPerSecond = 249318.3
12/20/2016 15:27:35:  Epoch[ 5 of 25]-Minibatch[  61-  70, 6.38%]: CE.SM = 1.29862213 * 10240; Err = 0.40449219 * 10240; time = 0.0408s; samplesPerSecond = 250968.1
12/20/2016 15:27:35:  Epoch[ 5 of 25]-Minibatch[  71-  80, 7.29%]: CE.SM = 1.30185699 * 10240; Err = 0.40732422 * 10240; time = 0.0409s; samplesPerSecond = 250415.7
12/20/2016 15:27:35:  Epoch[ 5 of 25]-Minibatch[  81-  90, 8.20%]: CE.SM = 1.28959579 * 10240; Err = 0.40546875 * 10240; time = 0.0415s; samplesPerSecond = 246830.3
12/20/2016 15:27:35:  Epoch[ 5 of 25]-Minibatch[  91- 100, 9.11%]: CE.SM = 1.29165726 * 10240; Err = 0.40449219 * 10240; time = 0.0403s; samplesPerSecond = 253968.3
12/20/2016 15:27:35:  Epoch[ 5 of 25]-Minibatch[ 101- 110, 10.02%]: CE.SM = 1.30497894 * 10240; Err = 0.40351562 * 10240; time = 0.0406s; samplesPerSecond = 252477.9
12/20/2016 15:27:35:  Epoch[ 5 of 25]-Minibatch[ 111- 120, 10.93%]: CE.SM = 1.30336304 * 10240; Err = 0.40800781 * 10240; time = 0.0410s; samplesPerSecond = 249859.7
12/20/2016 15:27:35:  Epoch[ 5 of 25]-Minibatch[ 121- 130, 11.85%]: CE.SM = 1.26690216 * 10240; Err = 0.40341797 * 10240; time = 0.0416s; samplesPerSecond = 245952.8
12/20/2016 15:27:35:  Epoch[ 5 of 25]-Minibatch[ 131- 140, 12.76%]: CE.SM = 1.31467896 * 10240; Err = 0.40917969 * 10240; time = 0.0410s; samplesPerSecond = 249530.9
12/20/2016 15:27:35:  Epoch[ 5 of 25]-Minibatch[ 141- 150, 13.67%]: CE.SM = 1.28602905 * 10240; Err = 0.40351562 * 10240; time = 0.0403s; samplesPerSecond = 253880.1
12/20/2016 15:27:35:  Epoch[ 5 of 25]-Minibatch[ 151- 160, 14.58%]: CE.SM = 1.29854126 * 10240; Err = 0.40234375 * 10240; time = 0.0412s; samplesPerSecond = 248592.0
12/20/2016 15:27:35:  Epoch[ 5 of 25]-Minibatch[ 161- 170, 15.49%]: CE.SM = 1.28179016 * 10240; Err = 0.39521484 * 10240; time = 0.0409s; samplesPerSecond = 250618.0
12/20/2016 15:27:35:  Epoch[ 5 of 25]-Minibatch[ 171- 180, 16.40%]: CE.SM = 1.29136200 * 10240; Err = 0.40849609 * 10240; time = 0.0413s; samplesPerSecond = 247833.9
12/20/2016 15:27:35:  Epoch[ 5 of 25]-Minibatch[ 181- 190, 17.31%]: CE.SM = 1.28007812 * 10240; Err = 0.40292969 * 10240; time = 0.0412s; samplesPerSecond = 248374.9
12/20/2016 15:27:35:  Epoch[ 5 of 25]-Minibatch[ 191- 200, 18.22%]: CE.SM = 1.29228210 * 10240; Err = 0.40996094 * 10240; time = 0.0401s; samplesPerSecond = 255240.7
12/20/2016 15:27:35:  Epoch[ 5 of 25]-Minibatch[ 201- 210, 19.13%]: CE.SM = 1.31485291 * 10240; Err = 0.40556641 * 10240; time = 0.0413s; samplesPerSecond = 248146.2
12/20/2016 15:27:36:  Epoch[ 5 of 25]-Minibatch[ 211- 220, 20.05%]: CE.SM = 1.29952698 * 10240; Err = 0.40830078 * 10240; time = 0.0404s; samplesPerSecond = 253321.1
12/20/2016 15:27:36:  Epoch[ 5 of 25]-Minibatch[ 221- 230, 20.96%]: CE.SM = 1.29268188 * 10240; Err = 0.41123047 * 10240; time = 0.0412s; samplesPerSecond = 248640.2
12/20/2016 15:27:36:  Epoch[ 5 of 25]-Minibatch[ 231- 240, 21.87%]: CE.SM = 1.30311584 * 10240; Err = 0.41103516 * 10240; time = 0.0415s; samplesPerSecond = 246860.0
12/20/2016 15:27:36:  Epoch[ 5 of 25]-Minibatch[ 241- 250, 22.78%]: CE.SM = 1.29775085 * 10240; Err = 0.41328125 * 10240; time = 0.0417s; samplesPerSecond = 245817.0
12/20/2016 15:27:36:  Epoch[ 5 of 25]-Minibatch[ 251- 260, 23.69%]: CE.SM = 1.29652710 * 10240; Err = 0.40751953 * 10240; time = 0.0416s; samplesPerSecond = 246248.6
12/20/2016 15:27:36:  Epoch[ 5 of 25]-Minibatch[ 261- 270, 24.60%]: CE.SM = 1.26312561 * 10240; Err = 0.40029297 * 10240; time = 0.0420s; samplesPerSecond = 243705.1
12/20/2016 15:27:36:  Epoch[ 5 of 25]-Minibatch[ 271- 280, 25.51%]: CE.SM = 1.31388550 * 10240; Err = 0.40322266 * 10240; time = 0.0420s; samplesPerSecond = 243786.3
12/20/2016 15:27:36:  Epoch[ 5 of 25]-Minibatch[ 281- 290, 26.42%]: CE.SM = 1.29604492 * 10240; Err = 0.40322266 * 10240; time = 0.0439s; samplesPerSecond = 233257.4
12/20/2016 15:27:36:  Epoch[ 5 of 25]-Minibatch[ 291- 300, 27.33%]: CE.SM = 1.30384216 * 10240; Err = 0.41113281 * 10240; time = 0.0410s; samplesPerSecond = 249975.6
12/20/2016 15:27:36:  Epoch[ 5 of 25]-Minibatch[ 301- 310, 28.25%]: CE.SM = 1.29468994 * 10240; Err = 0.40703125 * 10240; time = 0.0426s; samplesPerSecond = 240646.7
12/20/2016 15:27:36:  Epoch[ 5 of 25]-Minibatch[ 311- 320, 29.16%]: CE.SM = 1.27944946 * 10240; Err = 0.40673828 * 10240; time = 0.0416s; samplesPerSecond = 246248.6
12/20/2016 15:27:36:  Epoch[ 5 of 25]-Minibatch[ 321- 330, 30.07%]: CE.SM = 1.29638672 * 10240; Err = 0.40927734 * 10240; time = 0.0413s; samplesPerSecond = 248020.0
12/20/2016 15:27:36:  Epoch[ 5 of 25]-Minibatch[ 331- 340, 30.98%]: CE.SM = 1.28938904 * 10240; Err = 0.40400391 * 10240; time = 0.0420s; samplesPerSecond = 243972.2
12/20/2016 15:27:36:  Epoch[ 5 of 25]-Minibatch[ 341- 350, 31.89%]: CE.SM = 1.30159607 * 10240; Err = 0.40097656 * 10240; time = 0.0435s; samplesPerSecond = 235386.1
12/20/2016 15:27:36:  Epoch[ 5 of 25]-Minibatch[ 351- 360, 32.80%]: CE.SM = 1.28917847 * 10240; Err = 0.40410156 * 10240; time = 0.0459s; samplesPerSecond = 223127.7
12/20/2016 15:27:36:  Epoch[ 5 of 25]-Minibatch[ 361- 370, 33.71%]: CE.SM = 1.25986023 * 10240; Err = 0.39179687 * 10240; time = 0.0426s; samplesPerSecond = 240296.6
12/20/2016 15:27:36:  Epoch[ 5 of 25]-Minibatch[ 371- 380, 34.62%]: CE.SM = 1.29423218 * 10240; Err = 0.40654297 * 10240; time = 0.0434s; samplesPerSecond = 235678.6
12/20/2016 15:27:36:  Epoch[ 5 of 25]-Minibatch[ 381- 390, 35.54%]: CE.SM = 1.29831848 * 10240; Err = 0.40820312 * 10240; time = 0.0411s; samplesPerSecond = 249221.2
12/20/2016 15:27:36:  Epoch[ 5 of 25]-Minibatch[ 391- 400, 36.45%]: CE.SM = 1.27264404 * 10240; Err = 0.40156250 * 10240; time = 0.0420s; samplesPerSecond = 243873.4
12/20/2016 15:27:36:  Epoch[ 5 of 25]-Minibatch[ 401- 410, 37.36%]: CE.SM = 1.28237915 * 10240; Err = 0.39619141 * 10240; time = 0.0413s; samplesPerSecond = 247959.9
12/20/2016 15:27:36:  Epoch[ 5 of 25]-Minibatch[ 411- 420, 38.27%]: CE.SM = 1.28211670 * 10240; Err = 0.40224609 * 10240; time = 0.0405s; samplesPerSecond = 252777.1
12/20/2016 15:27:36:  Epoch[ 5 of 25]-Minibatch[ 421- 430, 39.18%]: CE.SM = 1.28646240 * 10240; Err = 0.40205078 * 10240; time = 0.0419s; samplesPerSecond = 244432.2
12/20/2016 15:27:36:  Epoch[ 5 of 25]-Minibatch[ 431- 440, 40.09%]: CE.SM = 1.27896118 * 10240; Err = 0.40468750 * 10240; time = 0.0408s; samplesPerSecond = 250716.2
12/20/2016 15:27:36:  Epoch[ 5 of 25]-Minibatch[ 441- 450, 41.00%]: CE.SM = 1.30128174 * 10240; Err = 0.40439453 * 10240; time = 0.0402s; samplesPerSecond = 254986.4
12/20/2016 15:27:37:  Epoch[ 5 of 25]-Minibatch[ 451- 460, 41.91%]: CE.SM = 1.28885498 * 10240; Err = 0.40654297 * 10240; time = 0.0567s; samplesPerSecond = 180708.0
12/20/2016 15:27:37:  Epoch[ 5 of 25]-Minibatch[ 461- 470, 42.82%]: CE.SM = 1.29921875 * 10240; Err = 0.40126953 * 10240; time = 0.0421s; samplesPerSecond = 243363.4
12/20/2016 15:27:37:  Epoch[ 5 of 25]-Minibatch[ 471- 480, 43.74%]: CE.SM = 1.29867554 * 10240; Err = 0.41142578 * 10240; time = 0.0447s; samplesPerSecond = 229082.8
12/20/2016 15:27:37:  Epoch[ 5 of 25]-Minibatch[ 481- 490, 44.65%]: CE.SM = 1.29379883 * 10240; Err = 0.40751953 * 10240; time = 0.0467s; samplesPerSecond = 219173.4
12/20/2016 15:27:37:  Epoch[ 5 of 25]-Minibatch[ 491- 500, 45.56%]: CE.SM = 1.28986206 * 10240; Err = 0.40478516 * 10240; time = 0.0410s; samplesPerSecond = 249932.9
12/20/2016 15:27:37:  Epoch[ 5 of 25]-Minibatch[ 501- 510, 46.47%]: CE.SM = 1.26298218 * 10240; Err = 0.39804688 * 10240; time = 0.0402s; samplesPerSecond = 254650.4
12/20/2016 15:27:37:  Epoch[ 5 of 25]-Minibatch[ 511- 520, 47.38%]: CE.SM = 1.26405640 * 10240; Err = 0.39941406 * 10240; time = 0.0416s; samplesPerSecond = 245935.1
12/20/2016 15:27:37:  Epoch[ 5 of 25]-Minibatch[ 521- 530, 48.29%]: CE.SM = 1.29113159 * 10240; Err = 0.40771484 * 10240; time = 0.0417s; samplesPerSecond = 245651.9
12/20/2016 15:27:37:  Epoch[ 5 of 25]-Minibatch[ 531- 540, 49.20%]: CE.SM = 1.28683472 * 10240; Err = 0.40449219 * 10240; time = 0.0404s; samplesPerSecond = 253477.9
12/20/2016 15:27:37:  Epoch[ 5 of 25]-Minibatch[ 541- 550, 50.11%]: CE.SM = 1.27340698 * 10240; Err = 0.40107422 * 10240; time = 0.0417s; samplesPerSecond = 245805.2
12/20/2016 15:27:37:  Epoch[ 5 of 25]-Minibatch[ 551- 560, 51.03%]: CE.SM = 1.28594360 * 10240; Err = 0.40117188 * 10240; time = 0.0413s; samplesPerSecond = 247959.9
12/20/2016 15:27:37:  Epoch[ 5 of 25]-Minibatch[ 561- 570, 51.94%]: CE.SM = 1.27941895 * 10240; Err = 0.39843750 * 10240; time = 0.0418s; samplesPerSecond = 244829.6
12/20/2016 15:27:37:  Epoch[ 5 of 25]-Minibatch[ 571- 580, 52.85%]: CE.SM = 1.28391113 * 10240; Err = 0.40156250 * 10240; time = 0.0410s; samplesPerSecond = 249963.4
12/20/2016 15:27:37:  Epoch[ 5 of 25]-Minibatch[ 581- 590, 53.76%]: CE.SM = 1.28911743 * 10240; Err = 0.40839844 * 10240; time = 0.0411s; samplesPerSecond = 249336.5
12/20/2016 15:27:37:  Epoch[ 5 of 25]-Minibatch[ 591- 600, 54.67%]: CE.SM = 1.28579712 * 10240; Err = 0.40214844 * 10240; time = 0.0410s; samplesPerSecond = 249464.0
12/20/2016 15:27:37:  Epoch[ 5 of 25]-Minibatch[ 601- 610, 55.58%]: CE.SM = 1.29443359 * 10240; Err = 0.41171875 * 10240; time = 0.0402s; samplesPerSecond = 254580.7
12/20/2016 15:27:37:  Epoch[ 5 of 25]-Minibatch[ 611- 620, 56.49%]: CE.SM = 1.28944092 * 10240; Err = 0.39746094 * 10240; time = 0.0414s; samplesPerSecond = 247271.3
12/20/2016 15:27:37:  Epoch[ 5 of 25]-Minibatch[ 621- 630, 57.40%]: CE.SM = 1.29895020 * 10240; Err = 0.41132812 * 10240; time = 0.0414s; samplesPerSecond = 247301.2
12/20/2016 15:27:37:  Epoch[ 5 of 25]-Minibatch[ 631- 640, 58.31%]: CE.SM = 1.28858643 * 10240; Err = 0.39941406 * 10240; time = 0.0404s; samplesPerSecond = 253459.1
12/20/2016 15:27:37:  Epoch[ 5 of 25]-Minibatch[ 641- 650, 59.23%]: CE.SM = 1.29472656 * 10240; Err = 0.40644531 * 10240; time = 0.0413s; samplesPerSecond = 248032.0
12/20/2016 15:27:37:  Epoch[ 5 of 25]-Minibatch[ 651- 660, 60.14%]: CE.SM = 1.26145630 * 10240; Err = 0.39609375 * 10240; time = 0.0405s; samplesPerSecond = 252752.1
12/20/2016 15:27:37:  Epoch[ 5 of 25]-Minibatch[ 661- 670, 61.05%]: CE.SM = 1.27737427 * 10240; Err = 0.40488281 * 10240; time = 0.0401s; samplesPerSecond = 255291.6
12/20/2016 15:27:37:  Epoch[ 5 of 25]-Minibatch[ 671- 680, 61.96%]: CE.SM = 1.29746704 * 10240; Err = 0.40615234 * 10240; time = 0.0416s; samplesPerSecond = 246408.5
12/20/2016 15:27:38:  Epoch[ 5 of 25]-Minibatch[ 681- 690, 62.87%]: CE.SM = 1.29533081 * 10240; Err = 0.40664062 * 10240; time = 0.0408s; samplesPerSecond = 251257.5
12/20/2016 15:27:38:  Epoch[ 5 of 25]-Minibatch[ 691- 700, 63.78%]: CE.SM = 1.28114014 * 10240; Err = 0.40283203 * 10240; time = 0.0414s; samplesPerSecond = 247235.5
12/20/2016 15:27:38:  Epoch[ 5 of 25]-Minibatch[ 701- 710, 64.69%]: CE.SM = 1.28918457 * 10240; Err = 0.40390625 * 10240; time = 0.0405s; samplesPerSecond = 252883.2
12/20/2016 15:27:38:  Epoch[ 5 of 25]-Minibatch[ 711- 720, 65.60%]: CE.SM = 1.29985962 * 10240; Err = 0.40673828 * 10240; time = 0.0400s; samplesPerSecond = 256012.8
12/20/2016 15:27:38:  Epoch[ 5 of 25]-Minibatch[ 721- 730, 66.51%]: CE.SM = 1.30622559 * 10240; Err = 0.40498047 * 10240; time = 0.0453s; samplesPerSecond = 225918.9
12/20/2016 15:27:38:  Epoch[ 5 of 25]-Minibatch[ 731- 740, 67.43%]: CE.SM = 1.28754272 * 10240; Err = 0.40205078 * 10240; time = 0.0426s; samplesPerSecond = 240093.8
12/20/2016 15:27:38:  Epoch[ 5 of 25]-Minibatch[ 741- 750, 68.34%]: CE.SM = 1.26129150 * 10240; Err = 0.39697266 * 10240; time = 0.0418s; samplesPerSecond = 245152.0
12/20/2016 15:27:38:  Epoch[ 5 of 25]-Minibatch[ 751- 760, 69.25%]: CE.SM = 1.23162842 * 10240; Err = 0.38525391 * 10240; time = 0.0409s; samplesPerSecond = 250183.2
12/20/2016 15:27:38:  Epoch[ 5 of 25]-Minibatch[ 761- 770, 70.16%]: CE.SM = 1.29881592 * 10240; Err = 0.40341797 * 10240; time = 0.0414s; samplesPerSecond = 247134.1
12/20/2016 15:27:38:  Epoch[ 5 of 25]-Minibatch[ 771- 780, 71.07%]: CE.SM = 1.30406494 * 10240; Err = 0.41152344 * 10240; time = 0.0411s; samplesPerSecond = 249433.7
12/20/2016 15:27:38:  Epoch[ 5 of 25]-Minibatch[ 781- 790, 71.98%]: CE.SM = 1.27133789 * 10240; Err = 0.40058594 * 10240; time = 0.0406s; samplesPerSecond = 252011.9
12/20/2016 15:27:38:  Epoch[ 5 of 25]-Minibatch[ 791- 800, 72.89%]: CE.SM = 1.27949829 * 10240; Err = 0.40859375 * 10240; time = 0.0411s; samplesPerSecond = 249045.4
12/20/2016 15:27:38:  Epoch[ 5 of 25]-Minibatch[ 801- 810, 73.80%]: CE.SM = 1.27689209 * 10240; Err = 0.40371094 * 10240; time = 0.0408s; samplesPerSecond = 251195.9
12/20/2016 15:27:38:  Epoch[ 5 of 25]-Minibatch[ 811- 820, 74.72%]: CE.SM = 1.29516602 * 10240; Err = 0.41455078 * 10240; time = 0.0437s; samplesPerSecond = 234464.4
12/20/2016 15:27:38:  Epoch[ 5 of 25]-Minibatch[ 821- 830, 75.63%]: CE.SM = 1.28485107 * 10240; Err = 0.40322266 * 10240; time = 0.0403s; samplesPerSecond = 254289.9
12/20/2016 15:27:38:  Epoch[ 5 of 25]-Minibatch[ 831- 840, 76.54%]: CE.SM = 1.27679443 * 10240; Err = 0.39765625 * 10240; time = 0.0462s; samplesPerSecond = 221865.9
12/20/2016 15:27:38:  Epoch[ 5 of 25]-Minibatch[ 841- 850, 77.45%]: CE.SM = 1.27832031 * 10240; Err = 0.40107422 * 10240; time = 0.0452s; samplesPerSecond = 226533.6
12/20/2016 15:27:38:  Epoch[ 5 of 25]-Minibatch[ 851- 860, 78.36%]: CE.SM = 1.27703857 * 10240; Err = 0.40927734 * 10240; time = 0.0477s; samplesPerSecond = 214648.1
12/20/2016 15:27:38:  Epoch[ 5 of 25]-Minibatch[ 861- 870, 79.27%]: CE.SM = 1.28828125 * 10240; Err = 0.39941406 * 10240; time = 0.0407s; samplesPerSecond = 251473.5
12/20/2016 15:27:38:  Epoch[ 5 of 25]-Minibatch[ 871- 880, 80.18%]: CE.SM = 1.28966064 * 10240; Err = 0.40263672 * 10240; time = 0.0399s; samplesPerSecond = 256358.9
12/20/2016 15:27:38:  Epoch[ 5 of 25]-Minibatch[ 881- 890, 81.09%]: CE.SM = 1.27938232 * 10240; Err = 0.39853516 * 10240; time = 0.0410s; samplesPerSecond = 249969.5
12/20/2016 15:27:38:  Epoch[ 5 of 25]-Minibatch[ 891- 900, 82.00%]: CE.SM = 1.27634277 * 10240; Err = 0.40742187 * 10240; time = 0.0410s; samplesPerSecond = 249652.6
12/20/2016 15:27:38:  Epoch[ 5 of 25]-Minibatch[ 901- 910, 82.92%]: CE.SM = 1.27301025 * 10240; Err = 0.40302734 * 10240; time = 0.0412s; samplesPerSecond = 248755.0
12/20/2016 15:27:38:  Epoch[ 5 of 25]-Minibatch[ 911- 920, 83.83%]: CE.SM = 1.28358154 * 10240; Err = 0.40517578 * 10240; time = 0.0415s; samplesPerSecond = 246949.3
12/20/2016 15:27:39:  Epoch[ 5 of 25]-Minibatch[ 921- 930, 84.74%]: CE.SM = 1.26082764 * 10240; Err = 0.39580078 * 10240; time = 0.0410s; samplesPerSecond = 249658.7
12/20/2016 15:27:39:  Epoch[ 5 of 25]-Minibatch[ 931- 940, 85.65%]: CE.SM = 1.25783691 * 10240; Err = 0.39755859 * 10240; time = 0.0417s; samplesPerSecond = 245811.1
12/20/2016 15:27:39:  Epoch[ 5 of 25]-Minibatch[ 941- 950, 86.56%]: CE.SM = 1.26722412 * 10240; Err = 0.39785156 * 10240; time = 0.0422s; samplesPerSecond = 242677.0
12/20/2016 15:27:39:  Epoch[ 5 of 25]-Minibatch[ 951- 960, 87.47%]: CE.SM = 1.26816406 * 10240; Err = 0.40244141 * 10240; time = 0.0401s; samplesPerSecond = 255597.4
12/20/2016 15:27:39:  Epoch[ 5 of 25]-Minibatch[ 961- 970, 88.38%]: CE.SM = 1.27711182 * 10240; Err = 0.40361328 * 10240; time = 0.0441s; samplesPerSecond = 232115.3
12/20/2016 15:27:39:  Epoch[ 5 of 25]-Minibatch[ 971- 980, 89.29%]: CE.SM = 1.24700928 * 10240; Err = 0.39404297 * 10240; time = 0.0422s; samplesPerSecond = 242562.1
12/20/2016 15:27:39:  Epoch[ 5 of 25]-Minibatch[ 981- 990, 90.21%]: CE.SM = 1.27526855 * 10240; Err = 0.40556641 * 10240; time = 0.0409s; samplesPerSecond = 250415.7
12/20/2016 15:27:39:  Epoch[ 5 of 25]-Minibatch[ 991-1000, 91.12%]: CE.SM = 1.25354004 * 10240; Err = 0.39765625 * 10240; time = 0.0400s; samplesPerSecond = 255948.8
12/20/2016 15:27:39:  Epoch[ 5 of 25]-Minibatch[1001-1010, 92.03%]: CE.SM = 1.28457031 * 10240; Err = 0.40888672 * 10240; time = 0.0398s; samplesPerSecond = 257428.7
12/20/2016 15:27:39:  Epoch[ 5 of 25]-Minibatch[1011-1020, 92.94%]: CE.SM = 1.30015869 * 10240; Err = 0.40712891 * 10240; time = 0.0404s; samplesPerSecond = 253741.7
12/20/2016 15:27:39:  Epoch[ 5 of 25]-Minibatch[1021-1030, 93.85%]: CE.SM = 1.27807617 * 10240; Err = 0.39912109 * 10240; time = 0.0394s; samplesPerSecond = 259957.9
12/20/2016 15:27:39:  Epoch[ 5 of 25]-Minibatch[1031-1040, 94.76%]: CE.SM = 1.28537598 * 10240; Err = 0.40537109 * 10240; time = 0.0391s; samplesPerSecond = 261939.5
12/20/2016 15:27:39:  Epoch[ 5 of 25]-Minibatch[1041-1050, 95.67%]: CE.SM = 1.27290039 * 10240; Err = 0.40488281 * 10240; time = 0.0395s; samplesPerSecond = 259174.9
12/20/2016 15:27:39:  Epoch[ 5 of 25]-Minibatch[1051-1060, 96.58%]: CE.SM = 1.27459717 * 10240; Err = 0.39580078 * 10240; time = 0.0399s; samplesPerSecond = 256802.5
12/20/2016 15:27:39:  Epoch[ 5 of 25]-Minibatch[1061-1070, 97.49%]: CE.SM = 1.25560303 * 10240; Err = 0.39941406 * 10240; time = 0.0393s; samplesPerSecond = 260652.6
12/20/2016 15:27:39:  Epoch[ 5 of 25]-Minibatch[1071-1080, 98.41%]: CE.SM = 1.27862549 * 10240; Err = 0.39833984 * 10240; time = 0.0394s; samplesPerSecond = 259760.0
12/20/2016 15:27:39:  Epoch[ 5 of 25]-Minibatch[1081-1090, 99.32%]: CE.SM = 1.27590332 * 10240; Err = 0.40185547 * 10240; time = 0.0392s; samplesPerSecond = 261511.4
12/20/2016 15:27:39: Finished Epoch[ 5 of 25]: [Training] CE.SM = 1.28588387 * 1124823; Err = 0.40376664 * 1124823; totalSamplesSeen = 5624115; learningRatePerSample = 0.003125; epochTime=4.8243s
12/20/2016 15:27:39: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn.5'

12/20/2016 15:27:39: Starting Epoch 6: learning rate per sample = 0.003125  effective momentum = 0.900000  momentum as time constant = 9719.0 samples

12/20/2016 15:27:39: Starting minibatch loop.
12/20/2016 15:27:40:  Epoch[ 6 of 25]-Minibatch[   1-  10, 0.91%]: CE.SM = 1.26279192 * 10240; Err = 0.39941406 * 10240; time = 0.0537s; samplesPerSecond = 190575.4
12/20/2016 15:27:40:  Epoch[ 6 of 25]-Minibatch[  11-  20, 1.82%]: CE.SM = 1.27049913 * 10240; Err = 0.39755859 * 10240; time = 0.0415s; samplesPerSecond = 246663.8
12/20/2016 15:27:40:  Epoch[ 6 of 25]-Minibatch[  21-  30, 2.73%]: CE.SM = 1.27578411 * 10240; Err = 0.39902344 * 10240; time = 0.0407s; samplesPerSecond = 251615.6
12/20/2016 15:27:40:  Epoch[ 6 of 25]-Minibatch[  31-  40, 3.64%]: CE.SM = 1.24446411 * 10240; Err = 0.38828125 * 10240; time = 0.0405s; samplesPerSecond = 252652.4
12/20/2016 15:27:40:  Epoch[ 6 of 25]-Minibatch[  41-  50, 4.56%]: CE.SM = 1.22486229 * 10240; Err = 0.39062500 * 10240; time = 0.0402s; samplesPerSecond = 254891.2
12/20/2016 15:27:40:  Epoch[ 6 of 25]-Minibatch[  51-  60, 5.47%]: CE.SM = 1.22671471 * 10240; Err = 0.39033203 * 10240; time = 0.0404s; samplesPerSecond = 253515.5
12/20/2016 15:27:40:  Epoch[ 6 of 25]-Minibatch[  61-  70, 6.38%]: CE.SM = 1.24966583 * 10240; Err = 0.39277344 * 10240; time = 0.0414s; samplesPerSecond = 247295.2
12/20/2016 15:27:40:  Epoch[ 6 of 25]-Minibatch[  71-  80, 7.29%]: CE.SM = 1.26407166 * 10240; Err = 0.39589844 * 10240; time = 0.0443s; samplesPerSecond = 231402.0
12/20/2016 15:27:40:  Epoch[ 6 of 25]-Minibatch[  81-  90, 8.20%]: CE.SM = 1.24431763 * 10240; Err = 0.39433594 * 10240; time = 0.0445s; samplesPerSecond = 229947.0
12/20/2016 15:27:40:  Epoch[ 6 of 25]-Minibatch[  91- 100, 9.11%]: CE.SM = 1.25016861 * 10240; Err = 0.39648438 * 10240; time = 0.0413s; samplesPerSecond = 248128.1
12/20/2016 15:27:40:  Epoch[ 6 of 25]-Minibatch[ 101- 110, 10.02%]: CE.SM = 1.25867386 * 10240; Err = 0.39697266 * 10240; time = 0.0413s; samplesPerSecond = 248068.0
12/20/2016 15:27:40:  Epoch[ 6 of 25]-Minibatch[ 111- 120, 10.93%]: CE.SM = 1.25494385 * 10240; Err = 0.39306641 * 10240; time = 0.0471s; samplesPerSecond = 217608.4
12/20/2016 15:27:40:  Epoch[ 6 of 25]-Minibatch[ 121- 130, 11.85%]: CE.SM = 1.25063629 * 10240; Err = 0.39843750 * 10240; time = 0.0429s; samplesPerSecond = 238705.8
12/20/2016 15:27:40:  Epoch[ 6 of 25]-Minibatch[ 131- 140, 12.76%]: CE.SM = 1.22840729 * 10240; Err = 0.38730469 * 10240; time = 0.0419s; samplesPerSecond = 244362.2
12/20/2016 15:27:40:  Epoch[ 6 of 25]-Minibatch[ 141- 150, 13.67%]: CE.SM = 1.23425446 * 10240; Err = 0.39296875 * 10240; time = 0.0424s; samplesPerSecond = 241549.3
12/20/2016 15:27:40:  Epoch[ 6 of 25]-Minibatch[ 151- 160, 14.58%]: CE.SM = 1.25425873 * 10240; Err = 0.39375000 * 10240; time = 0.0417s; samplesPerSecond = 245445.8
12/20/2016 15:27:40:  Epoch[ 6 of 25]-Minibatch[ 161- 170, 15.49%]: CE.SM = 1.25036316 * 10240; Err = 0.38867188 * 10240; time = 0.0420s; samplesPerSecond = 243693.5
12/20/2016 15:27:40:  Epoch[ 6 of 25]-Minibatch[ 171- 180, 16.40%]: CE.SM = 1.26905060 * 10240; Err = 0.40000000 * 10240; time = 0.0416s; samplesPerSecond = 246195.3
12/20/2016 15:27:40:  Epoch[ 6 of 25]-Minibatch[ 181- 190, 17.31%]: CE.SM = 1.23720703 * 10240; Err = 0.39199219 * 10240; time = 0.0408s; samplesPerSecond = 250974.2
12/20/2016 15:27:40:  Epoch[ 6 of 25]-Minibatch[ 191- 200, 18.22%]: CE.SM = 1.22660828 * 10240; Err = 0.39130859 * 10240; time = 0.0415s; samplesPerSecond = 246717.3
12/20/2016 15:27:40:  Epoch[ 6 of 25]-Minibatch[ 201- 210, 19.13%]: CE.SM = 1.25379944 * 10240; Err = 0.39746094 * 10240; time = 0.0414s; samplesPerSecond = 247271.3
12/20/2016 15:27:40:  Epoch[ 6 of 25]-Minibatch[ 211- 220, 20.05%]: CE.SM = 1.24004211 * 10240; Err = 0.39833984 * 10240; time = 0.0420s; samplesPerSecond = 243757.3
12/20/2016 15:27:40:  Epoch[ 6 of 25]-Minibatch[ 221- 230, 20.96%]: CE.SM = 1.25570984 * 10240; Err = 0.39511719 * 10240; time = 0.0412s; samplesPerSecond = 248610.1
12/20/2016 15:27:41:  Epoch[ 6 of 25]-Minibatch[ 231- 240, 21.87%]: CE.SM = 1.25778198 * 10240; Err = 0.39375000 * 10240; time = 0.0412s; samplesPerSecond = 248344.8
12/20/2016 15:27:41:  Epoch[ 6 of 25]-Minibatch[ 241- 250, 22.78%]: CE.SM = 1.22656860 * 10240; Err = 0.39560547 * 10240; time = 0.0412s; samplesPerSecond = 248616.1
12/20/2016 15:27:41:  Epoch[ 6 of 25]-Minibatch[ 251- 260, 23.69%]: CE.SM = 1.22251587 * 10240; Err = 0.39384766 * 10240; time = 0.0422s; samplesPerSecond = 242901.5
12/20/2016 15:27:41:  Epoch[ 6 of 25]-Minibatch[ 261- 270, 24.60%]: CE.SM = 1.24310303 * 10240; Err = 0.39492187 * 10240; time = 0.0417s; samplesPerSecond = 245504.7
12/20/2016 15:27:41:  Epoch[ 6 of 25]-Minibatch[ 271- 280, 25.51%]: CE.SM = 1.23923645 * 10240; Err = 0.38916016 * 10240; time = 0.0404s; samplesPerSecond = 253622.3
12/20/2016 15:27:41:  Epoch[ 6 of 25]-Minibatch[ 281- 290, 26.42%]: CE.SM = 1.23521729 * 10240; Err = 0.39169922 * 10240; time = 0.0406s; samplesPerSecond = 252453.0
12/20/2016 15:27:41:  Epoch[ 6 of 25]-Minibatch[ 291- 300, 27.33%]: CE.SM = 1.24726257 * 10240; Err = 0.39228516 * 10240; time = 0.0406s; samplesPerSecond = 251980.9
12/20/2016 15:27:41:  Epoch[ 6 of 25]-Minibatch[ 301- 310, 28.25%]: CE.SM = 1.26290894 * 10240; Err = 0.39736328 * 10240; time = 0.0415s; samplesPerSecond = 246485.7
12/20/2016 15:27:41:  Epoch[ 6 of 25]-Minibatch[ 311- 320, 29.16%]: CE.SM = 1.25978394 * 10240; Err = 0.39863281 * 10240; time = 0.0411s; samplesPerSecond = 248960.6
12/20/2016 15:27:41:  Epoch[ 6 of 25]-Minibatch[ 321- 330, 30.07%]: CE.SM = 1.25247498 * 10240; Err = 0.39511719 * 10240; time = 0.0409s; samplesPerSecond = 250122.1
12/20/2016 15:27:41:  Epoch[ 6 of 25]-Minibatch[ 331- 340, 30.98%]: CE.SM = 1.26035156 * 10240; Err = 0.39550781 * 10240; time = 0.0402s; samplesPerSecond = 254460.5
12/20/2016 15:27:41:  Epoch[ 6 of 25]-Minibatch[ 341- 350, 31.89%]: CE.SM = 1.25110474 * 10240; Err = 0.40205078 * 10240; time = 0.0415s; samplesPerSecond = 246836.2
12/20/2016 15:27:41:  Epoch[ 6 of 25]-Minibatch[ 351- 360, 32.80%]: CE.SM = 1.24007568 * 10240; Err = 0.39082031 * 10240; time = 0.0424s; samplesPerSecond = 241367.1
12/20/2016 15:27:41:  Epoch[ 6 of 25]-Minibatch[ 361- 370, 33.71%]: CE.SM = 1.25094910 * 10240; Err = 0.39365234 * 10240; time = 0.0421s; samplesPerSecond = 243363.4
12/20/2016 15:27:41:  Epoch[ 6 of 25]-Minibatch[ 371- 380, 34.62%]: CE.SM = 1.23955994 * 10240; Err = 0.39443359 * 10240; time = 0.0401s; samplesPerSecond = 255520.9
12/20/2016 15:27:41:  Epoch[ 6 of 25]-Minibatch[ 381- 390, 35.54%]: CE.SM = 1.24211121 * 10240; Err = 0.39306641 * 10240; time = 0.0402s; samplesPerSecond = 254428.9
12/20/2016 15:27:41:  Epoch[ 6 of 25]-Minibatch[ 391- 400, 36.45%]: CE.SM = 1.25071106 * 10240; Err = 0.39892578 * 10240; time = 0.0408s; samplesPerSecond = 250679.3
12/20/2016 15:27:41:  Epoch[ 6 of 25]-Minibatch[ 401- 410, 37.36%]: CE.SM = 1.24126892 * 10240; Err = 0.39394531 * 10240; time = 0.0407s; samplesPerSecond = 251294.5
12/20/2016 15:27:41:  Epoch[ 6 of 25]-Minibatch[ 411- 420, 38.27%]: CE.SM = 1.25873413 * 10240; Err = 0.39296875 * 10240; time = 0.0414s; samplesPerSecond = 247080.4
12/20/2016 15:27:41:  Epoch[ 6 of 25]-Minibatch[ 421- 430, 39.18%]: CE.SM = 1.23926392 * 10240; Err = 0.39238281 * 10240; time = 0.0403s; samplesPerSecond = 254031.3
12/20/2016 15:27:41:  Epoch[ 6 of 25]-Minibatch[ 431- 440, 40.09%]: CE.SM = 1.25000000 * 10240; Err = 0.39882812 * 10240; time = 0.0433s; samplesPerSecond = 236407.7
12/20/2016 15:27:41:  Epoch[ 6 of 25]-Minibatch[ 441- 450, 41.00%]: CE.SM = 1.24388428 * 10240; Err = 0.39160156 * 10240; time = 0.0405s; samplesPerSecond = 252920.7
12/20/2016 15:27:41:  Epoch[ 6 of 25]-Minibatch[ 451- 460, 41.91%]: CE.SM = 1.24476929 * 10240; Err = 0.39423828 * 10240; time = 0.0414s; samplesPerSecond = 247181.8
12/20/2016 15:27:41:  Epoch[ 6 of 25]-Minibatch[ 461- 470, 42.82%]: CE.SM = 1.22811890 * 10240; Err = 0.38554688 * 10240; time = 0.0404s; samplesPerSecond = 253716.6
12/20/2016 15:27:42:  Epoch[ 6 of 25]-Minibatch[ 471- 480, 43.74%]: CE.SM = 1.21256714 * 10240; Err = 0.39101562 * 10240; time = 0.0409s; samplesPerSecond = 250464.7
12/20/2016 15:27:42:  Epoch[ 6 of 25]-Minibatch[ 481- 490, 44.65%]: CE.SM = 1.24614258 * 10240; Err = 0.39462891 * 10240; time = 0.0402s; samplesPerSecond = 254713.7
12/20/2016 15:27:42:  Epoch[ 6 of 25]-Minibatch[ 491- 500, 45.56%]: CE.SM = 1.25689087 * 10240; Err = 0.39833984 * 10240; time = 0.0398s; samplesPerSecond = 257286.4
12/20/2016 15:27:42:  Epoch[ 6 of 25]-Minibatch[ 501- 510, 46.47%]: CE.SM = 1.23782349 * 10240; Err = 0.38710937 * 10240; time = 0.0401s; samplesPerSecond = 255514.5
12/20/2016 15:27:42:  Epoch[ 6 of 25]-Minibatch[ 511- 520, 47.38%]: CE.SM = 1.26912231 * 10240; Err = 0.39775391 * 10240; time = 0.0401s; samplesPerSecond = 255291.6
12/20/2016 15:27:42:  Epoch[ 6 of 25]-Minibatch[ 521- 530, 48.29%]: CE.SM = 1.24829102 * 10240; Err = 0.39082031 * 10240; time = 0.0413s; samplesPerSecond = 247827.9
12/20/2016 15:27:42:  Epoch[ 6 of 25]-Minibatch[ 531- 540, 49.20%]: CE.SM = 1.25938110 * 10240; Err = 0.39384766 * 10240; time = 0.0406s; samplesPerSecond = 252011.9
12/20/2016 15:27:42:  Epoch[ 6 of 25]-Minibatch[ 541- 550, 50.11%]: CE.SM = 1.23626099 * 10240; Err = 0.38964844 * 10240; time = 0.0416s; samplesPerSecond = 246272.2
12/20/2016 15:27:42:  Epoch[ 6 of 25]-Minibatch[ 551- 560, 51.03%]: CE.SM = 1.25032349 * 10240; Err = 0.39687500 * 10240; time = 0.0417s; samplesPerSecond = 245281.2
12/20/2016 15:27:42:  Epoch[ 6 of 25]-Minibatch[ 561- 570, 51.94%]: CE.SM = 1.23029785 * 10240; Err = 0.38886719 * 10240; time = 0.0414s; samplesPerSecond = 247265.4
12/20/2016 15:27:42:  Epoch[ 6 of 25]-Minibatch[ 571- 580, 52.85%]: CE.SM = 1.22753296 * 10240; Err = 0.39462891 * 10240; time = 0.0407s; samplesPerSecond = 251387.0
12/20/2016 15:27:42:  Epoch[ 6 of 25]-Minibatch[ 581- 590, 53.76%]: CE.SM = 1.22671509 * 10240; Err = 0.38964844 * 10240; time = 0.0412s; samplesPerSecond = 248815.5
12/20/2016 15:27:42:  Epoch[ 6 of 25]-Minibatch[ 591- 600, 54.67%]: CE.SM = 1.23577271 * 10240; Err = 0.38896484 * 10240; time = 0.0519s; samplesPerSecond = 197351.9
12/20/2016 15:27:42:  Epoch[ 6 of 25]-Minibatch[ 601- 610, 55.58%]: CE.SM = 1.23674927 * 10240; Err = 0.39199219 * 10240; time = 0.0418s; samplesPerSecond = 245146.2
12/20/2016 15:27:42:  Epoch[ 6 of 25]-Minibatch[ 611- 620, 56.49%]: CE.SM = 1.21119385 * 10240; Err = 0.38789062 * 10240; time = 0.0423s; samplesPerSecond = 242006.0
12/20/2016 15:27:42:  Epoch[ 6 of 25]-Minibatch[ 621- 630, 57.40%]: CE.SM = 1.23536987 * 10240; Err = 0.39257812 * 10240; time = 0.0472s; samplesPerSecond = 217059.5
12/20/2016 15:27:42:  Epoch[ 6 of 25]-Minibatch[ 631- 640, 58.31%]: CE.SM = 1.23206177 * 10240; Err = 0.38369141 * 10240; time = 0.0456s; samplesPerSecond = 224605.7
12/20/2016 15:27:42:  Epoch[ 6 of 25]-Minibatch[ 641- 650, 59.23%]: CE.SM = 1.25471191 * 10240; Err = 0.39716797 * 10240; time = 0.0416s; samplesPerSecond = 246367.0
12/20/2016 15:27:42:  Epoch[ 6 of 25]-Minibatch[ 651- 660, 60.14%]: CE.SM = 1.23814697 * 10240; Err = 0.39189453 * 10240; time = 0.0406s; samplesPerSecond = 252067.7
12/20/2016 15:27:42:  Epoch[ 6 of 25]-Minibatch[ 661- 670, 61.05%]: CE.SM = 1.23713989 * 10240; Err = 0.38750000 * 10240; time = 0.0455s; samplesPerSecond = 224980.8
12/20/2016 15:27:42:  Epoch[ 6 of 25]-Minibatch[ 671- 680, 61.96%]: CE.SM = 1.25257568 * 10240; Err = 0.40156250 * 10240; time = 0.0436s; samplesPerSecond = 235105.0
12/20/2016 15:27:42:  Epoch[ 6 of 25]-Minibatch[ 681- 690, 62.87%]: CE.SM = 1.23487549 * 10240; Err = 0.39111328 * 10240; time = 0.0412s; samplesPerSecond = 248368.9
12/20/2016 15:27:42:  Epoch[ 6 of 25]-Minibatch[ 691- 700, 63.78%]: CE.SM = 1.25864258 * 10240; Err = 0.39687500 * 10240; time = 0.0420s; samplesPerSecond = 244047.8
12/20/2016 15:27:42:  Epoch[ 6 of 25]-Minibatch[ 701- 710, 64.69%]: CE.SM = 1.22769165 * 10240; Err = 0.39228516 * 10240; time = 0.0449s; samplesPerSecond = 228047.1
12/20/2016 15:27:43:  Epoch[ 6 of 25]-Minibatch[ 711- 720, 65.60%]: CE.SM = 1.22996216 * 10240; Err = 0.38691406 * 10240; time = 0.0435s; samplesPerSecond = 235391.5
12/20/2016 15:27:43:  Epoch[ 6 of 25]-Minibatch[ 721- 730, 66.51%]: CE.SM = 1.23941040 * 10240; Err = 0.38662109 * 10240; time = 0.0528s; samplesPerSecond = 193968.8
12/20/2016 15:27:43:  Epoch[ 6 of 25]-Minibatch[ 731- 740, 67.43%]: CE.SM = 1.23883057 * 10240; Err = 0.39121094 * 10240; time = 0.0410s; samplesPerSecond = 249646.5
12/20/2016 15:27:43:  Epoch[ 6 of 25]-Minibatch[ 741- 750, 68.34%]: CE.SM = 1.24286499 * 10240; Err = 0.39443359 * 10240; time = 0.0413s; samplesPerSecond = 247857.9
12/20/2016 15:27:43:  Epoch[ 6 of 25]-Minibatch[ 751- 760, 69.25%]: CE.SM = 1.22816772 * 10240; Err = 0.38632813 * 10240; time = 0.0445s; samplesPerSecond = 230267.6
12/20/2016 15:27:43:  Epoch[ 6 of 25]-Minibatch[ 761- 770, 70.16%]: CE.SM = 1.24255981 * 10240; Err = 0.39072266 * 10240; time = 0.0438s; samplesPerSecond = 234041.1
12/20/2016 15:27:43:  Epoch[ 6 of 25]-Minibatch[ 771- 780, 71.07%]: CE.SM = 1.24633179 * 10240; Err = 0.39033203 * 10240; time = 0.0484s; samplesPerSecond = 211373.7
12/20/2016 15:27:43:  Epoch[ 6 of 25]-Minibatch[ 781- 790, 71.98%]: CE.SM = 1.24387817 * 10240; Err = 0.39101562 * 10240; time = 0.0407s; samplesPerSecond = 251547.6
12/20/2016 15:27:43:  Epoch[ 6 of 25]-Minibatch[ 791- 800, 72.89%]: CE.SM = 1.21024170 * 10240; Err = 0.38593750 * 10240; time = 0.0439s; samplesPerSecond = 233252.1
12/20/2016 15:27:43:  Epoch[ 6 of 25]-Minibatch[ 801- 810, 73.80%]: CE.SM = 1.25200806 * 10240; Err = 0.40185547 * 10240; time = 0.0411s; samplesPerSecond = 249269.7
12/20/2016 15:27:43:  Epoch[ 6 of 25]-Minibatch[ 811- 820, 74.72%]: CE.SM = 1.20735474 * 10240; Err = 0.38544922 * 10240; time = 0.0402s; samplesPerSecond = 254599.7
12/20/2016 15:27:43:  Epoch[ 6 of 25]-Minibatch[ 821- 830, 75.63%]: CE.SM = 1.24630127 * 10240; Err = 0.39482422 * 10240; time = 0.0423s; samplesPerSecond = 241937.4
12/20/2016 15:27:43:  Epoch[ 6 of 25]-Minibatch[ 831- 840, 76.54%]: CE.SM = 1.25319824 * 10240; Err = 0.39970703 * 10240; time = 0.0424s; samplesPerSecond = 241680.4
12/20/2016 15:27:43:  Epoch[ 6 of 25]-Minibatch[ 841- 850, 77.45%]: CE.SM = 1.23615723 * 10240; Err = 0.38769531 * 10240; time = 0.0405s; samplesPerSecond = 252590.0
12/20/2016 15:27:43:  Epoch[ 6 of 25]-Minibatch[ 851- 860, 78.36%]: CE.SM = 1.22944336 * 10240; Err = 0.38681641 * 10240; time = 0.0443s; samplesPerSecond = 231203.4
12/20/2016 15:27:43:  Epoch[ 6 of 25]-Minibatch[ 861- 870, 79.27%]: CE.SM = 1.23612061 * 10240; Err = 0.39404297 * 10240; time = 0.0465s; samplesPerSecond = 220144.0
12/20/2016 15:27:43:  Epoch[ 6 of 25]-Minibatch[ 871- 880, 80.18%]: CE.SM = 1.24631348 * 10240; Err = 0.39208984 * 10240; time = 0.0424s; samplesPerSecond = 241333.0
12/20/2016 15:27:43:  Epoch[ 6 of 25]-Minibatch[ 881- 890, 81.09%]: CE.SM = 1.21187744 * 10240; Err = 0.38886719 * 10240; time = 0.0413s; samplesPerSecond = 247911.9
12/20/2016 15:27:43:  Epoch[ 6 of 25]-Minibatch[ 891- 900, 82.00%]: CE.SM = 1.25064697 * 10240; Err = 0.39414063 * 10240; time = 0.0413s; samplesPerSecond = 247749.9
12/20/2016 15:27:43:  Epoch[ 6 of 25]-Minibatch[ 901- 910, 82.92%]: CE.SM = 1.21859131 * 10240; Err = 0.38320312 * 10240; time = 0.0419s; samplesPerSecond = 244549.0
12/20/2016 15:27:43:  Epoch[ 6 of 25]-Minibatch[ 911- 920, 83.83%]: CE.SM = 1.24074707 * 10240; Err = 0.39228516 * 10240; time = 0.0410s; samplesPerSecond = 249993.9
12/20/2016 15:27:43:  Epoch[ 6 of 25]-Minibatch[ 921- 930, 84.74%]: CE.SM = 1.23259277 * 10240; Err = 0.39453125 * 10240; time = 0.0410s; samplesPerSecond = 249628.2
12/20/2016 15:27:43:  Epoch[ 6 of 25]-Minibatch[ 931- 940, 85.65%]: CE.SM = 1.23616943 * 10240; Err = 0.38818359 * 10240; time = 0.0419s; samplesPerSecond = 244111.8
12/20/2016 15:27:44:  Epoch[ 6 of 25]-Minibatch[ 941- 950, 86.56%]: CE.SM = 1.21754150 * 10240; Err = 0.38808594 * 10240; time = 0.0398s; samplesPerSecond = 256989.4
12/20/2016 15:27:44:  Epoch[ 6 of 25]-Minibatch[ 951- 960, 87.47%]: CE.SM = 1.23448486 * 10240; Err = 0.39208984 * 10240; time = 0.0409s; samplesPerSecond = 250158.8
12/20/2016 15:27:44:  Epoch[ 6 of 25]-Minibatch[ 961- 970, 88.38%]: CE.SM = 1.24284668 * 10240; Err = 0.39648438 * 10240; time = 0.0413s; samplesPerSecond = 248158.2
12/20/2016 15:27:44:  Epoch[ 6 of 25]-Minibatch[ 971- 980, 89.29%]: CE.SM = 1.23588867 * 10240; Err = 0.39267578 * 10240; time = 0.0414s; samplesPerSecond = 247462.5
12/20/2016 15:27:44:  Epoch[ 6 of 25]-Minibatch[ 981- 990, 90.21%]: CE.SM = 1.23773193 * 10240; Err = 0.39384766 * 10240; time = 0.0412s; samplesPerSecond = 248839.6
12/20/2016 15:27:44:  Epoch[ 6 of 25]-Minibatch[ 991-1000, 91.12%]: CE.SM = 1.23610840 * 10240; Err = 0.38945313 * 10240; time = 0.0410s; samplesPerSecond = 249646.5
12/20/2016 15:27:44:  Epoch[ 6 of 25]-Minibatch[1001-1010, 92.03%]: CE.SM = 1.26147461 * 10240; Err = 0.39257812 * 10240; time = 0.0414s; samplesPerSecond = 247558.3
12/20/2016 15:27:44:  Epoch[ 6 of 25]-Minibatch[1011-1020, 92.94%]: CE.SM = 1.23276367 * 10240; Err = 0.39492187 * 10240; time = 0.0415s; samplesPerSecond = 246699.4
12/20/2016 15:27:44:  Epoch[ 6 of 25]-Minibatch[1021-1030, 93.85%]: CE.SM = 1.23044434 * 10240; Err = 0.39287109 * 10240; time = 0.0406s; samplesPerSecond = 252073.9
12/20/2016 15:27:44:  Epoch[ 6 of 25]-Minibatch[1031-1040, 94.76%]: CE.SM = 1.24459229 * 10240; Err = 0.39599609 * 10240; time = 0.0418s; samplesPerSecond = 245269.5
12/20/2016 15:27:44:  Epoch[ 6 of 25]-Minibatch[1041-1050, 95.67%]: CE.SM = 1.21468506 * 10240; Err = 0.39179687 * 10240; time = 0.0414s; samplesPerSecond = 247337.0
12/20/2016 15:27:44:  Epoch[ 6 of 25]-Minibatch[1051-1060, 96.58%]: CE.SM = 1.23815918 * 10240; Err = 0.38408203 * 10240; time = 0.0409s; samplesPerSecond = 250446.4
12/20/2016 15:27:44:  Epoch[ 6 of 25]-Minibatch[1061-1070, 97.49%]: CE.SM = 1.23948975 * 10240; Err = 0.39062500 * 10240; time = 0.0416s; samplesPerSecond = 246130.2
12/20/2016 15:27:44:  Epoch[ 6 of 25]-Minibatch[1071-1080, 98.41%]: CE.SM = 1.23822021 * 10240; Err = 0.39697266 * 10240; time = 0.0420s; samplesPerSecond = 244041.9
12/20/2016 15:27:44:  Epoch[ 6 of 25]-Minibatch[1081-1090, 99.32%]: CE.SM = 1.21701660 * 10240; Err = 0.38378906 * 10240; time = 0.0412s; samplesPerSecond = 248549.7
12/20/2016 15:27:44: Finished Epoch[ 6 of 25]: [Training] CE.SM = 1.24147333 * 1124823; Err = 0.39274979 * 1124823; totalSamplesSeen = 6748938; learningRatePerSample = 0.003125; epochTime=4.82279s
12/20/2016 15:27:44: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn.6'

12/20/2016 15:27:44: Starting Epoch 7: learning rate per sample = 0.003125  effective momentum = 0.900000  momentum as time constant = 9719.0 samples

12/20/2016 15:27:44: Starting minibatch loop.
12/20/2016 15:27:44:  Epoch[ 7 of 25]-Minibatch[   1-  10, 0.91%]: CE.SM = 1.16810532 * 10240; Err = 0.36787109 * 10240; time = 0.0491s; samplesPerSecond = 208596.5
12/20/2016 15:27:44:  Epoch[ 7 of 25]-Minibatch[  11-  20, 1.82%]: CE.SM = 1.19715309 * 10240; Err = 0.38291016 * 10240; time = 0.0410s; samplesPerSecond = 249658.7
12/20/2016 15:27:45:  Epoch[ 7 of 25]-Minibatch[  21-  30, 2.73%]: CE.SM = 1.19982262 * 10240; Err = 0.37968750 * 10240; time = 0.0406s; samplesPerSecond = 252011.9
12/20/2016 15:27:45:  Epoch[ 7 of 25]-Minibatch[  31-  40, 3.64%]: CE.SM = 1.17648392 * 10240; Err = 0.37070313 * 10240; time = 0.0410s; samplesPerSecond = 250061.1
12/20/2016 15:27:45:  Epoch[ 7 of 25]-Minibatch[  41-  50, 4.56%]: CE.SM = 1.20368462 * 10240; Err = 0.38525391 * 10240; time = 0.0407s; samplesPerSecond = 251646.5
12/20/2016 15:27:45:  Epoch[ 7 of 25]-Minibatch[  51-  60, 5.47%]: CE.SM = 1.20194969 * 10240; Err = 0.38417969 * 10240; time = 0.0410s; samplesPerSecond = 249676.9
12/20/2016 15:27:45:  Epoch[ 7 of 25]-Minibatch[  61-  70, 6.38%]: CE.SM = 1.21393433 * 10240; Err = 0.38994141 * 10240; time = 0.0403s; samplesPerSecond = 254220.5
12/20/2016 15:27:45:  Epoch[ 7 of 25]-Minibatch[  71-  80, 7.29%]: CE.SM = 1.20627670 * 10240; Err = 0.38945313 * 10240; time = 0.0407s; samplesPerSecond = 251368.5
12/20/2016 15:27:45:  Epoch[ 7 of 25]-Minibatch[  81-  90, 8.20%]: CE.SM = 1.21675415 * 10240; Err = 0.38349609 * 10240; time = 0.0404s; samplesPerSecond = 253358.7
12/20/2016 15:27:45:  Epoch[ 7 of 25]-Minibatch[  91- 100, 9.11%]: CE.SM = 1.22775345 * 10240; Err = 0.39140625 * 10240; time = 0.0408s; samplesPerSecond = 251134.3
12/20/2016 15:27:45:  Epoch[ 7 of 25]-Minibatch[ 101- 110, 10.02%]: CE.SM = 1.21372604 * 10240; Err = 0.38535156 * 10240; time = 0.0408s; samplesPerSecond = 251220.5
12/20/2016 15:27:45:  Epoch[ 7 of 25]-Minibatch[ 111- 120, 10.93%]: CE.SM = 1.19726105 * 10240; Err = 0.37968750 * 10240; time = 0.0416s; samplesPerSecond = 246313.7
12/20/2016 15:27:45:  Epoch[ 7 of 25]-Minibatch[ 121- 130, 11.85%]: CE.SM = 1.22784424 * 10240; Err = 0.38593750 * 10240; time = 0.0409s; samplesPerSecond = 250299.4
12/20/2016 15:27:45:  Epoch[ 7 of 25]-Minibatch[ 131- 140, 12.76%]: CE.SM = 1.21406250 * 10240; Err = 0.38173828 * 10240; time = 0.0409s; samplesPerSecond = 250213.8
12/20/2016 15:27:45:  Epoch[ 7 of 25]-Minibatch[ 141- 150, 13.67%]: CE.SM = 1.19874878 * 10240; Err = 0.37851563 * 10240; time = 0.0405s; samplesPerSecond = 252939.4
12/20/2016 15:27:45:  Epoch[ 7 of 25]-Minibatch[ 151- 160, 14.58%]: CE.SM = 1.20877075 * 10240; Err = 0.38378906 * 10240; time = 0.0408s; samplesPerSecond = 250703.9
12/20/2016 15:27:45:  Epoch[ 7 of 25]-Minibatch[ 161- 170, 15.49%]: CE.SM = 1.20868530 * 10240; Err = 0.38320312 * 10240; time = 0.0410s; samplesPerSecond = 249780.5
12/20/2016 15:27:45:  Epoch[ 7 of 25]-Minibatch[ 171- 180, 16.40%]: CE.SM = 1.18462677 * 10240; Err = 0.38193359 * 10240; time = 0.0401s; samplesPerSecond = 255100.8
12/20/2016 15:27:45:  Epoch[ 7 of 25]-Minibatch[ 181- 190, 17.31%]: CE.SM = 1.21729431 * 10240; Err = 0.38808594 * 10240; time = 0.0418s; samplesPerSecond = 245140.3
12/20/2016 15:27:45:  Epoch[ 7 of 25]-Minibatch[ 191- 200, 18.22%]: CE.SM = 1.19983063 * 10240; Err = 0.38222656 * 10240; time = 0.0407s; samplesPerSecond = 251535.2
12/20/2016 15:27:45:  Epoch[ 7 of 25]-Minibatch[ 201- 210, 19.13%]: CE.SM = 1.21074677 * 10240; Err = 0.38798828 * 10240; time = 0.0419s; samplesPerSecond = 244642.5
12/20/2016 15:27:45:  Epoch[ 7 of 25]-Minibatch[ 211- 220, 20.05%]: CE.SM = 1.20926208 * 10240; Err = 0.38671875 * 10240; time = 0.0425s; samplesPerSecond = 241100.0
12/20/2016 15:27:45:  Epoch[ 7 of 25]-Minibatch[ 221- 230, 20.96%]: CE.SM = 1.21681519 * 10240; Err = 0.38574219 * 10240; time = 0.0467s; samplesPerSecond = 219234.4
12/20/2016 15:27:45:  Epoch[ 7 of 25]-Minibatch[ 231- 240, 21.87%]: CE.SM = 1.20926819 * 10240; Err = 0.39072266 * 10240; time = 0.0408s; samplesPerSecond = 251041.9
12/20/2016 15:27:45:  Epoch[ 7 of 25]-Minibatch[ 241- 250, 22.78%]: CE.SM = 1.21136169 * 10240; Err = 0.38359375 * 10240; time = 0.0401s; samplesPerSecond = 255208.9
12/20/2016 15:27:45:  Epoch[ 7 of 25]-Minibatch[ 251- 260, 23.69%]: CE.SM = 1.19963074 * 10240; Err = 0.37763672 * 10240; time = 0.0407s; samplesPerSecond = 251894.1
12/20/2016 15:27:46:  Epoch[ 7 of 25]-Minibatch[ 261- 270, 24.60%]: CE.SM = 1.21356506 * 10240; Err = 0.38437500 * 10240; time = 0.0408s; samplesPerSecond = 251060.4
12/20/2016 15:27:46:  Epoch[ 7 of 25]-Minibatch[ 271- 280, 25.51%]: CE.SM = 1.21652832 * 10240; Err = 0.38554688 * 10240; time = 0.0407s; samplesPerSecond = 251819.8
12/20/2016 15:27:46:  Epoch[ 7 of 25]-Minibatch[ 281- 290, 26.42%]: CE.SM = 1.20802917 * 10240; Err = 0.37929687 * 10240; time = 0.0407s; samplesPerSecond = 251387.0
12/20/2016 15:27:46:  Epoch[ 7 of 25]-Minibatch[ 291- 300, 27.33%]: CE.SM = 1.22272034 * 10240; Err = 0.38710937 * 10240; time = 0.0403s; samplesPerSecond = 253899.0
12/20/2016 15:27:46:  Epoch[ 7 of 25]-Minibatch[ 301- 310, 28.25%]: CE.SM = 1.20689697 * 10240; Err = 0.38691406 * 10240; time = 0.0410s; samplesPerSecond = 249865.8
12/20/2016 15:27:46:  Epoch[ 7 of 25]-Minibatch[ 311- 320, 29.16%]: CE.SM = 1.19957581 * 10240; Err = 0.38457031 * 10240; time = 0.0412s; samplesPerSecond = 248694.6
12/20/2016 15:27:46:  Epoch[ 7 of 25]-Minibatch[ 321- 330, 30.07%]: CE.SM = 1.22283020 * 10240; Err = 0.38115234 * 10240; time = 0.0410s; samplesPerSecond = 249908.5
12/20/2016 15:27:46:  Epoch[ 7 of 25]-Minibatch[ 331- 340, 30.98%]: CE.SM = 1.20003357 * 10240; Err = 0.37978516 * 10240; time = 0.0399s; samplesPerSecond = 256461.6
12/20/2016 15:27:46:  Epoch[ 7 of 25]-Minibatch[ 341- 350, 31.89%]: CE.SM = 1.21448059 * 10240; Err = 0.38349609 * 10240; time = 0.0412s; samplesPerSecond = 248555.8
12/20/2016 15:27:46:  Epoch[ 7 of 25]-Minibatch[ 351- 360, 32.80%]: CE.SM = 1.20859375 * 10240; Err = 0.37978516 * 10240; time = 0.0416s; samplesPerSecond = 246420.4
12/20/2016 15:27:46:  Epoch[ 7 of 25]-Minibatch[ 361- 370, 33.71%]: CE.SM = 1.19839478 * 10240; Err = 0.38291016 * 10240; time = 0.0410s; samplesPerSecond = 249896.3
12/20/2016 15:27:46:  Epoch[ 7 of 25]-Minibatch[ 371- 380, 34.62%]: CE.SM = 1.20303040 * 10240; Err = 0.38867188 * 10240; time = 0.0404s; samplesPerSecond = 253678.8
12/20/2016 15:27:46:  Epoch[ 7 of 25]-Minibatch[ 381- 390, 35.54%]: CE.SM = 1.20429993 * 10240; Err = 0.38583984 * 10240; time = 0.0404s; samplesPerSecond = 253521.8
12/20/2016 15:27:46:  Epoch[ 7 of 25]-Minibatch[ 391- 400, 36.45%]: CE.SM = 1.21196289 * 10240; Err = 0.38808594 * 10240; time = 0.0408s; samplesPerSecond = 250814.4
12/20/2016 15:27:46:  Epoch[ 7 of 25]-Minibatch[ 401- 410, 37.36%]: CE.SM = 1.20273437 * 10240; Err = 0.37929687 * 10240; time = 0.0402s; samplesPerSecond = 254542.7
12/20/2016 15:27:46:  Epoch[ 7 of 25]-Minibatch[ 411- 420, 38.27%]: CE.SM = 1.21734314 * 10240; Err = 0.38789062 * 10240; time = 0.0410s; samplesPerSecond = 249676.9
12/20/2016 15:27:46:  Epoch[ 7 of 25]-Minibatch[ 421- 430, 39.18%]: CE.SM = 1.20904846 * 10240; Err = 0.38515625 * 10240; time = 0.0415s; samplesPerSecond = 246770.8
12/20/2016 15:27:46:  Epoch[ 7 of 25]-Minibatch[ 431- 440, 40.09%]: CE.SM = 1.19996948 * 10240; Err = 0.38642578 * 10240; time = 0.0412s; samplesPerSecond = 248447.2
12/20/2016 15:27:46:  Epoch[ 7 of 25]-Minibatch[ 441- 450, 41.00%]: CE.SM = 1.18839722 * 10240; Err = 0.38291016 * 10240; time = 0.0428s; samplesPerSecond = 239492.9
12/20/2016 15:27:46:  Epoch[ 7 of 25]-Minibatch[ 451- 460, 41.91%]: CE.SM = 1.20225220 * 10240; Err = 0.37968750 * 10240; time = 0.0409s; samplesPerSecond = 250281.1
12/20/2016 15:27:46:  Epoch[ 7 of 25]-Minibatch[ 461- 470, 42.82%]: CE.SM = 1.19744873 * 10240; Err = 0.37519531 * 10240; time = 0.0403s; samplesPerSecond = 254403.6
12/20/2016 15:27:46:  Epoch[ 7 of 25]-Minibatch[ 471- 480, 43.74%]: CE.SM = 1.19553833 * 10240; Err = 0.37285156 * 10240; time = 0.0399s; samplesPerSecond = 256487.3
12/20/2016 15:27:46:  Epoch[ 7 of 25]-Minibatch[ 481- 490, 44.65%]: CE.SM = 1.20729370 * 10240; Err = 0.38466797 * 10240; time = 0.0414s; samplesPerSecond = 247390.8
12/20/2016 15:27:46:  Epoch[ 7 of 25]-Minibatch[ 491- 500, 45.56%]: CE.SM = 1.20002441 * 10240; Err = 0.38759766 * 10240; time = 0.0486s; samplesPerSecond = 210487.4
12/20/2016 15:27:47:  Epoch[ 7 of 25]-Minibatch[ 501- 510, 46.47%]: CE.SM = 1.18764648 * 10240; Err = 0.38076172 * 10240; time = 0.0424s; samplesPerSecond = 241629.1
12/20/2016 15:27:47:  Epoch[ 7 of 25]-Minibatch[ 511- 520, 47.38%]: CE.SM = 1.21516113 * 10240; Err = 0.38349609 * 10240; time = 0.0424s; samplesPerSecond = 241406.9
12/20/2016 15:27:47:  Epoch[ 7 of 25]-Minibatch[ 521- 530, 48.29%]: CE.SM = 1.20580444 * 10240; Err = 0.38281250 * 10240; time = 0.0408s; samplesPerSecond = 251263.7
12/20/2016 15:27:47:  Epoch[ 7 of 25]-Minibatch[ 531- 540, 49.20%]: CE.SM = 1.18738403 * 10240; Err = 0.37841797 * 10240; time = 0.0409s; samplesPerSecond = 250599.6
12/20/2016 15:27:47:  Epoch[ 7 of 25]-Minibatch[ 541- 550, 50.11%]: CE.SM = 1.18766479 * 10240; Err = 0.37802734 * 10240; time = 0.0416s; samplesPerSecond = 246100.6
12/20/2016 15:27:47:  Epoch[ 7 of 25]-Minibatch[ 551- 560, 51.03%]: CE.SM = 1.20473022 * 10240; Err = 0.38613281 * 10240; time = 0.0414s; samplesPerSecond = 247408.7
12/20/2016 15:27:47:  Epoch[ 7 of 25]-Minibatch[ 561- 570, 51.94%]: CE.SM = 1.20039063 * 10240; Err = 0.38271484 * 10240; time = 0.0409s; samplesPerSecond = 250330.0
12/20/2016 15:27:47:  Epoch[ 7 of 25]-Minibatch[ 571- 580, 52.85%]: CE.SM = 1.23986206 * 10240; Err = 0.38574219 * 10240; time = 0.0423s; samplesPerSecond = 242046.0
12/20/2016 15:27:47:  Epoch[ 7 of 25]-Minibatch[ 581- 590, 53.76%]: CE.SM = 1.20167236 * 10240; Err = 0.38115234 * 10240; time = 0.0415s; samplesPerSecond = 246949.3
12/20/2016 15:27:47:  Epoch[ 7 of 25]-Minibatch[ 591- 600, 54.67%]: CE.SM = 1.17788086 * 10240; Err = 0.37441406 * 10240; time = 0.0408s; samplesPerSecond = 250685.5
12/20/2016 15:27:47:  Epoch[ 7 of 25]-Minibatch[ 601- 610, 55.58%]: CE.SM = 1.22758789 * 10240; Err = 0.38652344 * 10240; time = 0.0410s; samplesPerSecond = 249865.8
12/20/2016 15:27:47:  Epoch[ 7 of 25]-Minibatch[ 611- 620, 56.49%]: CE.SM = 1.19057007 * 10240; Err = 0.37431641 * 10240; time = 0.0417s; samplesPerSecond = 245545.9
12/20/2016 15:27:47:  Epoch[ 7 of 25]-Minibatch[ 621- 630, 57.40%]: CE.SM = 1.17855225 * 10240; Err = 0.37578125 * 10240; time = 0.0411s; samplesPerSecond = 248948.5
12/20/2016 15:27:47:  Epoch[ 7 of 25]-Minibatch[ 631- 640, 58.31%]: CE.SM = 1.22963867 * 10240; Err = 0.39453125 * 10240; time = 0.0414s; samplesPerSecond = 247050.6
12/20/2016 15:27:47:  Epoch[ 7 of 25]-Minibatch[ 641- 650, 59.23%]: CE.SM = 1.20941772 * 10240; Err = 0.37851563 * 10240; time = 0.0478s; samplesPerSecond = 214028.9
12/20/2016 15:27:47:  Epoch[ 7 of 25]-Minibatch[ 651- 660, 60.14%]: CE.SM = 1.17030029 * 10240; Err = 0.37099609 * 10240; time = 0.0454s; samplesPerSecond = 225789.4
12/20/2016 15:27:47:  Epoch[ 7 of 25]-Minibatch[ 661- 670, 61.05%]: CE.SM = 1.18946533 * 10240; Err = 0.37441406 * 10240; time = 0.0430s; samplesPerSecond = 238222.6
12/20/2016 15:27:47:  Epoch[ 7 of 25]-Minibatch[ 671- 680, 61.96%]: CE.SM = 1.19066162 * 10240; Err = 0.37529297 * 10240; time = 0.0415s; samplesPerSecond = 246776.7
12/20/2016 15:27:47:  Epoch[ 7 of 25]-Minibatch[ 681- 690, 62.87%]: CE.SM = 1.18230591 * 10240; Err = 0.37646484 * 10240; time = 0.0416s; samplesPerSecond = 246195.3
12/20/2016 15:27:47:  Epoch[ 7 of 25]-Minibatch[ 691- 700, 63.78%]: CE.SM = 1.22643433 * 10240; Err = 0.39296875 * 10240; time = 0.0417s; samplesPerSecond = 245387.0
12/20/2016 15:27:47:  Epoch[ 7 of 25]-Minibatch[ 701- 710, 64.69%]: CE.SM = 1.22045898 * 10240; Err = 0.39023438 * 10240; time = 0.0419s; samplesPerSecond = 244263.2
12/20/2016 15:27:47:  Epoch[ 7 of 25]-Minibatch[ 711- 720, 65.60%]: CE.SM = 1.20562134 * 10240; Err = 0.37822266 * 10240; time = 0.0406s; samplesPerSecond = 252490.4
12/20/2016 15:27:47:  Epoch[ 7 of 25]-Minibatch[ 721- 730, 66.51%]: CE.SM = 1.23485718 * 10240; Err = 0.38876953 * 10240; time = 0.0403s; samplesPerSecond = 253974.6
12/20/2016 15:27:47:  Epoch[ 7 of 25]-Minibatch[ 731- 740, 67.43%]: CE.SM = 1.22781982 * 10240; Err = 0.38457031 * 10240; time = 0.0426s; samplesPerSecond = 240658.0
12/20/2016 15:27:48:  Epoch[ 7 of 25]-Minibatch[ 741- 750, 68.34%]: CE.SM = 1.21111450 * 10240; Err = 0.38466797 * 10240; time = 0.0415s; samplesPerSecond = 246461.9
12/20/2016 15:27:48:  Epoch[ 7 of 25]-Minibatch[ 751- 760, 69.25%]: CE.SM = 1.20229492 * 10240; Err = 0.38154297 * 10240; time = 0.0416s; samplesPerSecond = 246367.0
12/20/2016 15:27:48:  Epoch[ 7 of 25]-Minibatch[ 761- 770, 70.16%]: CE.SM = 1.20404053 * 10240; Err = 0.38769531 * 10240; time = 0.0411s; samplesPerSecond = 248936.4
12/20/2016 15:27:48:  Epoch[ 7 of 25]-Minibatch[ 771- 780, 71.07%]: CE.SM = 1.22680664 * 10240; Err = 0.39316406 * 10240; time = 0.0399s; samplesPerSecond = 256551.6
12/20/2016 15:27:48:  Epoch[ 7 of 25]-Minibatch[ 781- 790, 71.98%]: CE.SM = 1.18895874 * 10240; Err = 0.37695312 * 10240; time = 0.0454s; samplesPerSecond = 225416.6
12/20/2016 15:27:48:  Epoch[ 7 of 25]-Minibatch[ 791- 800, 72.89%]: CE.SM = 1.19985962 * 10240; Err = 0.37939453 * 10240; time = 0.0422s; samplesPerSecond = 242430.0
12/20/2016 15:27:48:  Epoch[ 7 of 25]-Minibatch[ 801- 810, 73.80%]: CE.SM = 1.19721069 * 10240; Err = 0.38603516 * 10240; time = 0.0477s; samplesPerSecond = 214495.2
12/20/2016 15:27:48:  Epoch[ 7 of 25]-Minibatch[ 811- 820, 74.72%]: CE.SM = 1.18436890 * 10240; Err = 0.38105469 * 10240; time = 0.0464s; samplesPerSecond = 220689.7
12/20/2016 15:27:48:  Epoch[ 7 of 25]-Minibatch[ 821- 830, 75.63%]: CE.SM = 1.20135498 * 10240; Err = 0.37871094 * 10240; time = 0.0436s; samplesPerSecond = 235115.8
12/20/2016 15:27:48:  Epoch[ 7 of 25]-Minibatch[ 831- 840, 76.54%]: CE.SM = 1.19029541 * 10240; Err = 0.37500000 * 10240; time = 0.0432s; samplesPerSecond = 237152.3
12/20/2016 15:27:48:  Epoch[ 7 of 25]-Minibatch[ 841- 850, 77.45%]: CE.SM = 1.19329224 * 10240; Err = 0.37998047 * 10240; time = 0.0417s; samplesPerSecond = 245451.7
12/20/2016 15:27:48:  Epoch[ 7 of 25]-Minibatch[ 851- 860, 78.36%]: CE.SM = 1.20316772 * 10240; Err = 0.38251953 * 10240; time = 0.0408s; samplesPerSecond = 250998.8
12/20/2016 15:27:48:  Epoch[ 7 of 25]-Minibatch[ 861- 870, 79.27%]: CE.SM = 1.18341064 * 10240; Err = 0.37802734 * 10240; time = 0.0419s; samplesPerSecond = 244204.9
12/20/2016 15:27:48:  Epoch[ 7 of 25]-Minibatch[ 871- 880, 80.18%]: CE.SM = 1.20324707 * 10240; Err = 0.38642578 * 10240; time = 0.0411s; samplesPerSecond = 249166.6
12/20/2016 15:27:48:  Epoch[ 7 of 25]-Minibatch[ 881- 890, 81.09%]: CE.SM = 1.18505859 * 10240; Err = 0.37900391 * 10240; time = 0.0448s; samplesPerSecond = 228653.1
12/20/2016 15:27:48:  Epoch[ 7 of 25]-Minibatch[ 891- 900, 82.00%]: CE.SM = 1.24036865 * 10240; Err = 0.39482422 * 10240; time = 0.0407s; samplesPerSecond = 251566.1
12/20/2016 15:27:48:  Epoch[ 7 of 25]-Minibatch[ 901- 910, 82.92%]: CE.SM = 1.18140869 * 10240; Err = 0.37636719 * 10240; time = 0.0412s; samplesPerSecond = 248791.3
12/20/2016 15:27:48:  Epoch[ 7 of 25]-Minibatch[ 911- 920, 83.83%]: CE.SM = 1.21259766 * 10240; Err = 0.38759766 * 10240; time = 0.0418s; samplesPerSecond = 244882.3
12/20/2016 15:27:48:  Epoch[ 7 of 25]-Minibatch[ 921- 930, 84.74%]: CE.SM = 1.19184570 * 10240; Err = 0.38281250 * 10240; time = 0.0414s; samplesPerSecond = 247062.5
12/20/2016 15:27:48:  Epoch[ 7 of 25]-Minibatch[ 931- 940, 85.65%]: CE.SM = 1.21955566 * 10240; Err = 0.38984375 * 10240; time = 0.0412s; samplesPerSecond = 248495.4
12/20/2016 15:27:48:  Epoch[ 7 of 25]-Minibatch[ 941- 950, 86.56%]: CE.SM = 1.20374756 * 10240; Err = 0.38320312 * 10240; time = 0.0415s; samplesPerSecond = 246503.5
12/20/2016 15:27:48:  Epoch[ 7 of 25]-Minibatch[ 951- 960, 87.47%]: CE.SM = 1.22346191 * 10240; Err = 0.38564453 * 10240; time = 0.0422s; samplesPerSecond = 242723.0
12/20/2016 15:27:48:  Epoch[ 7 of 25]-Minibatch[ 961- 970, 88.38%]: CE.SM = 1.19930420 * 10240; Err = 0.38798828 * 10240; time = 0.0426s; samplesPerSecond = 240178.3
12/20/2016 15:27:49:  Epoch[ 7 of 25]-Minibatch[ 971- 980, 89.29%]: CE.SM = 1.20450439 * 10240; Err = 0.38457031 * 10240; time = 0.0412s; samplesPerSecond = 248616.1
12/20/2016 15:27:49:  Epoch[ 7 of 25]-Minibatch[ 981- 990, 90.21%]: CE.SM = 1.21894531 * 10240; Err = 0.38291016 * 10240; time = 0.0416s; samplesPerSecond = 246390.8
12/20/2016 15:27:49:  Epoch[ 7 of 25]-Minibatch[ 991-1000, 91.12%]: CE.SM = 1.21588135 * 10240; Err = 0.38681641 * 10240; time = 0.0403s; samplesPerSecond = 254062.8
12/20/2016 15:27:49:  Epoch[ 7 of 25]-Minibatch[1001-1010, 92.03%]: CE.SM = 1.20573730 * 10240; Err = 0.37939453 * 10240; time = 0.0420s; samplesPerSecond = 243896.6
12/20/2016 15:27:49:  Epoch[ 7 of 25]-Minibatch[1011-1020, 92.94%]: CE.SM = 1.20490723 * 10240; Err = 0.39082031 * 10240; time = 0.0402s; samplesPerSecond = 254999.1
12/20/2016 15:27:49:  Epoch[ 7 of 25]-Minibatch[1021-1030, 93.85%]: CE.SM = 1.17261963 * 10240; Err = 0.37353516 * 10240; time = 0.0406s; samplesPerSecond = 252092.6
12/20/2016 15:27:49:  Epoch[ 7 of 25]-Minibatch[1031-1040, 94.76%]: CE.SM = 1.18442383 * 10240; Err = 0.37841797 * 10240; time = 0.0417s; samplesPerSecond = 245328.2
12/20/2016 15:27:49:  Epoch[ 7 of 25]-Minibatch[1041-1050, 95.67%]: CE.SM = 1.20201416 * 10240; Err = 0.38476562 * 10240; time = 0.0401s; samplesPerSecond = 255368.0
12/20/2016 15:27:49:  Epoch[ 7 of 25]-Minibatch[1051-1060, 96.58%]: CE.SM = 1.22211914 * 10240; Err = 0.38652344 * 10240; time = 0.0412s; samplesPerSecond = 248447.2
12/20/2016 15:27:49:  Epoch[ 7 of 25]-Minibatch[1061-1070, 97.49%]: CE.SM = 1.18159180 * 10240; Err = 0.36972656 * 10240; time = 0.0413s; samplesPerSecond = 247672.0
12/20/2016 15:27:49:  Epoch[ 7 of 25]-Minibatch[1071-1080, 98.41%]: CE.SM = 1.21439209 * 10240; Err = 0.38242188 * 10240; time = 0.0403s; samplesPerSecond = 254226.8
12/20/2016 15:27:49:  Epoch[ 7 of 25]-Minibatch[1081-1090, 99.32%]: CE.SM = 1.16754150 * 10240; Err = 0.37753906 * 10240; time = 0.0405s; samplesPerSecond = 252633.7
12/20/2016 15:27:49: Finished Epoch[ 7 of 25]: [Training] CE.SM = 1.20391030 * 1124823; Err = 0.38275355 * 1124823; totalSamplesSeen = 7873761; learningRatePerSample = 0.003125; epochTime=4.77806s
12/20/2016 15:27:49: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn.7'

12/20/2016 15:27:49: Starting Epoch 8: learning rate per sample = 0.003125  effective momentum = 0.900000  momentum as time constant = 9719.0 samples

12/20/2016 15:27:49: Starting minibatch loop.
12/20/2016 15:27:49:  Epoch[ 8 of 25]-Minibatch[   1-  10, 0.91%]: CE.SM = 1.15187702 * 10240; Err = 0.36542969 * 10240; time = 0.0500s; samplesPerSecond = 204652.7
12/20/2016 15:27:49:  Epoch[ 8 of 25]-Minibatch[  11-  20, 1.82%]: CE.SM = 1.15552311 * 10240; Err = 0.37148437 * 10240; time = 0.0408s; samplesPerSecond = 250783.7
12/20/2016 15:27:49:  Epoch[ 8 of 25]-Minibatch[  21-  30, 2.73%]: CE.SM = 1.17213783 * 10240; Err = 0.37294922 * 10240; time = 0.0406s; samplesPerSecond = 252247.8
12/20/2016 15:27:49:  Epoch[ 8 of 25]-Minibatch[  31-  40, 3.64%]: CE.SM = 1.15169029 * 10240; Err = 0.36757812 * 10240; time = 0.0411s; samplesPerSecond = 249063.6
12/20/2016 15:27:49:  Epoch[ 8 of 25]-Minibatch[  41-  50, 4.56%]: CE.SM = 1.15237160 * 10240; Err = 0.36767578 * 10240; time = 0.0409s; samplesPerSecond = 250348.4
12/20/2016 15:27:50:  Epoch[ 8 of 25]-Minibatch[  51-  60, 5.47%]: CE.SM = 1.17786407 * 10240; Err = 0.37617187 * 10240; time = 0.0412s; samplesPerSecond = 248465.3
12/20/2016 15:27:50:  Epoch[ 8 of 25]-Minibatch[  61-  70, 6.38%]: CE.SM = 1.17709427 * 10240; Err = 0.37783203 * 10240; time = 0.0418s; samplesPerSecond = 244876.5
12/20/2016 15:27:50:  Epoch[ 8 of 25]-Minibatch[  71-  80, 7.29%]: CE.SM = 1.15776978 * 10240; Err = 0.37226562 * 10240; time = 0.0405s; samplesPerSecond = 252534.0
12/20/2016 15:27:50:  Epoch[ 8 of 25]-Minibatch[  81-  90, 8.20%]: CE.SM = 1.16990738 * 10240; Err = 0.37724609 * 10240; time = 0.0420s; samplesPerSecond = 243786.3
12/20/2016 15:27:50:  Epoch[ 8 of 25]-Minibatch[  91- 100, 9.11%]: CE.SM = 1.16154099 * 10240; Err = 0.37607422 * 10240; time = 0.0412s; samplesPerSecond = 248350.8
12/20/2016 15:27:50:  Epoch[ 8 of 25]-Minibatch[ 101- 110, 10.02%]: CE.SM = 1.17210236 * 10240; Err = 0.37900391 * 10240; time = 0.0415s; samplesPerSecond = 246556.9
12/20/2016 15:27:50:  Epoch[ 8 of 25]-Minibatch[ 111- 120, 10.93%]: CE.SM = 1.15027390 * 10240; Err = 0.37060547 * 10240; time = 0.0490s; samplesPerSecond = 209099.1
12/20/2016 15:27:50:  Epoch[ 8 of 25]-Minibatch[ 121- 130, 11.85%]: CE.SM = 1.19203186 * 10240; Err = 0.38095703 * 10240; time = 0.0512s; samplesPerSecond = 199945.3
12/20/2016 15:27:50:  Epoch[ 8 of 25]-Minibatch[ 131- 140, 12.76%]: CE.SM = 1.15579376 * 10240; Err = 0.36875000 * 10240; time = 0.0495s; samplesPerSecond = 207036.0
12/20/2016 15:27:50:  Epoch[ 8 of 25]-Minibatch[ 141- 150, 13.67%]: CE.SM = 1.17824860 * 10240; Err = 0.37880859 * 10240; time = 0.0418s; samplesPerSecond = 245075.7
12/20/2016 15:27:50:  Epoch[ 8 of 25]-Minibatch[ 151- 160, 14.58%]: CE.SM = 1.14957581 * 10240; Err = 0.36718750 * 10240; time = 0.0419s; samplesPerSecond = 244321.4
12/20/2016 15:27:50:  Epoch[ 8 of 25]-Minibatch[ 161- 170, 15.49%]: CE.SM = 1.18206329 * 10240; Err = 0.37675781 * 10240; time = 0.0415s; samplesPerSecond = 246723.2
12/20/2016 15:27:50:  Epoch[ 8 of 25]-Minibatch[ 171- 180, 16.40%]: CE.SM = 1.16797485 * 10240; Err = 0.36660156 * 10240; time = 0.0423s; samplesPerSecond = 242137.6
12/20/2016 15:27:50:  Epoch[ 8 of 25]-Minibatch[ 181- 190, 17.31%]: CE.SM = 1.19520264 * 10240; Err = 0.37470703 * 10240; time = 0.0433s; samplesPerSecond = 236445.9
12/20/2016 15:27:50:  Epoch[ 8 of 25]-Minibatch[ 191- 200, 18.22%]: CE.SM = 1.17863922 * 10240; Err = 0.37753906 * 10240; time = 0.0412s; samplesPerSecond = 248730.8
12/20/2016 15:27:50:  Epoch[ 8 of 25]-Minibatch[ 201- 210, 19.13%]: CE.SM = 1.18460541 * 10240; Err = 0.38271484 * 10240; time = 0.0411s; samplesPerSecond = 249391.1
12/20/2016 15:27:50:  Epoch[ 8 of 25]-Minibatch[ 211- 220, 20.05%]: CE.SM = 1.18921051 * 10240; Err = 0.37724609 * 10240; time = 0.0412s; samplesPerSecond = 248284.6
12/20/2016 15:27:50:  Epoch[ 8 of 25]-Minibatch[ 221- 230, 20.96%]: CE.SM = 1.18252258 * 10240; Err = 0.37666016 * 10240; time = 0.0415s; samplesPerSecond = 246574.7
12/20/2016 15:27:50:  Epoch[ 8 of 25]-Minibatch[ 231- 240, 21.87%]: CE.SM = 1.17423706 * 10240; Err = 0.38125000 * 10240; time = 0.0410s; samplesPerSecond = 249488.4
12/20/2016 15:27:50:  Epoch[ 8 of 25]-Minibatch[ 241- 250, 22.78%]: CE.SM = 1.15256958 * 10240; Err = 0.37294922 * 10240; time = 0.0410s; samplesPerSecond = 249847.5
12/20/2016 15:27:50:  Epoch[ 8 of 25]-Minibatch[ 251- 260, 23.69%]: CE.SM = 1.18058167 * 10240; Err = 0.37324219 * 10240; time = 0.0412s; samplesPerSecond = 248272.5
12/20/2016 15:27:50:  Epoch[ 8 of 25]-Minibatch[ 261- 270, 24.60%]: CE.SM = 1.17540894 * 10240; Err = 0.38144531 * 10240; time = 0.0425s; samplesPerSecond = 240703.3
12/20/2016 15:27:50:  Epoch[ 8 of 25]-Minibatch[ 271- 280, 25.51%]: CE.SM = 1.17577515 * 10240; Err = 0.36767578 * 10240; time = 0.0402s; samplesPerSecond = 254536.4
12/20/2016 15:27:50:  Epoch[ 8 of 25]-Minibatch[ 281- 290, 26.42%]: CE.SM = 1.17970581 * 10240; Err = 0.37734375 * 10240; time = 0.0406s; samplesPerSecond = 252179.5
12/20/2016 15:27:51:  Epoch[ 8 of 25]-Minibatch[ 291- 300, 27.33%]: CE.SM = 1.15234375 * 10240; Err = 0.36718750 * 10240; time = 0.0413s; samplesPerSecond = 248170.2
12/20/2016 15:27:51:  Epoch[ 8 of 25]-Minibatch[ 301- 310, 28.25%]: CE.SM = 1.17447815 * 10240; Err = 0.36767578 * 10240; time = 0.0419s; samplesPerSecond = 244554.8
12/20/2016 15:27:51:  Epoch[ 8 of 25]-Minibatch[ 311- 320, 29.16%]: CE.SM = 1.17292480 * 10240; Err = 0.37578125 * 10240; time = 0.0401s; samplesPerSecond = 255361.6
12/20/2016 15:27:51:  Epoch[ 8 of 25]-Minibatch[ 321- 330, 30.07%]: CE.SM = 1.16473083 * 10240; Err = 0.36826172 * 10240; time = 0.0496s; samplesPerSecond = 206534.9
12/20/2016 15:27:51:  Epoch[ 8 of 25]-Minibatch[ 331- 340, 30.98%]: CE.SM = 1.17460327 * 10240; Err = 0.37753906 * 10240; time = 0.0430s; samplesPerSecond = 238383.5
12/20/2016 15:27:51:  Epoch[ 8 of 25]-Minibatch[ 341- 350, 31.89%]: CE.SM = 1.16228943 * 10240; Err = 0.37470703 * 10240; time = 0.0423s; samplesPerSecond = 242321.0
12/20/2016 15:27:51:  Epoch[ 8 of 25]-Minibatch[ 351- 360, 32.80%]: CE.SM = 1.18336487 * 10240; Err = 0.38066406 * 10240; time = 0.0422s; samplesPerSecond = 242441.5
12/20/2016 15:27:51:  Epoch[ 8 of 25]-Minibatch[ 361- 370, 33.71%]: CE.SM = 1.17795105 * 10240; Err = 0.36933594 * 10240; time = 0.0412s; samplesPerSecond = 248561.8
12/20/2016 15:27:51:  Epoch[ 8 of 25]-Minibatch[ 371- 380, 34.62%]: CE.SM = 1.17995300 * 10240; Err = 0.37939453 * 10240; time = 0.0403s; samplesPerSecond = 254119.5
12/20/2016 15:27:51:  Epoch[ 8 of 25]-Minibatch[ 381- 390, 35.54%]: CE.SM = 1.16880188 * 10240; Err = 0.36855469 * 10240; time = 0.0417s; samplesPerSecond = 245416.4
12/20/2016 15:27:51:  Epoch[ 8 of 25]-Minibatch[ 391- 400, 36.45%]: CE.SM = 1.17364197 * 10240; Err = 0.37265625 * 10240; time = 0.0411s; samplesPerSecond = 249336.5
12/20/2016 15:27:51:  Epoch[ 8 of 25]-Minibatch[ 401- 410, 37.36%]: CE.SM = 1.19697876 * 10240; Err = 0.37753906 * 10240; time = 0.0407s; samplesPerSecond = 251498.2
12/20/2016 15:27:51:  Epoch[ 8 of 25]-Minibatch[ 411- 420, 38.27%]: CE.SM = 1.17404175 * 10240; Err = 0.37978516 * 10240; time = 0.0402s; samplesPerSecond = 254872.2
12/20/2016 15:27:51:  Epoch[ 8 of 25]-Minibatch[ 421- 430, 39.18%]: CE.SM = 1.17275696 * 10240; Err = 0.37539062 * 10240; time = 0.0414s; samplesPerSecond = 247540.3
12/20/2016 15:27:51:  Epoch[ 8 of 25]-Minibatch[ 431- 440, 40.09%]: CE.SM = 1.15722351 * 10240; Err = 0.36923828 * 10240; time = 0.0417s; samplesPerSecond = 245822.9
12/20/2016 15:27:51:  Epoch[ 8 of 25]-Minibatch[ 441- 450, 41.00%]: CE.SM = 1.17506714 * 10240; Err = 0.37753906 * 10240; time = 0.0413s; samplesPerSecond = 248206.3
12/20/2016 15:27:51:  Epoch[ 8 of 25]-Minibatch[ 451- 460, 41.91%]: CE.SM = 1.14228516 * 10240; Err = 0.37128906 * 10240; time = 0.0411s; samplesPerSecond = 248851.7
12/20/2016 15:27:51:  Epoch[ 8 of 25]-Minibatch[ 461- 470, 42.82%]: CE.SM = 1.16901855 * 10240; Err = 0.37177734 * 10240; time = 0.0408s; samplesPerSecond = 251140.4
12/20/2016 15:27:51:  Epoch[ 8 of 25]-Minibatch[ 471- 480, 43.74%]: CE.SM = 1.19158325 * 10240; Err = 0.38007812 * 10240; time = 0.0412s; samplesPerSecond = 248809.4
12/20/2016 15:27:51:  Epoch[ 8 of 25]-Minibatch[ 481- 490, 44.65%]: CE.SM = 1.16333008 * 10240; Err = 0.37128906 * 10240; time = 0.0411s; samplesPerSecond = 249233.3
12/20/2016 15:27:51:  Epoch[ 8 of 25]-Minibatch[ 491- 500, 45.56%]: CE.SM = 1.17387695 * 10240; Err = 0.37910156 * 10240; time = 0.0421s; samplesPerSecond = 243068.7
12/20/2016 15:27:51:  Epoch[ 8 of 25]-Minibatch[ 501- 510, 46.47%]: CE.SM = 1.18383179 * 10240; Err = 0.37890625 * 10240; time = 0.0409s; samplesPerSecond = 250207.7
12/20/2016 15:27:51:  Epoch[ 8 of 25]-Minibatch[ 511- 520, 47.38%]: CE.SM = 1.16779785 * 10240; Err = 0.37578125 * 10240; time = 0.0405s; samplesPerSecond = 252671.1
12/20/2016 15:27:51:  Epoch[ 8 of 25]-Minibatch[ 521- 530, 48.29%]: CE.SM = 1.20620728 * 10240; Err = 0.37988281 * 10240; time = 0.0426s; samplesPerSecond = 240398.2
12/20/2016 15:27:52:  Epoch[ 8 of 25]-Minibatch[ 531- 540, 49.20%]: CE.SM = 1.17899170 * 10240; Err = 0.37353516 * 10240; time = 0.0415s; samplesPerSecond = 246979.1
12/20/2016 15:27:52:  Epoch[ 8 of 25]-Minibatch[ 541- 550, 50.11%]: CE.SM = 1.14906006 * 10240; Err = 0.36728516 * 10240; time = 0.0410s; samplesPerSecond = 249707.4
12/20/2016 15:27:52:  Epoch[ 8 of 25]-Minibatch[ 551- 560, 51.03%]: CE.SM = 1.20245361 * 10240; Err = 0.38007812 * 10240; time = 0.0403s; samplesPerSecond = 254050.2
12/20/2016 15:27:52:  Epoch[ 8 of 25]-Minibatch[ 561- 570, 51.94%]: CE.SM = 1.18029785 * 10240; Err = 0.37958984 * 10240; time = 0.0411s; samplesPerSecond = 249324.3
12/20/2016 15:27:52:  Epoch[ 8 of 25]-Minibatch[ 571- 580, 52.85%]: CE.SM = 1.16741943 * 10240; Err = 0.37500000 * 10240; time = 0.0399s; samplesPerSecond = 256513.0
12/20/2016 15:27:52:  Epoch[ 8 of 25]-Minibatch[ 581- 590, 53.76%]: CE.SM = 1.17124634 * 10240; Err = 0.37636719 * 10240; time = 0.0404s; samplesPerSecond = 253333.7
12/20/2016 15:27:52:  Epoch[ 8 of 25]-Minibatch[ 591- 600, 54.67%]: CE.SM = 1.18756104 * 10240; Err = 0.38037109 * 10240; time = 0.0441s; samplesPerSecond = 232341.8
12/20/2016 15:27:52:  Epoch[ 8 of 25]-Minibatch[ 601- 610, 55.58%]: CE.SM = 1.18463135 * 10240; Err = 0.37890625 * 10240; time = 0.0413s; samplesPerSecond = 247714.0
12/20/2016 15:27:52:  Epoch[ 8 of 25]-Minibatch[ 611- 620, 56.49%]: CE.SM = 1.15816650 * 10240; Err = 0.36904297 * 10240; time = 0.0422s; samplesPerSecond = 242378.3
12/20/2016 15:27:52:  Epoch[ 8 of 25]-Minibatch[ 621- 630, 57.40%]: CE.SM = 1.16782837 * 10240; Err = 0.36660156 * 10240; time = 0.0416s; samplesPerSecond = 246331.5
12/20/2016 15:27:52:  Epoch[ 8 of 25]-Minibatch[ 631- 640, 58.31%]: CE.SM = 1.15578003 * 10240; Err = 0.36845703 * 10240; time = 0.0417s; samplesPerSecond = 245675.5
12/20/2016 15:27:52:  Epoch[ 8 of 25]-Minibatch[ 641- 650, 59.23%]: CE.SM = 1.19362183 * 10240; Err = 0.37734375 * 10240; time = 0.0412s; samplesPerSecond = 248374.9
12/20/2016 15:27:52:  Epoch[ 8 of 25]-Minibatch[ 651- 660, 60.14%]: CE.SM = 1.18231201 * 10240; Err = 0.37714844 * 10240; time = 0.0413s; samplesPerSecond = 247953.9
12/20/2016 15:27:52:  Epoch[ 8 of 25]-Minibatch[ 661- 670, 61.05%]: CE.SM = 1.17754517 * 10240; Err = 0.37568359 * 10240; time = 0.0410s; samplesPerSecond = 249573.5
12/20/2016 15:27:52:  Epoch[ 8 of 25]-Minibatch[ 671- 680, 61.96%]: CE.SM = 1.17971802 * 10240; Err = 0.37500000 * 10240; time = 0.0400s; samplesPerSecond = 256314.0
12/20/2016 15:27:52:  Epoch[ 8 of 25]-Minibatch[ 681- 690, 62.87%]: CE.SM = 1.16259766 * 10240; Err = 0.37460938 * 10240; time = 0.0399s; samplesPerSecond = 256378.2
12/20/2016 15:27:52:  Epoch[ 8 of 25]-Minibatch[ 691- 700, 63.78%]: CE.SM = 1.18628540 * 10240; Err = 0.37744141 * 10240; time = 0.0406s; samplesPerSecond = 252018.1
12/20/2016 15:27:52:  Epoch[ 8 of 25]-Minibatch[ 701- 710, 64.69%]: CE.SM = 1.16718140 * 10240; Err = 0.37714844 * 10240; time = 0.0428s; samplesPerSecond = 239174.1
12/20/2016 15:27:52:  Epoch[ 8 of 25]-Minibatch[ 711- 720, 65.60%]: CE.SM = 1.17827759 * 10240; Err = 0.37333984 * 10240; time = 0.0403s; samplesPerSecond = 254144.7
12/20/2016 15:27:52:  Epoch[ 8 of 25]-Minibatch[ 721- 730, 66.51%]: CE.SM = 1.17711182 * 10240; Err = 0.37783203 * 10240; time = 0.0414s; samplesPerSecond = 247624.1
12/20/2016 15:27:52:  Epoch[ 8 of 25]-Minibatch[ 731- 740, 67.43%]: CE.SM = 1.17451172 * 10240; Err = 0.37382813 * 10240; time = 0.0406s; samplesPerSecond = 251980.9
12/20/2016 15:27:52:  Epoch[ 8 of 25]-Minibatch[ 741- 750, 68.34%]: CE.SM = 1.16453857 * 10240; Err = 0.37792969 * 10240; time = 0.0406s; samplesPerSecond = 252111.2
12/20/2016 15:27:52:  Epoch[ 8 of 25]-Minibatch[ 751- 760, 69.25%]: CE.SM = 1.16255493 * 10240; Err = 0.37392578 * 10240; time = 0.0413s; samplesPerSecond = 247953.9
12/20/2016 15:27:52:  Epoch[ 8 of 25]-Minibatch[ 761- 770, 70.16%]: CE.SM = 1.19238281 * 10240; Err = 0.38330078 * 10240; time = 0.0406s; samplesPerSecond = 252502.8
12/20/2016 15:27:53:  Epoch[ 8 of 25]-Minibatch[ 771- 780, 71.07%]: CE.SM = 1.15292358 * 10240; Err = 0.36767578 * 10240; time = 0.0407s; samplesPerSecond = 251455.0
12/20/2016 15:27:53:  Epoch[ 8 of 25]-Minibatch[ 781- 790, 71.98%]: CE.SM = 1.13427124 * 10240; Err = 0.36152344 * 10240; time = 0.0456s; samplesPerSecond = 224699.4
12/20/2016 15:27:53:  Epoch[ 8 of 25]-Minibatch[ 791- 800, 72.89%]: CE.SM = 1.17636108 * 10240; Err = 0.37304688 * 10240; time = 0.0453s; samplesPerSecond = 226223.4
12/20/2016 15:27:53:  Epoch[ 8 of 25]-Minibatch[ 801- 810, 73.80%]: CE.SM = 1.17728882 * 10240; Err = 0.37431641 * 10240; time = 0.0414s; samplesPerSecond = 247181.8
12/20/2016 15:27:53:  Epoch[ 8 of 25]-Minibatch[ 811- 820, 74.72%]: CE.SM = 1.19040527 * 10240; Err = 0.38007812 * 10240; time = 0.0402s; samplesPerSecond = 254669.4
12/20/2016 15:27:53:  Epoch[ 8 of 25]-Minibatch[ 821- 830, 75.63%]: CE.SM = 1.14154053 * 10240; Err = 0.36806641 * 10240; time = 0.0408s; samplesPerSecond = 250777.6
12/20/2016 15:27:53:  Epoch[ 8 of 25]-Minibatch[ 831- 840, 76.54%]: CE.SM = 1.18237305 * 10240; Err = 0.37958984 * 10240; time = 0.0416s; samplesPerSecond = 245882.0
12/20/2016 15:27:53:  Epoch[ 8 of 25]-Minibatch[ 841- 850, 77.45%]: CE.SM = 1.16619263 * 10240; Err = 0.37304688 * 10240; time = 0.0400s; samplesPerSecond = 255974.4
12/20/2016 15:27:53:  Epoch[ 8 of 25]-Minibatch[ 851- 860, 78.36%]: CE.SM = 1.15889893 * 10240; Err = 0.36484375 * 10240; time = 0.0408s; samplesPerSecond = 250949.6
12/20/2016 15:27:53:  Epoch[ 8 of 25]-Minibatch[ 861- 870, 79.27%]: CE.SM = 1.16030273 * 10240; Err = 0.37714844 * 10240; time = 0.0414s; samplesPerSecond = 247307.2
12/20/2016 15:27:53:  Epoch[ 8 of 25]-Minibatch[ 871- 880, 80.18%]: CE.SM = 1.16898804 * 10240; Err = 0.36884766 * 10240; time = 0.0407s; samplesPerSecond = 251757.9
12/20/2016 15:27:53:  Epoch[ 8 of 25]-Minibatch[ 881- 890, 81.09%]: CE.SM = 1.15755615 * 10240; Err = 0.36982422 * 10240; time = 0.0409s; samplesPerSecond = 250513.7
12/20/2016 15:27:53:  Epoch[ 8 of 25]-Minibatch[ 891- 900, 82.00%]: CE.SM = 1.16673584 * 10240; Err = 0.37773438 * 10240; time = 0.0415s; samplesPerSecond = 246949.3
12/20/2016 15:27:53:  Epoch[ 8 of 25]-Minibatch[ 901- 910, 82.92%]: CE.SM = 1.18425293 * 10240; Err = 0.37666016 * 10240; time = 0.0416s; samplesPerSecond = 246153.8
12/20/2016 15:27:53:  Epoch[ 8 of 25]-Minibatch[ 911- 920, 83.83%]: CE.SM = 1.16461182 * 10240; Err = 0.37226562 * 10240; time = 0.0415s; samplesPerSecond = 246985.0
12/20/2016 15:27:53:  Epoch[ 8 of 25]-Minibatch[ 921- 930, 84.74%]: CE.SM = 1.16403809 * 10240; Err = 0.36875000 * 10240; time = 0.0419s; samplesPerSecond = 244549.0
12/20/2016 15:27:53:  Epoch[ 8 of 25]-Minibatch[ 931- 940, 85.65%]: CE.SM = 1.16501465 * 10240; Err = 0.37109375 * 10240; time = 0.0412s; samplesPerSecond = 248664.4
12/20/2016 15:27:53:  Epoch[ 8 of 25]-Minibatch[ 941- 950, 86.56%]: CE.SM = 1.18261719 * 10240; Err = 0.37705078 * 10240; time = 0.0406s; samplesPerSecond = 252223.0
12/20/2016 15:27:53:  Epoch[ 8 of 25]-Minibatch[ 951- 960, 87.47%]: CE.SM = 1.13518066 * 10240; Err = 0.36171875 * 10240; time = 0.0417s; samplesPerSecond = 245316.5
12/20/2016 15:27:53:  Epoch[ 8 of 25]-Minibatch[ 961- 970, 88.38%]: CE.SM = 1.15676270 * 10240; Err = 0.36943359 * 10240; time = 0.0411s; samplesPerSecond = 248936.4
12/20/2016 15:27:53:  Epoch[ 8 of 25]-Minibatch[ 971- 980, 89.29%]: CE.SM = 1.19014893 * 10240; Err = 0.37519531 * 10240; time = 0.0409s; samplesPerSecond = 250116.0
12/20/2016 15:27:53:  Epoch[ 8 of 25]-Minibatch[ 981- 990, 90.21%]: CE.SM = 1.17581787 * 10240; Err = 0.37333984 * 10240; time = 0.0404s; samplesPerSecond = 253471.6
12/20/2016 15:27:53:  Epoch[ 8 of 25]-Minibatch[ 991-1000, 91.12%]: CE.SM = 1.15682373 * 10240; Err = 0.36591797 * 10240; time = 0.0400s; samplesPerSecond = 256108.8
12/20/2016 15:27:53:  Epoch[ 8 of 25]-Minibatch[1001-1010, 92.03%]: CE.SM = 1.17191162 * 10240; Err = 0.37441406 * 10240; time = 0.0408s; samplesPerSecond = 251005.0
12/20/2016 15:27:54:  Epoch[ 8 of 25]-Minibatch[1011-1020, 92.94%]: CE.SM = 1.15561523 * 10240; Err = 0.36718750 * 10240; time = 0.0419s; samplesPerSecond = 244432.2
12/20/2016 15:27:54:  Epoch[ 8 of 25]-Minibatch[1021-1030, 93.85%]: CE.SM = 1.16855469 * 10240; Err = 0.37861328 * 10240; time = 0.0402s; samplesPerSecond = 254897.6
12/20/2016 15:27:54:  Epoch[ 8 of 25]-Minibatch[1031-1040, 94.76%]: CE.SM = 1.15213623 * 10240; Err = 0.36767578 * 10240; time = 0.0407s; samplesPerSecond = 251529.1
12/20/2016 15:27:54:  Epoch[ 8 of 25]-Minibatch[1041-1050, 95.67%]: CE.SM = 1.14560547 * 10240; Err = 0.37177734 * 10240; time = 0.0414s; samplesPerSecond = 247211.6
12/20/2016 15:27:54:  Epoch[ 8 of 25]-Minibatch[1051-1060, 96.58%]: CE.SM = 1.16243896 * 10240; Err = 0.37216797 * 10240; time = 0.0418s; samplesPerSecond = 244689.2
12/20/2016 15:27:54:  Epoch[ 8 of 25]-Minibatch[1061-1070, 97.49%]: CE.SM = 1.17281494 * 10240; Err = 0.37851563 * 10240; time = 0.0415s; samplesPerSecond = 247026.8
12/20/2016 15:27:54:  Epoch[ 8 of 25]-Minibatch[1071-1080, 98.41%]: CE.SM = 1.16181641 * 10240; Err = 0.37460938 * 10240; time = 0.0403s; samplesPerSecond = 254334.1
12/20/2016 15:27:54:  Epoch[ 8 of 25]-Minibatch[1081-1090, 99.32%]: CE.SM = 1.16320801 * 10240; Err = 0.36835937 * 10240; time = 0.0405s; samplesPerSecond = 253127.0
12/20/2016 15:27:54: Finished Epoch[ 8 of 25]: [Training] CE.SM = 1.17047127 * 1124823; Err = 0.37393617 * 1124823; totalSamplesSeen = 8998584; learningRatePerSample = 0.003125; epochTime=4.77717s
12/20/2016 15:27:54: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn.8'

12/20/2016 15:27:54: Starting Epoch 9: learning rate per sample = 0.003125  effective momentum = 0.900000  momentum as time constant = 9719.0 samples

12/20/2016 15:27:54: Starting minibatch loop.
12/20/2016 15:27:54:  Epoch[ 9 of 25]-Minibatch[   1-  10, 0.91%]: CE.SM = 1.11439342 * 10240; Err = 0.35898438 * 10240; time = 0.0453s; samplesPerSecond = 225894.0
12/20/2016 15:27:54:  Epoch[ 9 of 25]-Minibatch[  11-  20, 1.82%]: CE.SM = 1.13200550 * 10240; Err = 0.35947266 * 10240; time = 0.0440s; samplesPerSecond = 232663.8
12/20/2016 15:27:54:  Epoch[ 9 of 25]-Minibatch[  21-  30, 2.73%]: CE.SM = 1.11156120 * 10240; Err = 0.35996094 * 10240; time = 0.0408s; samplesPerSecond = 251128.1
12/20/2016 15:27:54:  Epoch[ 9 of 25]-Minibatch[  31-  40, 3.64%]: CE.SM = 1.13190804 * 10240; Err = 0.36328125 * 10240; time = 0.0397s; samplesPerSecond = 258038.5
12/20/2016 15:27:54:  Epoch[ 9 of 25]-Minibatch[  41-  50, 4.56%]: CE.SM = 1.11050797 * 10240; Err = 0.35566406 * 10240; time = 0.0403s; samplesPerSecond = 254043.9
12/20/2016 15:27:54:  Epoch[ 9 of 25]-Minibatch[  51-  60, 5.47%]: CE.SM = 1.14945068 * 10240; Err = 0.36748047 * 10240; time = 0.0399s; samplesPerSecond = 256346.1
12/20/2016 15:27:54:  Epoch[ 9 of 25]-Minibatch[  61-  70, 6.38%]: CE.SM = 1.15015945 * 10240; Err = 0.36718750 * 10240; time = 0.0407s; samplesPerSecond = 251801.2
12/20/2016 15:27:54:  Epoch[ 9 of 25]-Minibatch[  71-  80, 7.29%]: CE.SM = 1.16514130 * 10240; Err = 0.37167969 * 10240; time = 0.0441s; samplesPerSecond = 232257.5
12/20/2016 15:27:54:  Epoch[ 9 of 25]-Minibatch[  81-  90, 8.20%]: CE.SM = 1.13193817 * 10240; Err = 0.36621094 * 10240; time = 0.0446s; samplesPerSecond = 229766.4
12/20/2016 15:27:55:  Epoch[ 9 of 25]-Minibatch[  91- 100, 9.11%]: CE.SM = 1.15028305 * 10240; Err = 0.37255859 * 10240; time = 0.0429s; samplesPerSecond = 238494.5
12/20/2016 15:27:55:  Epoch[ 9 of 25]-Minibatch[ 101- 110, 10.02%]: CE.SM = 1.13765869 * 10240; Err = 0.36298828 * 10240; time = 0.0424s; samplesPerSecond = 241737.5
12/20/2016 15:27:55:  Epoch[ 9 of 25]-Minibatch[ 111- 120, 10.93%]: CE.SM = 1.14276657 * 10240; Err = 0.36640625 * 10240; time = 0.0414s; samplesPerSecond = 247265.4
12/20/2016 15:27:55:  Epoch[ 9 of 25]-Minibatch[ 121- 130, 11.85%]: CE.SM = 1.12703705 * 10240; Err = 0.36425781 * 10240; time = 0.0416s; samplesPerSecond = 246367.0
12/20/2016 15:27:55:  Epoch[ 9 of 25]-Minibatch[ 131- 140, 12.76%]: CE.SM = 1.13758240 * 10240; Err = 0.36923828 * 10240; time = 0.0415s; samplesPerSecond = 247002.9
12/20/2016 15:27:55:  Epoch[ 9 of 25]-Minibatch[ 141- 150, 13.67%]: CE.SM = 1.15263519 * 10240; Err = 0.37060547 * 10240; time = 0.0419s; samplesPerSecond = 244514.0
12/20/2016 15:27:55:  Epoch[ 9 of 25]-Minibatch[ 151- 160, 14.58%]: CE.SM = 1.14361877 * 10240; Err = 0.36738281 * 10240; time = 0.0418s; samplesPerSecond = 244788.7
12/20/2016 15:27:55:  Epoch[ 9 of 25]-Minibatch[ 161- 170, 15.49%]: CE.SM = 1.15400543 * 10240; Err = 0.37021484 * 10240; time = 0.0410s; samplesPerSecond = 249762.2
12/20/2016 15:27:55:  Epoch[ 9 of 25]-Minibatch[ 171- 180, 16.40%]: CE.SM = 1.17128448 * 10240; Err = 0.37353516 * 10240; time = 0.0403s; samplesPerSecond = 254220.5
12/20/2016 15:27:55:  Epoch[ 9 of 25]-Minibatch[ 181- 190, 17.31%]: CE.SM = 1.13865051 * 10240; Err = 0.36806641 * 10240; time = 0.0416s; samplesPerSecond = 246082.9
12/20/2016 15:27:55:  Epoch[ 9 of 25]-Minibatch[ 191- 200, 18.22%]: CE.SM = 1.11599426 * 10240; Err = 0.36308594 * 10240; time = 0.0413s; samplesPerSecond = 247773.9
12/20/2016 15:27:55:  Epoch[ 9 of 25]-Minibatch[ 201- 210, 19.13%]: CE.SM = 1.12889252 * 10240; Err = 0.36142578 * 10240; time = 0.0410s; samplesPerSecond = 249664.8
12/20/2016 15:27:55:  Epoch[ 9 of 25]-Minibatch[ 211- 220, 20.05%]: CE.SM = 1.15295715 * 10240; Err = 0.37187500 * 10240; time = 0.0408s; samplesPerSecond = 250968.1
12/20/2016 15:27:55:  Epoch[ 9 of 25]-Minibatch[ 221- 230, 20.96%]: CE.SM = 1.12399750 * 10240; Err = 0.35312500 * 10240; time = 0.0413s; samplesPerSecond = 247791.9
12/20/2016 15:27:55:  Epoch[ 9 of 25]-Minibatch[ 231- 240, 21.87%]: CE.SM = 1.12522888 * 10240; Err = 0.36093750 * 10240; time = 0.0412s; samplesPerSecond = 248640.2
12/20/2016 15:27:55:  Epoch[ 9 of 25]-Minibatch[ 241- 250, 22.78%]: CE.SM = 1.13381958 * 10240; Err = 0.36884766 * 10240; time = 0.0407s; samplesPerSecond = 251677.4
12/20/2016 15:27:55:  Epoch[ 9 of 25]-Minibatch[ 251- 260, 23.69%]: CE.SM = 1.14318848 * 10240; Err = 0.36250000 * 10240; time = 0.0412s; samplesPerSecond = 248459.3
12/20/2016 15:27:55:  Epoch[ 9 of 25]-Minibatch[ 261- 270, 24.60%]: CE.SM = 1.11600952 * 10240; Err = 0.36171875 * 10240; time = 0.0415s; samplesPerSecond = 246735.1
12/20/2016 15:27:55:  Epoch[ 9 of 25]-Minibatch[ 271- 280, 25.51%]: CE.SM = 1.14188538 * 10240; Err = 0.36386719 * 10240; time = 0.0407s; samplesPerSecond = 251566.1
12/20/2016 15:27:55:  Epoch[ 9 of 25]-Minibatch[ 281- 290, 26.42%]: CE.SM = 1.16066895 * 10240; Err = 0.37041016 * 10240; time = 0.0409s; samplesPerSecond = 250079.4
12/20/2016 15:27:55:  Epoch[ 9 of 25]-Minibatch[ 291- 300, 27.33%]: CE.SM = 1.15304260 * 10240; Err = 0.36806641 * 10240; time = 0.0407s; samplesPerSecond = 251702.2
12/20/2016 15:27:55:  Epoch[ 9 of 25]-Minibatch[ 301- 310, 28.25%]: CE.SM = 1.14992065 * 10240; Err = 0.37246094 * 10240; time = 0.0406s; samplesPerSecond = 252409.5
12/20/2016 15:27:55:  Epoch[ 9 of 25]-Minibatch[ 311- 320, 29.16%]: CE.SM = 1.18078613 * 10240; Err = 0.37734375 * 10240; time = 0.0406s; samplesPerSecond = 252378.4
12/20/2016 15:27:55:  Epoch[ 9 of 25]-Minibatch[ 321- 330, 30.07%]: CE.SM = 1.12718201 * 10240; Err = 0.36201172 * 10240; time = 0.0410s; samplesPerSecond = 249756.1
12/20/2016 15:27:56:  Epoch[ 9 of 25]-Minibatch[ 331- 340, 30.98%]: CE.SM = 1.13842773 * 10240; Err = 0.36806641 * 10240; time = 0.0398s; samplesPerSecond = 257144.3
12/20/2016 15:27:56:  Epoch[ 9 of 25]-Minibatch[ 341- 350, 31.89%]: CE.SM = 1.13220215 * 10240; Err = 0.36591797 * 10240; time = 0.0400s; samplesPerSecond = 255757.0
12/20/2016 15:27:56:  Epoch[ 9 of 25]-Minibatch[ 351- 360, 32.80%]: CE.SM = 1.14487000 * 10240; Err = 0.36650391 * 10240; time = 0.0480s; samplesPerSecond = 213484.6
12/20/2016 15:27:56:  Epoch[ 9 of 25]-Minibatch[ 361- 370, 33.71%]: CE.SM = 1.14629517 * 10240; Err = 0.36728516 * 10240; time = 0.0422s; samplesPerSecond = 242659.8
12/20/2016 15:27:56:  Epoch[ 9 of 25]-Minibatch[ 371- 380, 34.62%]: CE.SM = 1.12862854 * 10240; Err = 0.36816406 * 10240; time = 0.0406s; samplesPerSecond = 252353.5
12/20/2016 15:27:56:  Epoch[ 9 of 25]-Minibatch[ 381- 390, 35.54%]: CE.SM = 1.14126282 * 10240; Err = 0.36171875 * 10240; time = 0.0404s; samplesPerSecond = 253371.3
12/20/2016 15:27:56:  Epoch[ 9 of 25]-Minibatch[ 391- 400, 36.45%]: CE.SM = 1.13685303 * 10240; Err = 0.36816406 * 10240; time = 0.0413s; samplesPerSecond = 248200.3
12/20/2016 15:27:56:  Epoch[ 9 of 25]-Minibatch[ 401- 410, 37.36%]: CE.SM = 1.14827271 * 10240; Err = 0.36962891 * 10240; time = 0.0409s; samplesPerSecond = 250618.0
12/20/2016 15:27:56:  Epoch[ 9 of 25]-Minibatch[ 411- 420, 38.27%]: CE.SM = 1.16770325 * 10240; Err = 0.37578125 * 10240; time = 0.0402s; samplesPerSecond = 254726.4
12/20/2016 15:27:56:  Epoch[ 9 of 25]-Minibatch[ 421- 430, 39.18%]: CE.SM = 1.12505798 * 10240; Err = 0.35781250 * 10240; time = 0.0408s; samplesPerSecond = 250845.1
12/20/2016 15:27:56:  Epoch[ 9 of 25]-Minibatch[ 431- 440, 40.09%]: CE.SM = 1.15171204 * 10240; Err = 0.36884766 * 10240; time = 0.0412s; samplesPerSecond = 248604.0
12/20/2016 15:27:56:  Epoch[ 9 of 25]-Minibatch[ 441- 450, 41.00%]: CE.SM = 1.13880920 * 10240; Err = 0.36259766 * 10240; time = 0.0414s; samplesPerSecond = 247408.7
12/20/2016 15:27:56:  Epoch[ 9 of 25]-Minibatch[ 451- 460, 41.91%]: CE.SM = 1.12868042 * 10240; Err = 0.36035156 * 10240; time = 0.0413s; samplesPerSecond = 247785.9
12/20/2016 15:27:56:  Epoch[ 9 of 25]-Minibatch[ 461- 470, 42.82%]: CE.SM = 1.14776611 * 10240; Err = 0.36748047 * 10240; time = 0.0407s; samplesPerSecond = 251881.7
12/20/2016 15:27:56:  Epoch[ 9 of 25]-Minibatch[ 471- 480, 43.74%]: CE.SM = 1.15319824 * 10240; Err = 0.36757812 * 10240; time = 0.0404s; samplesPerSecond = 253189.6
12/20/2016 15:27:56:  Epoch[ 9 of 25]-Minibatch[ 481- 490, 44.65%]: CE.SM = 1.16716309 * 10240; Err = 0.37871094 * 10240; time = 0.0406s; samplesPerSecond = 251974.7
12/20/2016 15:27:56:  Epoch[ 9 of 25]-Minibatch[ 491- 500, 45.56%]: CE.SM = 1.14012451 * 10240; Err = 0.37148437 * 10240; time = 0.0419s; samplesPerSecond = 244671.7
12/20/2016 15:27:56:  Epoch[ 9 of 25]-Minibatch[ 501- 510, 46.47%]: CE.SM = 1.13063354 * 10240; Err = 0.36337891 * 10240; time = 0.0401s; samplesPerSecond = 255654.9
12/20/2016 15:27:56:  Epoch[ 9 of 25]-Minibatch[ 511- 520, 47.38%]: CE.SM = 1.16026611 * 10240; Err = 0.37031250 * 10240; time = 0.0409s; samplesPerSecond = 250642.5
12/20/2016 15:27:56:  Epoch[ 9 of 25]-Minibatch[ 521- 530, 48.29%]: CE.SM = 1.14611816 * 10240; Err = 0.36181641 * 10240; time = 0.0407s; samplesPerSecond = 251640.3
12/20/2016 15:27:56:  Epoch[ 9 of 25]-Minibatch[ 531- 540, 49.20%]: CE.SM = 1.13996582 * 10240; Err = 0.36787109 * 10240; time = 0.0401s; samplesPerSecond = 255113.5
12/20/2016 15:27:56:  Epoch[ 9 of 25]-Minibatch[ 541- 550, 50.11%]: CE.SM = 1.15352173 * 10240; Err = 0.36894531 * 10240; time = 0.0400s; samplesPerSecond = 255891.2
12/20/2016 15:27:56:  Epoch[ 9 of 25]-Minibatch[ 551- 560, 51.03%]: CE.SM = 1.14992676 * 10240; Err = 0.36992188 * 10240; time = 0.0400s; samplesPerSecond = 256314.0
12/20/2016 15:27:56:  Epoch[ 9 of 25]-Minibatch[ 561- 570, 51.94%]: CE.SM = 1.14282837 * 10240; Err = 0.36953125 * 10240; time = 0.0399s; samplesPerSecond = 256570.9
12/20/2016 15:27:57:  Epoch[ 9 of 25]-Minibatch[ 571- 580, 52.85%]: CE.SM = 1.14742432 * 10240; Err = 0.36972656 * 10240; time = 0.0399s; samplesPerSecond = 256712.4
12/20/2016 15:27:57:  Epoch[ 9 of 25]-Minibatch[ 581- 590, 53.76%]: CE.SM = 1.12858276 * 10240; Err = 0.36328125 * 10240; time = 0.0401s; samplesPerSecond = 255674.0
12/20/2016 15:27:57:  Epoch[ 9 of 25]-Minibatch[ 591- 600, 54.67%]: CE.SM = 1.16118164 * 10240; Err = 0.37382813 * 10240; time = 0.0399s; samplesPerSecond = 256551.6
12/20/2016 15:27:57:  Epoch[ 9 of 25]-Minibatch[ 601- 610, 55.58%]: CE.SM = 1.13369751 * 10240; Err = 0.36884766 * 10240; time = 0.0400s; samplesPerSecond = 255955.2
12/20/2016 15:27:57:  Epoch[ 9 of 25]-Minibatch[ 611- 620, 56.49%]: CE.SM = 1.18173218 * 10240; Err = 0.37236328 * 10240; time = 0.0418s; samplesPerSecond = 245064.0
12/20/2016 15:27:57:  Epoch[ 9 of 25]-Minibatch[ 621- 630, 57.40%]: CE.SM = 1.14647217 * 10240; Err = 0.37050781 * 10240; time = 0.0413s; samplesPerSecond = 248170.2
12/20/2016 15:27:57:  Epoch[ 9 of 25]-Minibatch[ 631- 640, 58.31%]: CE.SM = 1.16033936 * 10240; Err = 0.36660156 * 10240; time = 0.0408s; samplesPerSecond = 250912.7
12/20/2016 15:27:57:  Epoch[ 9 of 25]-Minibatch[ 641- 650, 59.23%]: CE.SM = 1.14703369 * 10240; Err = 0.36943359 * 10240; time = 0.0409s; samplesPerSecond = 250452.5
12/20/2016 15:27:57:  Epoch[ 9 of 25]-Minibatch[ 651- 660, 60.14%]: CE.SM = 1.15239258 * 10240; Err = 0.37294922 * 10240; time = 0.0403s; samplesPerSecond = 254308.8
12/20/2016 15:27:57:  Epoch[ 9 of 25]-Minibatch[ 661- 670, 61.05%]: CE.SM = 1.14221802 * 10240; Err = 0.36337891 * 10240; time = 0.0406s; samplesPerSecond = 252434.4
12/20/2016 15:27:57:  Epoch[ 9 of 25]-Minibatch[ 671- 680, 61.96%]: CE.SM = 1.13240967 * 10240; Err = 0.36660156 * 10240; time = 0.0408s; samplesPerSecond = 251220.5
12/20/2016 15:27:57:  Epoch[ 9 of 25]-Minibatch[ 681- 690, 62.87%]: CE.SM = 1.16109619 * 10240; Err = 0.36894531 * 10240; time = 0.0402s; samplesPerSecond = 254789.7
12/20/2016 15:27:57:  Epoch[ 9 of 25]-Minibatch[ 691- 700, 63.78%]: CE.SM = 1.13227539 * 10240; Err = 0.36533203 * 10240; time = 0.0400s; samplesPerSecond = 256006.4
12/20/2016 15:27:57:  Epoch[ 9 of 25]-Minibatch[ 701- 710, 64.69%]: CE.SM = 1.12026978 * 10240; Err = 0.36132812 * 10240; time = 0.0409s; samplesPerSecond = 250183.2
12/20/2016 15:27:57:  Epoch[ 9 of 25]-Minibatch[ 711- 720, 65.60%]: CE.SM = 1.11231689 * 10240; Err = 0.35888672 * 10240; time = 0.0409s; samplesPerSecond = 250201.6
12/20/2016 15:27:57:  Epoch[ 9 of 25]-Minibatch[ 721- 730, 66.51%]: CE.SM = 1.11784058 * 10240; Err = 0.35986328 * 10240; time = 0.0408s; samplesPerSecond = 251128.1
12/20/2016 15:27:57:  Epoch[ 9 of 25]-Minibatch[ 731- 740, 67.43%]: CE.SM = 1.15226440 * 10240; Err = 0.36679688 * 10240; time = 0.0412s; samplesPerSecond = 248362.8
12/20/2016 15:27:57:  Epoch[ 9 of 25]-Minibatch[ 741- 750, 68.34%]: CE.SM = 1.13311157 * 10240; Err = 0.35800781 * 10240; time = 0.0411s; samplesPerSecond = 249251.5
12/20/2016 15:27:57:  Epoch[ 9 of 25]-Minibatch[ 751- 760, 69.25%]: CE.SM = 1.16543579 * 10240; Err = 0.37197266 * 10240; time = 0.0406s; samplesPerSecond = 251987.1
12/20/2016 15:27:57:  Epoch[ 9 of 25]-Minibatch[ 761- 770, 70.16%]: CE.SM = 1.14399414 * 10240; Err = 0.36689453 * 10240; time = 0.0409s; samplesPerSecond = 250477.0
12/20/2016 15:27:57:  Epoch[ 9 of 25]-Minibatch[ 771- 780, 71.07%]: CE.SM = 1.15466309 * 10240; Err = 0.36943359 * 10240; time = 0.0404s; samplesPerSecond = 253691.4
12/20/2016 15:27:57:  Epoch[ 9 of 25]-Minibatch[ 781- 790, 71.98%]: CE.SM = 1.13218994 * 10240; Err = 0.36162109 * 10240; time = 0.0408s; samplesPerSecond = 250980.4
12/20/2016 15:27:57:  Epoch[ 9 of 25]-Minibatch[ 791- 800, 72.89%]: CE.SM = 1.13908081 * 10240; Err = 0.36552734 * 10240; time = 0.0407s; samplesPerSecond = 251337.7
12/20/2016 15:27:57:  Epoch[ 9 of 25]-Minibatch[ 801- 810, 73.80%]: CE.SM = 1.16975708 * 10240; Err = 0.37656250 * 10240; time = 0.0412s; samplesPerSecond = 248640.2
12/20/2016 15:27:57:  Epoch[ 9 of 25]-Minibatch[ 811- 820, 74.72%]: CE.SM = 1.12988281 * 10240; Err = 0.36777344 * 10240; time = 0.0409s; samplesPerSecond = 250091.6
12/20/2016 15:27:58:  Epoch[ 9 of 25]-Minibatch[ 821- 830, 75.63%]: CE.SM = 1.12719727 * 10240; Err = 0.36406250 * 10240; time = 0.0396s; samplesPerSecond = 258598.9
12/20/2016 15:27:58:  Epoch[ 9 of 25]-Minibatch[ 831- 840, 76.54%]: CE.SM = 1.16270142 * 10240; Err = 0.37148437 * 10240; time = 0.0421s; samplesPerSecond = 243432.8
12/20/2016 15:27:58:  Epoch[ 9 of 25]-Minibatch[ 841- 850, 77.45%]: CE.SM = 1.14605103 * 10240; Err = 0.36904297 * 10240; time = 0.0404s; samplesPerSecond = 253452.8
12/20/2016 15:27:58:  Epoch[ 9 of 25]-Minibatch[ 851- 860, 78.36%]: CE.SM = 1.14787598 * 10240; Err = 0.36884766 * 10240; time = 0.0408s; samplesPerSecond = 250753.0
12/20/2016 15:27:58:  Epoch[ 9 of 25]-Minibatch[ 861- 870, 79.27%]: CE.SM = 1.14149170 * 10240; Err = 0.36718750 * 10240; time = 0.0413s; samplesPerSecond = 247935.9
12/20/2016 15:27:58:  Epoch[ 9 of 25]-Minibatch[ 871- 880, 80.18%]: CE.SM = 1.16851807 * 10240; Err = 0.37373047 * 10240; time = 0.0406s; samplesPerSecond = 251925.1
12/20/2016 15:27:58:  Epoch[ 9 of 25]-Minibatch[ 881- 890, 81.09%]: CE.SM = 1.14067383 * 10240; Err = 0.36904297 * 10240; time = 0.0402s; samplesPerSecond = 254447.9
12/20/2016 15:27:58:  Epoch[ 9 of 25]-Minibatch[ 891- 900, 82.00%]: CE.SM = 1.12813721 * 10240; Err = 0.36679688 * 10240; time = 0.0406s; samplesPerSecond = 251993.3
12/20/2016 15:27:58:  Epoch[ 9 of 25]-Minibatch[ 901- 910, 82.92%]: CE.SM = 1.14923096 * 10240; Err = 0.36708984 * 10240; time = 0.0405s; samplesPerSecond = 252914.4
12/20/2016 15:27:58:  Epoch[ 9 of 25]-Minibatch[ 911- 920, 83.83%]: CE.SM = 1.13627930 * 10240; Err = 0.36406250 * 10240; time = 0.0398s; samplesPerSecond = 257086.2
12/20/2016 15:27:58:  Epoch[ 9 of 25]-Minibatch[ 921- 930, 84.74%]: CE.SM = 1.11568604 * 10240; Err = 0.35498047 * 10240; time = 0.0406s; samplesPerSecond = 251993.3
12/20/2016 15:27:58:  Epoch[ 9 of 25]-Minibatch[ 931- 940, 85.65%]: CE.SM = 1.13935547 * 10240; Err = 0.35966797 * 10240; time = 0.0403s; samplesPerSecond = 254195.2
12/20/2016 15:27:58:  Epoch[ 9 of 25]-Minibatch[ 941- 950, 86.56%]: CE.SM = 1.14739990 * 10240; Err = 0.37050781 * 10240; time = 0.0405s; samplesPerSecond = 253033.2
12/20/2016 15:27:58:  Epoch[ 9 of 25]-Minibatch[ 951- 960, 87.47%]: CE.SM = 1.16165771 * 10240; Err = 0.37109375 * 10240; time = 0.0403s; samplesPerSecond = 253911.6
12/20/2016 15:27:58:  Epoch[ 9 of 25]-Minibatch[ 961- 970, 88.38%]: CE.SM = 1.14221191 * 10240; Err = 0.36865234 * 10240; time = 0.0395s; samplesPerSecond = 259024.1
12/20/2016 15:27:58:  Epoch[ 9 of 25]-Minibatch[ 971- 980, 89.29%]: CE.SM = 1.11970215 * 10240; Err = 0.36494141 * 10240; time = 0.0392s; samplesPerSecond = 261284.5
12/20/2016 15:27:58:  Epoch[ 9 of 25]-Minibatch[ 981- 990, 90.21%]: CE.SM = 1.12823486 * 10240; Err = 0.36582031 * 10240; time = 0.0397s; samplesPerSecond = 257765.7
12/20/2016 15:27:58:  Epoch[ 9 of 25]-Minibatch[ 991-1000, 91.12%]: CE.SM = 1.14385986 * 10240; Err = 0.36474609 * 10240; time = 0.0397s; samplesPerSecond = 258129.6
12/20/2016 15:27:58:  Epoch[ 9 of 25]-Minibatch[1001-1010, 92.03%]: CE.SM = 1.13854980 * 10240; Err = 0.36416016 * 10240; time = 0.0398s; samplesPerSecond = 256989.4
12/20/2016 15:27:58:  Epoch[ 9 of 25]-Minibatch[1011-1020, 92.94%]: CE.SM = 1.11375732 * 10240; Err = 0.35683594 * 10240; time = 0.0396s; samplesPerSecond = 258403.1
12/20/2016 15:27:58:  Epoch[ 9 of 25]-Minibatch[1021-1030, 93.85%]: CE.SM = 1.12630615 * 10240; Err = 0.36054687 * 10240; time = 0.0410s; samplesPerSecond = 249810.9
12/20/2016 15:27:58:  Epoch[ 9 of 25]-Minibatch[1031-1040, 94.76%]: CE.SM = 1.14398193 * 10240; Err = 0.36552734 * 10240; time = 0.0401s; samplesPerSecond = 255138.9
12/20/2016 15:27:58:  Epoch[ 9 of 25]-Minibatch[1041-1050, 95.67%]: CE.SM = 1.13605957 * 10240; Err = 0.36347656 * 10240; time = 0.0404s; samplesPerSecond = 253660.0
12/20/2016 15:27:58:  Epoch[ 9 of 25]-Minibatch[1051-1060, 96.58%]: CE.SM = 1.14685059 * 10240; Err = 0.36562500 * 10240; time = 0.0433s; samplesPerSecond = 236298.6
12/20/2016 15:27:58:  Epoch[ 9 of 25]-Minibatch[1061-1070, 97.49%]: CE.SM = 1.12934570 * 10240; Err = 0.36552734 * 10240; time = 0.0431s; samplesPerSecond = 237758.0
12/20/2016 15:27:59:  Epoch[ 9 of 25]-Minibatch[1071-1080, 98.41%]: CE.SM = 1.13002930 * 10240; Err = 0.36464844 * 10240; time = 0.0436s; samplesPerSecond = 235002.5
12/20/2016 15:27:59:  Epoch[ 9 of 25]-Minibatch[1081-1090, 99.32%]: CE.SM = 1.12952881 * 10240; Err = 0.36035156 * 10240; time = 0.0471s; samplesPerSecond = 217640.8
12/20/2016 15:27:59: Finished Epoch[ 9 of 25]: [Training] CE.SM = 1.14140236 * 1124823; Err = 0.36639987 * 1124823; totalSamplesSeen = 10123407; learningRatePerSample = 0.003125; epochTime=4.70282s
12/20/2016 15:27:59: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn.9'

12/20/2016 15:27:59: Starting Epoch 10: learning rate per sample = 0.003125  effective momentum = 0.900000  momentum as time constant = 9719.0 samples

12/20/2016 15:27:59: Starting minibatch loop.
12/20/2016 15:27:59:  Epoch[10 of 25]-Minibatch[   1-  10, 0.91%]: CE.SM = 1.09907522 * 10240; Err = 0.35800781 * 10240; time = 0.0800s; samplesPerSecond = 127974.4
12/20/2016 15:27:59:  Epoch[10 of 25]-Minibatch[  11-  20, 1.82%]: CE.SM = 1.10512896 * 10240; Err = 0.36064453 * 10240; time = 0.0486s; samplesPerSecond = 210795.0
12/20/2016 15:27:59:  Epoch[10 of 25]-Minibatch[  21-  30, 2.73%]: CE.SM = 1.10691795 * 10240; Err = 0.35507813 * 10240; time = 0.0431s; samplesPerSecond = 237691.8
12/20/2016 15:27:59:  Epoch[10 of 25]-Minibatch[  31-  40, 3.64%]: CE.SM = 1.09731865 * 10240; Err = 0.35673828 * 10240; time = 0.0426s; samplesPerSecond = 240477.2
12/20/2016 15:27:59:  Epoch[10 of 25]-Minibatch[  41-  50, 4.56%]: CE.SM = 1.12551575 * 10240; Err = 0.36074219 * 10240; time = 0.0422s; samplesPerSecond = 242447.2
12/20/2016 15:27:59:  Epoch[10 of 25]-Minibatch[  51-  60, 5.47%]: CE.SM = 1.08489799 * 10240; Err = 0.35039063 * 10240; time = 0.0414s; samplesPerSecond = 247062.5
12/20/2016 15:27:59:  Epoch[10 of 25]-Minibatch[  61-  70, 6.38%]: CE.SM = 1.10882187 * 10240; Err = 0.35703125 * 10240; time = 0.0407s; samplesPerSecond = 251795.0
12/20/2016 15:27:59:  Epoch[10 of 25]-Minibatch[  71-  80, 7.29%]: CE.SM = 1.11975403 * 10240; Err = 0.36513672 * 10240; time = 0.0443s; samplesPerSecond = 231203.4
12/20/2016 15:27:59:  Epoch[10 of 25]-Minibatch[  81-  90, 8.20%]: CE.SM = 1.11434708 * 10240; Err = 0.35908203 * 10240; time = 0.0417s; samplesPerSecond = 245504.7
12/20/2016 15:27:59:  Epoch[10 of 25]-Minibatch[  91- 100, 9.11%]: CE.SM = 1.09469223 * 10240; Err = 0.35742188 * 10240; time = 0.0409s; samplesPerSecond = 250238.3
12/20/2016 15:27:59:  Epoch[10 of 25]-Minibatch[ 101- 110, 10.02%]: CE.SM = 1.09051132 * 10240; Err = 0.35341797 * 10240; time = 0.0462s; samplesPerSecond = 221491.6
12/20/2016 15:27:59:  Epoch[10 of 25]-Minibatch[ 111- 120, 10.93%]: CE.SM = 1.12907028 * 10240; Err = 0.36503906 * 10240; time = 0.0412s; samplesPerSecond = 248658.4
12/20/2016 15:27:59:  Epoch[10 of 25]-Minibatch[ 121- 130, 11.85%]: CE.SM = 1.11979980 * 10240; Err = 0.36269531 * 10240; time = 0.0434s; samplesPerSecond = 236075.2
12/20/2016 15:28:00:  Epoch[10 of 25]-Minibatch[ 131- 140, 12.76%]: CE.SM = 1.10125122 * 10240; Err = 0.35078125 * 10240; time = 0.0415s; samplesPerSecond = 246568.7
12/20/2016 15:28:00:  Epoch[10 of 25]-Minibatch[ 141- 150, 13.67%]: CE.SM = 1.12119751 * 10240; Err = 0.35371094 * 10240; time = 0.0409s; samplesPerSecond = 250299.4
12/20/2016 15:28:00:  Epoch[10 of 25]-Minibatch[ 151- 160, 14.58%]: CE.SM = 1.11571808 * 10240; Err = 0.36611328 * 10240; time = 0.0428s; samplesPerSecond = 239135.0
12/20/2016 15:28:00:  Epoch[10 of 25]-Minibatch[ 161- 170, 15.49%]: CE.SM = 1.09401398 * 10240; Err = 0.34912109 * 10240; time = 0.0418s; samplesPerSecond = 244706.8
12/20/2016 15:28:00:  Epoch[10 of 25]-Minibatch[ 171- 180, 16.40%]: CE.SM = 1.09755707 * 10240; Err = 0.35244141 * 10240; time = 0.0417s; samplesPerSecond = 245475.2
12/20/2016 15:28:00:  Epoch[10 of 25]-Minibatch[ 181- 190, 17.31%]: CE.SM = 1.11199341 * 10240; Err = 0.36074219 * 10240; time = 0.0413s; samplesPerSecond = 247654.1
12/20/2016 15:28:00:  Epoch[10 of 25]-Minibatch[ 191- 200, 18.22%]: CE.SM = 1.11345825 * 10240; Err = 0.35927734 * 10240; time = 0.0422s; samplesPerSecond = 242682.8
12/20/2016 15:28:00:  Epoch[10 of 25]-Minibatch[ 201- 210, 19.13%]: CE.SM = 1.11300354 * 10240; Err = 0.36103516 * 10240; time = 0.0411s; samplesPerSecond = 249354.7
12/20/2016 15:28:00:  Epoch[10 of 25]-Minibatch[ 211- 220, 20.05%]: CE.SM = 1.13068695 * 10240; Err = 0.36464844 * 10240; time = 0.0417s; samplesPerSecond = 245651.9
12/20/2016 15:28:00:  Epoch[10 of 25]-Minibatch[ 221- 230, 20.96%]: CE.SM = 1.12184296 * 10240; Err = 0.36132812 * 10240; time = 0.0422s; samplesPerSecond = 242401.3
12/20/2016 15:28:00:  Epoch[10 of 25]-Minibatch[ 231- 240, 21.87%]: CE.SM = 1.12500305 * 10240; Err = 0.36044922 * 10240; time = 0.0421s; samplesPerSecond = 243068.7
12/20/2016 15:28:00:  Epoch[10 of 25]-Minibatch[ 241- 250, 22.78%]: CE.SM = 1.12481384 * 10240; Err = 0.36074219 * 10240; time = 0.0418s; samplesPerSecond = 244812.1
12/20/2016 15:28:00:  Epoch[10 of 25]-Minibatch[ 251- 260, 23.69%]: CE.SM = 1.11848755 * 10240; Err = 0.35937500 * 10240; time = 0.0413s; samplesPerSecond = 248116.1
12/20/2016 15:28:00:  Epoch[10 of 25]-Minibatch[ 261- 270, 24.60%]: CE.SM = 1.11231079 * 10240; Err = 0.35449219 * 10240; time = 0.0419s; samplesPerSecond = 244665.9
12/20/2016 15:28:00:  Epoch[10 of 25]-Minibatch[ 271- 280, 25.51%]: CE.SM = 1.10046692 * 10240; Err = 0.35156250 * 10240; time = 0.0427s; samplesPerSecond = 239560.2
12/20/2016 15:28:00:  Epoch[10 of 25]-Minibatch[ 281- 290, 26.42%]: CE.SM = 1.11512756 * 10240; Err = 0.35791016 * 10240; time = 0.0417s; samplesPerSecond = 245522.3
12/20/2016 15:28:00:  Epoch[10 of 25]-Minibatch[ 291- 300, 27.33%]: CE.SM = 1.11817017 * 10240; Err = 0.36005859 * 10240; time = 0.0412s; samplesPerSecond = 248598.0
12/20/2016 15:28:00:  Epoch[10 of 25]-Minibatch[ 301- 310, 28.25%]: CE.SM = 1.09852295 * 10240; Err = 0.35341797 * 10240; time = 0.0418s; samplesPerSecond = 245163.8
12/20/2016 15:28:00:  Epoch[10 of 25]-Minibatch[ 311- 320, 29.16%]: CE.SM = 1.12539063 * 10240; Err = 0.35761719 * 10240; time = 0.0414s; samplesPerSecond = 247068.5
12/20/2016 15:28:00:  Epoch[10 of 25]-Minibatch[ 321- 330, 30.07%]: CE.SM = 1.12762146 * 10240; Err = 0.36904297 * 10240; time = 0.0417s; samplesPerSecond = 245811.1
12/20/2016 15:28:00:  Epoch[10 of 25]-Minibatch[ 331- 340, 30.98%]: CE.SM = 1.09183350 * 10240; Err = 0.35283203 * 10240; time = 0.0425s; samplesPerSecond = 241077.3
12/20/2016 15:28:00:  Epoch[10 of 25]-Minibatch[ 341- 350, 31.89%]: CE.SM = 1.10688477 * 10240; Err = 0.35576172 * 10240; time = 0.0423s; samplesPerSecond = 242160.5
12/20/2016 15:28:00:  Epoch[10 of 25]-Minibatch[ 351- 360, 32.80%]: CE.SM = 1.11831055 * 10240; Err = 0.36074219 * 10240; time = 0.0423s; samplesPerSecond = 242332.4
12/20/2016 15:28:00:  Epoch[10 of 25]-Minibatch[ 361- 370, 33.71%]: CE.SM = 1.13406982 * 10240; Err = 0.36103516 * 10240; time = 0.0416s; samplesPerSecond = 246337.4
12/20/2016 15:28:01:  Epoch[10 of 25]-Minibatch[ 371- 380, 34.62%]: CE.SM = 1.11441956 * 10240; Err = 0.35927734 * 10240; time = 0.0416s; samplesPerSecond = 246183.4
12/20/2016 15:28:01:  Epoch[10 of 25]-Minibatch[ 381- 390, 35.54%]: CE.SM = 1.12025757 * 10240; Err = 0.36279297 * 10240; time = 0.0424s; samplesPerSecond = 241412.6
12/20/2016 15:28:01:  Epoch[10 of 25]-Minibatch[ 391- 400, 36.45%]: CE.SM = 1.11292114 * 10240; Err = 0.35781250 * 10240; time = 0.0434s; samplesPerSecond = 235711.2
12/20/2016 15:28:01:  Epoch[10 of 25]-Minibatch[ 401- 410, 37.36%]: CE.SM = 1.09807129 * 10240; Err = 0.34960938 * 10240; time = 0.0427s; samplesPerSecond = 240048.8
12/20/2016 15:28:01:  Epoch[10 of 25]-Minibatch[ 411- 420, 38.27%]: CE.SM = 1.11203003 * 10240; Err = 0.35478516 * 10240; time = 0.0412s; samplesPerSecond = 248694.6
12/20/2016 15:28:01:  Epoch[10 of 25]-Minibatch[ 421- 430, 39.18%]: CE.SM = 1.11154785 * 10240; Err = 0.36064453 * 10240; time = 0.0407s; samplesPerSecond = 251560.0
12/20/2016 15:28:01:  Epoch[10 of 25]-Minibatch[ 431- 440, 40.09%]: CE.SM = 1.11183777 * 10240; Err = 0.36044922 * 10240; time = 0.0424s; samplesPerSecond = 241537.9
12/20/2016 15:28:01:  Epoch[10 of 25]-Minibatch[ 441- 450, 41.00%]: CE.SM = 1.10666199 * 10240; Err = 0.35830078 * 10240; time = 0.0422s; samplesPerSecond = 242441.5
12/20/2016 15:28:01:  Epoch[10 of 25]-Minibatch[ 451- 460, 41.91%]: CE.SM = 1.09715576 * 10240; Err = 0.35117188 * 10240; time = 0.0428s; samplesPerSecond = 239375.4
12/20/2016 15:28:01:  Epoch[10 of 25]-Minibatch[ 461- 470, 42.82%]: CE.SM = 1.11263733 * 10240; Err = 0.36074219 * 10240; time = 0.0429s; samplesPerSecond = 238917.4
12/20/2016 15:28:01:  Epoch[10 of 25]-Minibatch[ 471- 480, 43.74%]: CE.SM = 1.11500854 * 10240; Err = 0.36103516 * 10240; time = 0.0425s; samplesPerSecond = 241048.9
12/20/2016 15:28:01:  Epoch[10 of 25]-Minibatch[ 481- 490, 44.65%]: CE.SM = 1.13566895 * 10240; Err = 0.36591797 * 10240; time = 0.0423s; samplesPerSecond = 242074.7
12/20/2016 15:28:01:  Epoch[10 of 25]-Minibatch[ 491- 500, 45.56%]: CE.SM = 1.10542603 * 10240; Err = 0.35976562 * 10240; time = 0.0428s; samplesPerSecond = 239230.0
12/20/2016 15:28:01:  Epoch[10 of 25]-Minibatch[ 501- 510, 46.47%]: CE.SM = 1.10762329 * 10240; Err = 0.35781250 * 10240; time = 0.0417s; samplesPerSecond = 245699.1
12/20/2016 15:28:01:  Epoch[10 of 25]-Minibatch[ 511- 520, 47.38%]: CE.SM = 1.11765137 * 10240; Err = 0.35761719 * 10240; time = 0.0409s; samplesPerSecond = 250244.4
12/20/2016 15:28:01:  Epoch[10 of 25]-Minibatch[ 521- 530, 48.29%]: CE.SM = 1.09895630 * 10240; Err = 0.35419922 * 10240; time = 0.0424s; samplesPerSecond = 241310.2
12/20/2016 15:28:01:  Epoch[10 of 25]-Minibatch[ 531- 540, 49.20%]: CE.SM = 1.12516479 * 10240; Err = 0.36875000 * 10240; time = 0.0424s; samplesPerSecond = 241259.1
12/20/2016 15:28:01:  Epoch[10 of 25]-Minibatch[ 541- 550, 50.11%]: CE.SM = 1.11702881 * 10240; Err = 0.36279297 * 10240; time = 0.0417s; samplesPerSecond = 245752.1
12/20/2016 15:28:01:  Epoch[10 of 25]-Minibatch[ 551- 560, 51.03%]: CE.SM = 1.10577393 * 10240; Err = 0.35546875 * 10240; time = 0.0420s; samplesPerSecond = 243681.9
12/20/2016 15:28:01:  Epoch[10 of 25]-Minibatch[ 561- 570, 51.94%]: CE.SM = 1.12163696 * 10240; Err = 0.36220703 * 10240; time = 0.0413s; samplesPerSecond = 247983.9
12/20/2016 15:28:01:  Epoch[10 of 25]-Minibatch[ 571- 580, 52.85%]: CE.SM = 1.11975098 * 10240; Err = 0.35703125 * 10240; time = 0.0416s; samplesPerSecond = 245946.9
12/20/2016 15:28:01:  Epoch[10 of 25]-Minibatch[ 581- 590, 53.76%]: CE.SM = 1.11815796 * 10240; Err = 0.35644531 * 10240; time = 0.0427s; samplesPerSecond = 239728.4
12/20/2016 15:28:01:  Epoch[10 of 25]-Minibatch[ 591- 600, 54.67%]: CE.SM = 1.12558594 * 10240; Err = 0.36718750 * 10240; time = 0.0428s; samplesPerSecond = 239207.6
12/20/2016 15:28:02:  Epoch[10 of 25]-Minibatch[ 601- 610, 55.58%]: CE.SM = 1.12739868 * 10240; Err = 0.36152344 * 10240; time = 0.0412s; samplesPerSecond = 248368.9
12/20/2016 15:28:02:  Epoch[10 of 25]-Minibatch[ 611- 620, 56.49%]: CE.SM = 1.09552612 * 10240; Err = 0.35947266 * 10240; time = 0.0422s; samplesPerSecond = 242728.8
12/20/2016 15:28:02:  Epoch[10 of 25]-Minibatch[ 621- 630, 57.40%]: CE.SM = 1.12642822 * 10240; Err = 0.36718750 * 10240; time = 0.0415s; samplesPerSecond = 246937.4
12/20/2016 15:28:02:  Epoch[10 of 25]-Minibatch[ 631- 640, 58.31%]: CE.SM = 1.14060059 * 10240; Err = 0.36455078 * 10240; time = 0.0423s; samplesPerSecond = 242051.8
12/20/2016 15:28:02:  Epoch[10 of 25]-Minibatch[ 641- 650, 59.23%]: CE.SM = 1.10703735 * 10240; Err = 0.35605469 * 10240; time = 0.0430s; samplesPerSecond = 238328.0
12/20/2016 15:28:02:  Epoch[10 of 25]-Minibatch[ 651- 660, 60.14%]: CE.SM = 1.09460449 * 10240; Err = 0.35273437 * 10240; time = 0.0425s; samplesPerSecond = 240941.2
12/20/2016 15:28:02:  Epoch[10 of 25]-Minibatch[ 661- 670, 61.05%]: CE.SM = 1.12447510 * 10240; Err = 0.36357422 * 10240; time = 0.0427s; samplesPerSecond = 240071.3
12/20/2016 15:28:02:  Epoch[10 of 25]-Minibatch[ 671- 680, 61.96%]: CE.SM = 1.12467041 * 10240; Err = 0.36416016 * 10240; time = 0.0425s; samplesPerSecond = 240788.2
12/20/2016 15:28:02:  Epoch[10 of 25]-Minibatch[ 681- 690, 62.87%]: CE.SM = 1.12784424 * 10240; Err = 0.36367187 * 10240; time = 0.0424s; samplesPerSecond = 241367.1
12/20/2016 15:28:02:  Epoch[10 of 25]-Minibatch[ 691- 700, 63.78%]: CE.SM = 1.09744263 * 10240; Err = 0.35195312 * 10240; time = 0.0421s; samplesPerSecond = 243403.9
12/20/2016 15:28:02:  Epoch[10 of 25]-Minibatch[ 701- 710, 64.69%]: CE.SM = 1.11209717 * 10240; Err = 0.35839844 * 10240; time = 0.0419s; samplesPerSecond = 244111.8
12/20/2016 15:28:02:  Epoch[10 of 25]-Minibatch[ 711- 720, 65.60%]: CE.SM = 1.10394287 * 10240; Err = 0.35517578 * 10240; time = 0.0411s; samplesPerSecond = 248906.2
12/20/2016 15:28:02:  Epoch[10 of 25]-Minibatch[ 721- 730, 66.51%]: CE.SM = 1.11655273 * 10240; Err = 0.36044922 * 10240; time = 0.0421s; samplesPerSecond = 243109.1
12/20/2016 15:28:02:  Epoch[10 of 25]-Minibatch[ 731- 740, 67.43%]: CE.SM = 1.12000732 * 10240; Err = 0.36396484 * 10240; time = 0.0424s; samplesPerSecond = 241543.6
12/20/2016 15:28:02:  Epoch[10 of 25]-Minibatch[ 741- 750, 68.34%]: CE.SM = 1.12407837 * 10240; Err = 0.35849609 * 10240; time = 0.0414s; samplesPerSecond = 247420.7
12/20/2016 15:28:02:  Epoch[10 of 25]-Minibatch[ 751- 760, 69.25%]: CE.SM = 1.12064819 * 10240; Err = 0.36416016 * 10240; time = 0.0421s; samplesPerSecond = 242964.9
12/20/2016 15:28:02:  Epoch[10 of 25]-Minibatch[ 761- 770, 70.16%]: CE.SM = 1.11834106 * 10240; Err = 0.36142578 * 10240; time = 0.0424s; samplesPerSecond = 241726.1
12/20/2016 15:28:02:  Epoch[10 of 25]-Minibatch[ 771- 780, 71.07%]: CE.SM = 1.13437500 * 10240; Err = 0.36171875 * 10240; time = 0.0426s; samplesPerSecond = 240307.9
12/20/2016 15:28:02:  Epoch[10 of 25]-Minibatch[ 781- 790, 71.98%]: CE.SM = 1.11066895 * 10240; Err = 0.35869141 * 10240; time = 0.0411s; samplesPerSecond = 249427.6
12/20/2016 15:28:02:  Epoch[10 of 25]-Minibatch[ 791- 800, 72.89%]: CE.SM = 1.11502686 * 10240; Err = 0.36064453 * 10240; time = 0.0413s; samplesPerSecond = 248122.1
12/20/2016 15:28:02:  Epoch[10 of 25]-Minibatch[ 801- 810, 73.80%]: CE.SM = 1.11593628 * 10240; Err = 0.36035156 * 10240; time = 0.0407s; samplesPerSecond = 251313.0
12/20/2016 15:28:02:  Epoch[10 of 25]-Minibatch[ 811- 820, 74.72%]: CE.SM = 1.12749023 * 10240; Err = 0.36328125 * 10240; time = 0.0421s; samplesPerSecond = 242970.7
12/20/2016 15:28:02:  Epoch[10 of 25]-Minibatch[ 821- 830, 75.63%]: CE.SM = 1.10866089 * 10240; Err = 0.35390625 * 10240; time = 0.0419s; samplesPerSecond = 244601.6
12/20/2016 15:28:02:  Epoch[10 of 25]-Minibatch[ 831- 840, 76.54%]: CE.SM = 1.11759033 * 10240; Err = 0.35634766 * 10240; time = 0.0417s; samplesPerSecond = 245457.6
12/20/2016 15:28:03:  Epoch[10 of 25]-Minibatch[ 841- 850, 77.45%]: CE.SM = 1.11689453 * 10240; Err = 0.36054687 * 10240; time = 0.0451s; samplesPerSecond = 227091.3
12/20/2016 15:28:03:  Epoch[10 of 25]-Minibatch[ 851- 860, 78.36%]: CE.SM = 1.11553345 * 10240; Err = 0.35312500 * 10240; time = 0.0437s; samplesPerSecond = 234089.2
12/20/2016 15:28:03:  Epoch[10 of 25]-Minibatch[ 861- 870, 79.27%]: CE.SM = 1.10444946 * 10240; Err = 0.35449219 * 10240; time = 0.0428s; samplesPerSecond = 239436.9
12/20/2016 15:28:03:  Epoch[10 of 25]-Minibatch[ 871- 880, 80.18%]: CE.SM = 1.11693726 * 10240; Err = 0.36318359 * 10240; time = 0.0420s; samplesPerSecond = 243722.5
12/20/2016 15:28:03:  Epoch[10 of 25]-Minibatch[ 881- 890, 81.09%]: CE.SM = 1.13313599 * 10240; Err = 0.36738281 * 10240; time = 0.0423s; samplesPerSecond = 242355.4
12/20/2016 15:28:03:  Epoch[10 of 25]-Minibatch[ 891- 900, 82.00%]: CE.SM = 1.10369263 * 10240; Err = 0.35312500 * 10240; time = 0.0422s; samplesPerSecond = 242619.5
12/20/2016 15:28:03:  Epoch[10 of 25]-Minibatch[ 901- 910, 82.92%]: CE.SM = 1.10064087 * 10240; Err = 0.35244141 * 10240; time = 0.0436s; samplesPerSecond = 235040.3
12/20/2016 15:28:03:  Epoch[10 of 25]-Minibatch[ 911- 920, 83.83%]: CE.SM = 1.11740723 * 10240; Err = 0.35722656 * 10240; time = 0.0426s; samplesPerSecond = 240477.2
12/20/2016 15:28:03:  Epoch[10 of 25]-Minibatch[ 921- 930, 84.74%]: CE.SM = 1.12454834 * 10240; Err = 0.36142578 * 10240; time = 0.0412s; samplesPerSecond = 248399.0
12/20/2016 15:28:03:  Epoch[10 of 25]-Minibatch[ 931- 940, 85.65%]: CE.SM = 1.09973145 * 10240; Err = 0.35654297 * 10240; time = 0.0416s; samplesPerSecond = 246035.6
12/20/2016 15:28:03:  Epoch[10 of 25]-Minibatch[ 941- 950, 86.56%]: CE.SM = 1.11234131 * 10240; Err = 0.35664062 * 10240; time = 0.0408s; samplesPerSecond = 251226.7
12/20/2016 15:28:03:  Epoch[10 of 25]-Minibatch[ 951- 960, 87.47%]: CE.SM = 1.14152832 * 10240; Err = 0.36250000 * 10240; time = 0.0419s; samplesPerSecond = 244338.9
12/20/2016 15:28:03:  Epoch[10 of 25]-Minibatch[ 961- 970, 88.38%]: CE.SM = 1.11563721 * 10240; Err = 0.35751953 * 10240; time = 0.0421s; samplesPerSecond = 243508.0
12/20/2016 15:28:03:  Epoch[10 of 25]-Minibatch[ 971- 980, 89.29%]: CE.SM = 1.11085205 * 10240; Err = 0.35703125 * 10240; time = 0.0427s; samplesPerSecond = 239885.7
12/20/2016 15:28:03:  Epoch[10 of 25]-Minibatch[ 981- 990, 90.21%]: CE.SM = 1.12639160 * 10240; Err = 0.35888672 * 10240; time = 0.0456s; samplesPerSecond = 224428.5
12/20/2016 15:28:03:  Epoch[10 of 25]-Minibatch[ 991-1000, 91.12%]: CE.SM = 1.09903564 * 10240; Err = 0.35517578 * 10240; time = 0.0444s; samplesPerSecond = 230615.0
12/20/2016 15:28:03:  Epoch[10 of 25]-Minibatch[1001-1010, 92.03%]: CE.SM = 1.09277344 * 10240; Err = 0.35673828 * 10240; time = 0.0423s; samplesPerSecond = 242349.7
12/20/2016 15:28:03:  Epoch[10 of 25]-Minibatch[1011-1020, 92.94%]: CE.SM = 1.11043701 * 10240; Err = 0.36328125 * 10240; time = 0.0420s; samplesPerSecond = 243914.1
12/20/2016 15:28:03:  Epoch[10 of 25]-Minibatch[1021-1030, 93.85%]: CE.SM = 1.10891113 * 10240; Err = 0.35869141 * 10240; time = 0.0477s; samplesPerSecond = 214801.1
12/20/2016 15:28:03:  Epoch[10 of 25]-Minibatch[1031-1040, 94.76%]: CE.SM = 1.10446777 * 10240; Err = 0.35869141 * 10240; time = 0.0413s; samplesPerSecond = 247696.0
12/20/2016 15:28:03:  Epoch[10 of 25]-Minibatch[1041-1050, 95.67%]: CE.SM = 1.09616699 * 10240; Err = 0.35673828 * 10240; time = 0.0450s; samplesPerSecond = 227454.5
12/20/2016 15:28:03:  Epoch[10 of 25]-Minibatch[1051-1060, 96.58%]: CE.SM = 1.10015869 * 10240; Err = 0.35527344 * 10240; time = 0.0413s; samplesPerSecond = 248032.0
12/20/2016 15:28:03:  Epoch[10 of 25]-Minibatch[1061-1070, 97.49%]: CE.SM = 1.10728760 * 10240; Err = 0.34990234 * 10240; time = 0.0447s; samplesPerSecond = 228985.4
12/20/2016 15:28:04:  Epoch[10 of 25]-Minibatch[1071-1080, 98.41%]: CE.SM = 1.08934326 * 10240; Err = 0.35654297 * 10240; time = 0.0412s; samplesPerSecond = 248441.2
12/20/2016 15:28:04:  Epoch[10 of 25]-Minibatch[1081-1090, 99.32%]: CE.SM = 1.12740479 * 10240; Err = 0.36767578 * 10240; time = 0.0435s; samplesPerSecond = 235526.8
12/20/2016 15:28:04: Finished Epoch[10 of 25]: [Training] CE.SM = 1.11314280 * 1124823; Err = 0.35881379 * 1124823; totalSamplesSeen = 11248230; learningRatePerSample = 0.003125; epochTime=4.88641s
12/20/2016 15:28:04: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn.10'

12/20/2016 15:28:04: Starting Epoch 11: learning rate per sample = 0.003125  effective momentum = 0.900000  momentum as time constant = 9719.0 samples

12/20/2016 15:28:04: Starting minibatch loop.
12/20/2016 15:28:04:  Epoch[11 of 25]-Minibatch[   1-  10, 0.91%]: CE.SM = 1.07408009 * 10240; Err = 0.34589844 * 10240; time = 0.0498s; samplesPerSecond = 205573.0
12/20/2016 15:28:04:  Epoch[11 of 25]-Minibatch[  11-  20, 1.82%]: CE.SM = 1.08015461 * 10240; Err = 0.35361328 * 10240; time = 0.0538s; samplesPerSecond = 190366.4
12/20/2016 15:28:04:  Epoch[11 of 25]-Minibatch[  21-  30, 2.73%]: CE.SM = 1.06647720 * 10240; Err = 0.34082031 * 10240; time = 0.0417s; samplesPerSecond = 245369.4
12/20/2016 15:28:04:  Epoch[11 of 25]-Minibatch[  31-  40, 3.64%]: CE.SM = 1.09674568 * 10240; Err = 0.35732422 * 10240; time = 0.0405s; samplesPerSecond = 252902.0
12/20/2016 15:28:04:  Epoch[11 of 25]-Minibatch[  41-  50, 4.56%]: CE.SM = 1.08890533 * 10240; Err = 0.35976562 * 10240; time = 0.0412s; samplesPerSecond = 248549.7
12/20/2016 15:28:04:  Epoch[11 of 25]-Minibatch[  51-  60, 5.47%]: CE.SM = 1.09185944 * 10240; Err = 0.35468750 * 10240; time = 0.0483s; samplesPerSecond = 211789.0
12/20/2016 15:28:04:  Epoch[11 of 25]-Minibatch[  61-  70, 6.38%]: CE.SM = 1.09326401 * 10240; Err = 0.35361328 * 10240; time = 0.0440s; samplesPerSecond = 232515.9
12/20/2016 15:28:04:  Epoch[11 of 25]-Minibatch[  71-  80, 7.29%]: CE.SM = 1.07386780 * 10240; Err = 0.35224609 * 10240; time = 0.0422s; samplesPerSecond = 242677.0
12/20/2016 15:28:04:  Epoch[11 of 25]-Minibatch[  81-  90, 8.20%]: CE.SM = 1.10821152 * 10240; Err = 0.35166016 * 10240; time = 0.0405s; samplesPerSecond = 252945.7
12/20/2016 15:28:04:  Epoch[11 of 25]-Minibatch[  91- 100, 9.11%]: CE.SM = 1.07245789 * 10240; Err = 0.34609375 * 10240; time = 0.0403s; samplesPerSecond = 253943.1
12/20/2016 15:28:04:  Epoch[11 of 25]-Minibatch[ 101- 110, 10.02%]: CE.SM = 1.09525909 * 10240; Err = 0.35292969 * 10240; time = 0.0397s; samplesPerSecond = 257824.1
12/20/2016 15:28:04:  Epoch[11 of 25]-Minibatch[ 111- 120, 10.93%]: CE.SM = 1.08380585 * 10240; Err = 0.35009766 * 10240; time = 0.0402s; samplesPerSecond = 254428.9
12/20/2016 15:28:04:  Epoch[11 of 25]-Minibatch[ 121- 130, 11.85%]: CE.SM = 1.07916870 * 10240; Err = 0.35126953 * 10240; time = 0.0399s; samplesPerSecond = 256442.4
12/20/2016 15:28:04:  Epoch[11 of 25]-Minibatch[ 131- 140, 12.76%]: CE.SM = 1.07398834 * 10240; Err = 0.34765625 * 10240; time = 0.0400s; samplesPerSecond = 256211.4
12/20/2016 15:28:04:  Epoch[11 of 25]-Minibatch[ 141- 150, 13.67%]: CE.SM = 1.07526703 * 10240; Err = 0.34775391 * 10240; time = 0.0407s; samplesPerSecond = 251826.0
12/20/2016 15:28:05:  Epoch[11 of 25]-Minibatch[ 151- 160, 14.58%]: CE.SM = 1.09406433 * 10240; Err = 0.35009766 * 10240; time = 0.0412s; samplesPerSecond = 248573.9
12/20/2016 15:28:05:  Epoch[11 of 25]-Minibatch[ 161- 170, 15.49%]: CE.SM = 1.08153839 * 10240; Err = 0.35039063 * 10240; time = 0.0410s; samplesPerSecond = 249610.0
12/20/2016 15:28:05:  Epoch[11 of 25]-Minibatch[ 171- 180, 16.40%]: CE.SM = 1.10425873 * 10240; Err = 0.35810547 * 10240; time = 0.0400s; samplesPerSecond = 256096.0
12/20/2016 15:28:05:  Epoch[11 of 25]-Minibatch[ 181- 190, 17.31%]: CE.SM = 1.08450012 * 10240; Err = 0.34882812 * 10240; time = 0.0404s; samplesPerSecond = 253421.4
12/20/2016 15:28:05:  Epoch[11 of 25]-Minibatch[ 191- 200, 18.22%]: CE.SM = 1.06751709 * 10240; Err = 0.34531250 * 10240; time = 0.0410s; samplesPerSecond = 249981.7
12/20/2016 15:28:05:  Epoch[11 of 25]-Minibatch[ 201- 210, 19.13%]: CE.SM = 1.06950226 * 10240; Err = 0.34453125 * 10240; time = 0.0403s; samplesPerSecond = 254403.6
12/20/2016 15:28:05:  Epoch[11 of 25]-Minibatch[ 211- 220, 20.05%]: CE.SM = 1.09516602 * 10240; Err = 0.34931641 * 10240; time = 0.0411s; samplesPerSecond = 249269.7
12/20/2016 15:28:05:  Epoch[11 of 25]-Minibatch[ 221- 230, 20.96%]: CE.SM = 1.09115448 * 10240; Err = 0.34707031 * 10240; time = 0.0409s; samplesPerSecond = 250470.9
12/20/2016 15:28:05:  Epoch[11 of 25]-Minibatch[ 231- 240, 21.87%]: CE.SM = 1.11629639 * 10240; Err = 0.35654297 * 10240; time = 0.0404s; samplesPerSecond = 253748.0
12/20/2016 15:28:05:  Epoch[11 of 25]-Minibatch[ 241- 250, 22.78%]: CE.SM = 1.09538879 * 10240; Err = 0.35361328 * 10240; time = 0.0405s; samplesPerSecond = 252889.5
12/20/2016 15:28:05:  Epoch[11 of 25]-Minibatch[ 251- 260, 23.69%]: CE.SM = 1.08605652 * 10240; Err = 0.34775391 * 10240; time = 0.0409s; samplesPerSecond = 250667.1
12/20/2016 15:28:05:  Epoch[11 of 25]-Minibatch[ 261- 270, 24.60%]: CE.SM = 1.10585022 * 10240; Err = 0.35654297 * 10240; time = 0.0406s; samplesPerSecond = 252434.4
12/20/2016 15:28:05:  Epoch[11 of 25]-Minibatch[ 271- 280, 25.51%]: CE.SM = 1.07723694 * 10240; Err = 0.34990234 * 10240; time = 0.0405s; samplesPerSecond = 252877.0
12/20/2016 15:28:05:  Epoch[11 of 25]-Minibatch[ 281- 290, 26.42%]: CE.SM = 1.10373535 * 10240; Err = 0.36093750 * 10240; time = 0.0407s; samplesPerSecond = 251313.0
12/20/2016 15:28:05:  Epoch[11 of 25]-Minibatch[ 291- 300, 27.33%]: CE.SM = 1.07893066 * 10240; Err = 0.34882812 * 10240; time = 0.0410s; samplesPerSecond = 249664.8
12/20/2016 15:28:05:  Epoch[11 of 25]-Minibatch[ 301- 310, 28.25%]: CE.SM = 1.08729858 * 10240; Err = 0.35087891 * 10240; time = 0.0404s; samplesPerSecond = 253697.7
12/20/2016 15:28:05:  Epoch[11 of 25]-Minibatch[ 311- 320, 29.16%]: CE.SM = 1.08742065 * 10240; Err = 0.35097656 * 10240; time = 0.0412s; samplesPerSecond = 248700.6
12/20/2016 15:28:05:  Epoch[11 of 25]-Minibatch[ 321- 330, 30.07%]: CE.SM = 1.09171753 * 10240; Err = 0.34814453 * 10240; time = 0.0405s; samplesPerSecond = 253039.4
12/20/2016 15:28:05:  Epoch[11 of 25]-Minibatch[ 331- 340, 30.98%]: CE.SM = 1.10382690 * 10240; Err = 0.35449219 * 10240; time = 0.0412s; samplesPerSecond = 248507.5
12/20/2016 15:28:05:  Epoch[11 of 25]-Minibatch[ 341- 350, 31.89%]: CE.SM = 1.07641602 * 10240; Err = 0.34609375 * 10240; time = 0.0403s; samplesPerSecond = 253993.5
12/20/2016 15:28:05:  Epoch[11 of 25]-Minibatch[ 351- 360, 32.80%]: CE.SM = 1.08507385 * 10240; Err = 0.35263672 * 10240; time = 0.0401s; samplesPerSecond = 255075.4
12/20/2016 15:28:05:  Epoch[11 of 25]-Minibatch[ 361- 370, 33.71%]: CE.SM = 1.08111877 * 10240; Err = 0.35224609 * 10240; time = 0.0403s; samplesPerSecond = 254012.4
12/20/2016 15:28:05:  Epoch[11 of 25]-Minibatch[ 371- 380, 34.62%]: CE.SM = 1.07869873 * 10240; Err = 0.35273437 * 10240; time = 0.0400s; samplesPerSecond = 256198.6
12/20/2016 15:28:05:  Epoch[11 of 25]-Minibatch[ 381- 390, 35.54%]: CE.SM = 1.09391479 * 10240; Err = 0.35253906 * 10240; time = 0.0405s; samplesPerSecond = 252558.9
12/20/2016 15:28:06:  Epoch[11 of 25]-Minibatch[ 391- 400, 36.45%]: CE.SM = 1.09148254 * 10240; Err = 0.34882812 * 10240; time = 0.0396s; samplesPerSecond = 258435.8
12/20/2016 15:28:06:  Epoch[11 of 25]-Minibatch[ 401- 410, 37.36%]: CE.SM = 1.10493469 * 10240; Err = 0.35117188 * 10240; time = 0.0407s; samplesPerSecond = 251665.1
12/20/2016 15:28:06:  Epoch[11 of 25]-Minibatch[ 411- 420, 38.27%]: CE.SM = 1.05597534 * 10240; Err = 0.34003906 * 10240; time = 0.0404s; samplesPerSecond = 253220.9
12/20/2016 15:28:06:  Epoch[11 of 25]-Minibatch[ 421- 430, 39.18%]: CE.SM = 1.10039673 * 10240; Err = 0.35654297 * 10240; time = 0.0409s; samplesPerSecond = 250397.4
12/20/2016 15:28:06:  Epoch[11 of 25]-Minibatch[ 431- 440, 40.09%]: CE.SM = 1.09311218 * 10240; Err = 0.34648438 * 10240; time = 0.0407s; samplesPerSecond = 251720.7
12/20/2016 15:28:06:  Epoch[11 of 25]-Minibatch[ 441- 450, 41.00%]: CE.SM = 1.08522949 * 10240; Err = 0.35087891 * 10240; time = 0.0407s; samplesPerSecond = 251461.1
12/20/2016 15:28:06:  Epoch[11 of 25]-Minibatch[ 451- 460, 41.91%]: CE.SM = 1.06863403 * 10240; Err = 0.34541016 * 10240; time = 0.0406s; samplesPerSecond = 252359.7
12/20/2016 15:28:06:  Epoch[11 of 25]-Minibatch[ 461- 470, 42.82%]: CE.SM = 1.10346375 * 10240; Err = 0.35507813 * 10240; time = 0.0410s; samplesPerSecond = 249701.3
12/20/2016 15:28:06:  Epoch[11 of 25]-Minibatch[ 471- 480, 43.74%]: CE.SM = 1.09988708 * 10240; Err = 0.36044922 * 10240; time = 0.0408s; samplesPerSecond = 251177.4
12/20/2016 15:28:06:  Epoch[11 of 25]-Minibatch[ 481- 490, 44.65%]: CE.SM = 1.09394531 * 10240; Err = 0.34775391 * 10240; time = 0.0411s; samplesPerSecond = 248984.9
12/20/2016 15:28:06:  Epoch[11 of 25]-Minibatch[ 491- 500, 45.56%]: CE.SM = 1.07499390 * 10240; Err = 0.35087891 * 10240; time = 0.0406s; samplesPerSecond = 252073.9
12/20/2016 15:28:06:  Epoch[11 of 25]-Minibatch[ 501- 510, 46.47%]: CE.SM = 1.10949707 * 10240; Err = 0.35781250 * 10240; time = 0.0411s; samplesPerSecond = 249287.9
12/20/2016 15:28:06:  Epoch[11 of 25]-Minibatch[ 511- 520, 47.38%]: CE.SM = 1.07921143 * 10240; Err = 0.34960938 * 10240; time = 0.0409s; samplesPerSecond = 250158.8
12/20/2016 15:28:06:  Epoch[11 of 25]-Minibatch[ 521- 530, 48.29%]: CE.SM = 1.08793945 * 10240; Err = 0.35195312 * 10240; time = 0.0409s; samplesPerSecond = 250636.4
12/20/2016 15:28:06:  Epoch[11 of 25]-Minibatch[ 531- 540, 49.20%]: CE.SM = 1.10858154 * 10240; Err = 0.35517578 * 10240; time = 0.0411s; samplesPerSecond = 249093.9
12/20/2016 15:28:06:  Epoch[11 of 25]-Minibatch[ 541- 550, 50.11%]: CE.SM = 1.08657227 * 10240; Err = 0.35390625 * 10240; time = 0.0402s; samplesPerSecond = 254745.4
12/20/2016 15:28:06:  Epoch[11 of 25]-Minibatch[ 551- 560, 51.03%]: CE.SM = 1.07840576 * 10240; Err = 0.35185547 * 10240; time = 0.0437s; samplesPerSecond = 234566.5
12/20/2016 15:28:06:  Epoch[11 of 25]-Minibatch[ 561- 570, 51.94%]: CE.SM = 1.10354614 * 10240; Err = 0.35117188 * 10240; time = 0.0409s; samplesPerSecond = 250421.9
12/20/2016 15:28:06:  Epoch[11 of 25]-Minibatch[ 571- 580, 52.85%]: CE.SM = 1.11474609 * 10240; Err = 0.35292969 * 10240; time = 0.0399s; samplesPerSecond = 256564.4
12/20/2016 15:28:06:  Epoch[11 of 25]-Minibatch[ 581- 590, 53.76%]: CE.SM = 1.09022827 * 10240; Err = 0.34843750 * 10240; time = 0.0406s; samplesPerSecond = 251962.3
12/20/2016 15:28:06:  Epoch[11 of 25]-Minibatch[ 591- 600, 54.67%]: CE.SM = 1.09744873 * 10240; Err = 0.35371094 * 10240; time = 0.0410s; samplesPerSecond = 249658.7
12/20/2016 15:28:06:  Epoch[11 of 25]-Minibatch[ 601- 610, 55.58%]: CE.SM = 1.10689697 * 10240; Err = 0.34794922 * 10240; time = 0.0402s; samplesPerSecond = 254504.8
12/20/2016 15:28:06:  Epoch[11 of 25]-Minibatch[ 611- 620, 56.49%]: CE.SM = 1.08624268 * 10240; Err = 0.34951172 * 10240; time = 0.0409s; samplesPerSecond = 250673.2
12/20/2016 15:28:06:  Epoch[11 of 25]-Minibatch[ 621- 630, 57.40%]: CE.SM = 1.10289917 * 10240; Err = 0.35312500 * 10240; time = 0.0408s; samplesPerSecond = 250986.5
12/20/2016 15:28:06:  Epoch[11 of 25]-Minibatch[ 631- 640, 58.31%]: CE.SM = 1.06251831 * 10240; Err = 0.33701172 * 10240; time = 0.0408s; samplesPerSecond = 250783.7
12/20/2016 15:28:07:  Epoch[11 of 25]-Minibatch[ 641- 650, 59.23%]: CE.SM = 1.08501587 * 10240; Err = 0.34648438 * 10240; time = 0.0401s; samplesPerSecond = 255202.5
12/20/2016 15:28:07:  Epoch[11 of 25]-Minibatch[ 651- 660, 60.14%]: CE.SM = 1.09047852 * 10240; Err = 0.35244141 * 10240; time = 0.0412s; samplesPerSecond = 248598.0
12/20/2016 15:28:07:  Epoch[11 of 25]-Minibatch[ 661- 670, 61.05%]: CE.SM = 1.09100952 * 10240; Err = 0.35644531 * 10240; time = 0.0410s; samplesPerSecond = 249957.3
12/20/2016 15:28:07:  Epoch[11 of 25]-Minibatch[ 671- 680, 61.96%]: CE.SM = 1.09338989 * 10240; Err = 0.35068359 * 10240; time = 0.0402s; samplesPerSecond = 254808.8
12/20/2016 15:28:07:  Epoch[11 of 25]-Minibatch[ 681- 690, 62.87%]: CE.SM = 1.08222656 * 10240; Err = 0.34960938 * 10240; time = 0.0406s; samplesPerSecond = 252173.3
12/20/2016 15:28:07:  Epoch[11 of 25]-Minibatch[ 691- 700, 63.78%]: CE.SM = 1.07398682 * 10240; Err = 0.35097656 * 10240; time = 0.0398s; samplesPerSecond = 257015.2
12/20/2016 15:28:07:  Epoch[11 of 25]-Minibatch[ 701- 710, 64.69%]: CE.SM = 1.09643555 * 10240; Err = 0.36083984 * 10240; time = 0.0409s; samplesPerSecond = 250317.8
12/20/2016 15:28:07:  Epoch[11 of 25]-Minibatch[ 711- 720, 65.60%]: CE.SM = 1.09683838 * 10240; Err = 0.35566406 * 10240; time = 0.0401s; samplesPerSecond = 255253.4
12/20/2016 15:28:07:  Epoch[11 of 25]-Minibatch[ 721- 730, 66.51%]: CE.SM = 1.09250488 * 10240; Err = 0.35371094 * 10240; time = 0.0399s; samplesPerSecond = 256673.8
12/20/2016 15:28:07:  Epoch[11 of 25]-Minibatch[ 731- 740, 67.43%]: CE.SM = 1.05962524 * 10240; Err = 0.34355469 * 10240; time = 0.0400s; samplesPerSecond = 255910.4
12/20/2016 15:28:07:  Epoch[11 of 25]-Minibatch[ 741- 750, 68.34%]: CE.SM = 1.08675537 * 10240; Err = 0.35839844 * 10240; time = 0.0404s; samplesPerSecond = 253578.3
12/20/2016 15:28:07:  Epoch[11 of 25]-Minibatch[ 751- 760, 69.25%]: CE.SM = 1.08573608 * 10240; Err = 0.34814453 * 10240; time = 0.0404s; samplesPerSecond = 253421.4
12/20/2016 15:28:07:  Epoch[11 of 25]-Minibatch[ 761- 770, 70.16%]: CE.SM = 1.10641479 * 10240; Err = 0.35849609 * 10240; time = 0.0399s; samplesPerSecond = 256506.6
12/20/2016 15:28:07:  Epoch[11 of 25]-Minibatch[ 771- 780, 71.07%]: CE.SM = 1.08406982 * 10240; Err = 0.34814453 * 10240; time = 0.0403s; samplesPerSecond = 254302.5
12/20/2016 15:28:07:  Epoch[11 of 25]-Minibatch[ 781- 790, 71.98%]: CE.SM = 1.10072632 * 10240; Err = 0.35693359 * 10240; time = 0.0399s; samplesPerSecond = 256435.9
12/20/2016 15:28:07:  Epoch[11 of 25]-Minibatch[ 791- 800, 72.89%]: CE.SM = 1.09235840 * 10240; Err = 0.35156250 * 10240; time = 0.0405s; samplesPerSecond = 252758.4
12/20/2016 15:28:07:  Epoch[11 of 25]-Minibatch[ 801- 810, 73.80%]: CE.SM = 1.09342651 * 10240; Err = 0.35390625 * 10240; time = 0.0399s; samplesPerSecond = 256513.0
12/20/2016 15:28:07:  Epoch[11 of 25]-Minibatch[ 811- 820, 74.72%]: CE.SM = 1.08021240 * 10240; Err = 0.35556641 * 10240; time = 0.0404s; samplesPerSecond = 253477.9
12/20/2016 15:28:07:  Epoch[11 of 25]-Minibatch[ 821- 830, 75.63%]: CE.SM = 1.08281860 * 10240; Err = 0.35078125 * 10240; time = 0.0409s; samplesPerSecond = 250146.6
12/20/2016 15:28:07:  Epoch[11 of 25]-Minibatch[ 831- 840, 76.54%]: CE.SM = 1.08178101 * 10240; Err = 0.35009766 * 10240; time = 0.0404s; samplesPerSecond = 253559.5
12/20/2016 15:28:07:  Epoch[11 of 25]-Minibatch[ 841- 850, 77.45%]: CE.SM = 1.09431152 * 10240; Err = 0.35976562 * 10240; time = 0.0402s; samplesPerSecond = 254865.8
12/20/2016 15:28:07:  Epoch[11 of 25]-Minibatch[ 851- 860, 78.36%]: CE.SM = 1.09564819 * 10240; Err = 0.35664062 * 10240; time = 0.0405s; samplesPerSecond = 252951.9
12/20/2016 15:28:07:  Epoch[11 of 25]-Minibatch[ 861- 870, 79.27%]: CE.SM = 1.08950195 * 10240; Err = 0.35195312 * 10240; time = 0.0407s; samplesPerSecond = 251479.7
12/20/2016 15:28:07:  Epoch[11 of 25]-Minibatch[ 871- 880, 80.18%]: CE.SM = 1.09063110 * 10240; Err = 0.34902344 * 10240; time = 0.0404s; samplesPerSecond = 253377.5
12/20/2016 15:28:08:  Epoch[11 of 25]-Minibatch[ 881- 890, 81.09%]: CE.SM = 1.08436279 * 10240; Err = 0.35000000 * 10240; time = 0.0403s; samplesPerSecond = 253798.3
12/20/2016 15:28:08:  Epoch[11 of 25]-Minibatch[ 891- 900, 82.00%]: CE.SM = 1.09423218 * 10240; Err = 0.34804687 * 10240; time = 0.0406s; samplesPerSecond = 252278.9
12/20/2016 15:28:08:  Epoch[11 of 25]-Minibatch[ 901- 910, 82.92%]: CE.SM = 1.10844727 * 10240; Err = 0.36201172 * 10240; time = 0.0400s; samplesPerSecond = 256025.6
12/20/2016 15:28:08:  Epoch[11 of 25]-Minibatch[ 911- 920, 83.83%]: CE.SM = 1.08268433 * 10240; Err = 0.34990234 * 10240; time = 0.0403s; samplesPerSecond = 253792.0
12/20/2016 15:28:08:  Epoch[11 of 25]-Minibatch[ 921- 930, 84.74%]: CE.SM = 1.10942383 * 10240; Err = 0.36083984 * 10240; time = 0.0405s; samplesPerSecond = 253001.9
12/20/2016 15:28:08:  Epoch[11 of 25]-Minibatch[ 931- 940, 85.65%]: CE.SM = 1.08257446 * 10240; Err = 0.35146484 * 10240; time = 0.0401s; samplesPerSecond = 255227.9
12/20/2016 15:28:08:  Epoch[11 of 25]-Minibatch[ 941- 950, 86.56%]: CE.SM = 1.08376465 * 10240; Err = 0.35488281 * 10240; time = 0.0406s; samplesPerSecond = 252428.1
12/20/2016 15:28:08:  Epoch[11 of 25]-Minibatch[ 951- 960, 87.47%]: CE.SM = 1.09954834 * 10240; Err = 0.35263672 * 10240; time = 0.0402s; samplesPerSecond = 254675.7
12/20/2016 15:28:08:  Epoch[11 of 25]-Minibatch[ 961- 970, 88.38%]: CE.SM = 1.08288574 * 10240; Err = 0.34667969 * 10240; time = 0.0409s; samplesPerSecond = 250483.1
12/20/2016 15:28:08:  Epoch[11 of 25]-Minibatch[ 971- 980, 89.29%]: CE.SM = 1.09716797 * 10240; Err = 0.35732422 * 10240; time = 0.0415s; samplesPerSecond = 246723.2
12/20/2016 15:28:08:  Epoch[11 of 25]-Minibatch[ 981- 990, 90.21%]: CE.SM = 1.09445801 * 10240; Err = 0.35693359 * 10240; time = 0.0401s; samplesPerSecond = 255603.8
12/20/2016 15:28:08:  Epoch[11 of 25]-Minibatch[ 991-1000, 91.12%]: CE.SM = 1.09107666 * 10240; Err = 0.35683594 * 10240; time = 0.0413s; samplesPerSecond = 247881.9
12/20/2016 15:28:08:  Epoch[11 of 25]-Minibatch[1001-1010, 92.03%]: CE.SM = 1.06068115 * 10240; Err = 0.34335938 * 10240; time = 0.0398s; samplesPerSecond = 257305.8
12/20/2016 15:28:08:  Epoch[11 of 25]-Minibatch[1011-1020, 92.94%]: CE.SM = 1.08872070 * 10240; Err = 0.34921875 * 10240; time = 0.0402s; samplesPerSecond = 254929.3
12/20/2016 15:28:08:  Epoch[11 of 25]-Minibatch[1021-1030, 93.85%]: CE.SM = 1.10858154 * 10240; Err = 0.36318359 * 10240; time = 0.0408s; samplesPerSecond = 251066.5
12/20/2016 15:28:08:  Epoch[11 of 25]-Minibatch[1031-1040, 94.76%]: CE.SM = 1.08892822 * 10240; Err = 0.35126953 * 10240; time = 0.0411s; samplesPerSecond = 249342.6
12/20/2016 15:28:08:  Epoch[11 of 25]-Minibatch[1041-1050, 95.67%]: CE.SM = 1.09562988 * 10240; Err = 0.35087891 * 10240; time = 0.0403s; samplesPerSecond = 253785.7
12/20/2016 15:28:08:  Epoch[11 of 25]-Minibatch[1051-1060, 96.58%]: CE.SM = 1.08806152 * 10240; Err = 0.35771484 * 10240; time = 0.0401s; samplesPerSecond = 255374.3
12/20/2016 15:28:08:  Epoch[11 of 25]-Minibatch[1061-1070, 97.49%]: CE.SM = 1.07353516 * 10240; Err = 0.35058594 * 10240; time = 0.0402s; samplesPerSecond = 254656.7
12/20/2016 15:28:08:  Epoch[11 of 25]-Minibatch[1071-1080, 98.41%]: CE.SM = 1.08787842 * 10240; Err = 0.35058594 * 10240; time = 0.0408s; samplesPerSecond = 250691.6
12/20/2016 15:28:08:  Epoch[11 of 25]-Minibatch[1081-1090, 99.32%]: CE.SM = 1.08133545 * 10240; Err = 0.35429688 * 10240; time = 0.0401s; samplesPerSecond = 255431.7
12/20/2016 15:28:08: Finished Epoch[11 of 25]: [Training] CE.SM = 1.08881631 * 1124823; Err = 0.35184025 * 1124823; totalSamplesSeen = 12373053; learningRatePerSample = 0.003125; epochTime=4.68332s
12/20/2016 15:28:08: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn.11'

12/20/2016 15:28:08: Starting Epoch 12: learning rate per sample = 0.003125  effective momentum = 0.900000  momentum as time constant = 9719.0 samples

12/20/2016 15:28:09: Starting minibatch loop.
12/20/2016 15:28:09:  Epoch[12 of 25]-Minibatch[   1-  10, 0.91%]: CE.SM = 1.08012915 * 10240; Err = 0.34892578 * 10240; time = 0.0503s; samplesPerSecond = 203505.7
12/20/2016 15:28:09:  Epoch[12 of 25]-Minibatch[  11-  20, 1.82%]: CE.SM = 1.05921087 * 10240; Err = 0.34492187 * 10240; time = 0.0420s; samplesPerSecond = 243618.1
12/20/2016 15:28:09:  Epoch[12 of 25]-Minibatch[  21-  30, 2.73%]: CE.SM = 1.04964142 * 10240; Err = 0.34345703 * 10240; time = 0.0409s; samplesPerSecond = 250293.3
12/20/2016 15:28:09:  Epoch[12 of 25]-Minibatch[  31-  40, 3.64%]: CE.SM = 1.04519901 * 10240; Err = 0.34023437 * 10240; time = 0.0412s; samplesPerSecond = 248386.9
12/20/2016 15:28:09:  Epoch[12 of 25]-Minibatch[  41-  50, 4.56%]: CE.SM = 1.03802223 * 10240; Err = 0.34414062 * 10240; time = 0.0407s; samplesPerSecond = 251492.0
12/20/2016 15:28:09:  Epoch[12 of 25]-Minibatch[  51-  60, 5.47%]: CE.SM = 1.05004997 * 10240; Err = 0.34267578 * 10240; time = 0.0410s; samplesPerSecond = 249750.0
12/20/2016 15:28:09:  Epoch[12 of 25]-Minibatch[  61-  70, 6.38%]: CE.SM = 1.06498260 * 10240; Err = 0.34140625 * 10240; time = 0.0416s; samplesPerSecond = 246017.8
12/20/2016 15:28:09:  Epoch[12 of 25]-Minibatch[  71-  80, 7.29%]: CE.SM = 1.04932098 * 10240; Err = 0.33867188 * 10240; time = 0.0415s; samplesPerSecond = 246758.9
12/20/2016 15:28:09:  Epoch[12 of 25]-Minibatch[  81-  90, 8.20%]: CE.SM = 1.06014557 * 10240; Err = 0.34531250 * 10240; time = 0.0412s; samplesPerSecond = 248773.1
12/20/2016 15:28:09:  Epoch[12 of 25]-Minibatch[  91- 100, 9.11%]: CE.SM = 1.05608521 * 10240; Err = 0.33916016 * 10240; time = 0.0407s; samplesPerSecond = 251362.4
12/20/2016 15:28:09:  Epoch[12 of 25]-Minibatch[ 101- 110, 10.02%]: CE.SM = 1.06178818 * 10240; Err = 0.33808594 * 10240; time = 0.0401s; samplesPerSecond = 255088.1
12/20/2016 15:28:09:  Epoch[12 of 25]-Minibatch[ 111- 120, 10.93%]: CE.SM = 1.03916779 * 10240; Err = 0.33925781 * 10240; time = 0.0410s; samplesPerSecond = 250024.4
12/20/2016 15:28:09:  Epoch[12 of 25]-Minibatch[ 121- 130, 11.85%]: CE.SM = 1.03408966 * 10240; Err = 0.33125000 * 10240; time = 0.0403s; samplesPerSecond = 254075.4
12/20/2016 15:28:09:  Epoch[12 of 25]-Minibatch[ 131- 140, 12.76%]: CE.SM = 1.07578278 * 10240; Err = 0.34960938 * 10240; time = 0.0403s; samplesPerSecond = 253911.6
12/20/2016 15:28:09:  Epoch[12 of 25]-Minibatch[ 141- 150, 13.67%]: CE.SM = 1.06610718 * 10240; Err = 0.34550781 * 10240; time = 0.0400s; samplesPerSecond = 256243.4
12/20/2016 15:28:09:  Epoch[12 of 25]-Minibatch[ 151- 160, 14.58%]: CE.SM = 1.06508484 * 10240; Err = 0.34531250 * 10240; time = 0.0442s; samplesPerSecond = 231847.3
12/20/2016 15:28:09:  Epoch[12 of 25]-Minibatch[ 161- 170, 15.49%]: CE.SM = 1.06285095 * 10240; Err = 0.35097656 * 10240; time = 0.0394s; samplesPerSecond = 259746.8
12/20/2016 15:28:09:  Epoch[12 of 25]-Minibatch[ 171- 180, 16.40%]: CE.SM = 1.09321289 * 10240; Err = 0.35332031 * 10240; time = 0.0401s; samplesPerSecond = 255399.8
12/20/2016 15:28:09:  Epoch[12 of 25]-Minibatch[ 181- 190, 17.31%]: CE.SM = 1.02884979 * 10240; Err = 0.33330078 * 10240; time = 0.0409s; samplesPerSecond = 250556.7
12/20/2016 15:28:09:  Epoch[12 of 25]-Minibatch[ 191- 200, 18.22%]: CE.SM = 1.07676086 * 10240; Err = 0.34863281 * 10240; time = 0.0400s; samplesPerSecond = 255968.0
12/20/2016 15:28:09:  Epoch[12 of 25]-Minibatch[ 201- 210, 19.13%]: CE.SM = 1.04379578 * 10240; Err = 0.33779297 * 10240; time = 0.0398s; samplesPerSecond = 257487.0
12/20/2016 15:28:10:  Epoch[12 of 25]-Minibatch[ 211- 220, 20.05%]: CE.SM = 1.07596893 * 10240; Err = 0.35019531 * 10240; time = 0.0414s; samplesPerSecond = 247564.2
12/20/2016 15:28:10:  Epoch[12 of 25]-Minibatch[ 221- 230, 20.96%]: CE.SM = 1.08504028 * 10240; Err = 0.34833984 * 10240; time = 0.0420s; samplesPerSecond = 244065.2
12/20/2016 15:28:10:  Epoch[12 of 25]-Minibatch[ 231- 240, 21.87%]: CE.SM = 1.07382660 * 10240; Err = 0.35117188 * 10240; time = 0.0404s; samplesPerSecond = 253616.0
12/20/2016 15:28:10:  Epoch[12 of 25]-Minibatch[ 241- 250, 22.78%]: CE.SM = 1.05936890 * 10240; Err = 0.34472656 * 10240; time = 0.0403s; samplesPerSecond = 253999.8
12/20/2016 15:28:10:  Epoch[12 of 25]-Minibatch[ 251- 260, 23.69%]: CE.SM = 1.06685486 * 10240; Err = 0.34462891 * 10240; time = 0.0407s; samplesPerSecond = 251325.3
12/20/2016 15:28:10:  Epoch[12 of 25]-Minibatch[ 261- 270, 24.60%]: CE.SM = 1.05781555 * 10240; Err = 0.34375000 * 10240; time = 0.0423s; samplesPerSecond = 242114.7
12/20/2016 15:28:10:  Epoch[12 of 25]-Minibatch[ 271- 280, 25.51%]: CE.SM = 1.05258179 * 10240; Err = 0.34443359 * 10240; time = 0.0406s; samplesPerSecond = 251937.5
12/20/2016 15:28:10:  Epoch[12 of 25]-Minibatch[ 281- 290, 26.42%]: CE.SM = 1.06917114 * 10240; Err = 0.34892578 * 10240; time = 0.0401s; samplesPerSecond = 255348.9
12/20/2016 15:28:10:  Epoch[12 of 25]-Minibatch[ 291- 300, 27.33%]: CE.SM = 1.06975708 * 10240; Err = 0.34628906 * 10240; time = 0.0400s; samplesPerSecond = 256301.2
12/20/2016 15:28:10:  Epoch[12 of 25]-Minibatch[ 301- 310, 28.25%]: CE.SM = 1.06213379 * 10240; Err = 0.34794922 * 10240; time = 0.0407s; samplesPerSecond = 251362.4
12/20/2016 15:28:10:  Epoch[12 of 25]-Minibatch[ 311- 320, 29.16%]: CE.SM = 1.08763428 * 10240; Err = 0.35058594 * 10240; time = 0.0406s; samplesPerSecond = 252235.4
12/20/2016 15:28:10:  Epoch[12 of 25]-Minibatch[ 321- 330, 30.07%]: CE.SM = 1.08007812 * 10240; Err = 0.34746094 * 10240; time = 0.0410s; samplesPerSecond = 249908.5
12/20/2016 15:28:10:  Epoch[12 of 25]-Minibatch[ 331- 340, 30.98%]: CE.SM = 1.04908752 * 10240; Err = 0.34443359 * 10240; time = 0.0409s; samplesPerSecond = 250421.9
12/20/2016 15:28:10:  Epoch[12 of 25]-Minibatch[ 341- 350, 31.89%]: CE.SM = 1.08341980 * 10240; Err = 0.35058594 * 10240; time = 0.0406s; samplesPerSecond = 251974.7
12/20/2016 15:28:10:  Epoch[12 of 25]-Minibatch[ 351- 360, 32.80%]: CE.SM = 1.06504517 * 10240; Err = 0.34716797 * 10240; time = 0.0413s; samplesPerSecond = 247827.9
12/20/2016 15:28:10:  Epoch[12 of 25]-Minibatch[ 361- 370, 33.71%]: CE.SM = 1.07315674 * 10240; Err = 0.34804687 * 10240; time = 0.0406s; samplesPerSecond = 252049.1
12/20/2016 15:28:10:  Epoch[12 of 25]-Minibatch[ 371- 380, 34.62%]: CE.SM = 1.07452393 * 10240; Err = 0.35234375 * 10240; time = 0.0402s; samplesPerSecond = 254530.1
12/20/2016 15:28:10:  Epoch[12 of 25]-Minibatch[ 381- 390, 35.54%]: CE.SM = 1.10630798 * 10240; Err = 0.35625000 * 10240; time = 0.0405s; samplesPerSecond = 253058.2
12/20/2016 15:28:10:  Epoch[12 of 25]-Minibatch[ 391- 400, 36.45%]: CE.SM = 1.02747192 * 10240; Err = 0.33662109 * 10240; time = 0.0404s; samplesPerSecond = 253590.9
12/20/2016 15:28:10:  Epoch[12 of 25]-Minibatch[ 401- 410, 37.36%]: CE.SM = 1.05968323 * 10240; Err = 0.34687500 * 10240; time = 0.0406s; samplesPerSecond = 252322.4
12/20/2016 15:28:10:  Epoch[12 of 25]-Minibatch[ 411- 420, 38.27%]: CE.SM = 1.06748657 * 10240; Err = 0.34462891 * 10240; time = 0.0403s; samplesPerSecond = 253943.1
12/20/2016 15:28:10:  Epoch[12 of 25]-Minibatch[ 421- 430, 39.18%]: CE.SM = 1.06712952 * 10240; Err = 0.34638672 * 10240; time = 0.0401s; samplesPerSecond = 255056.3
12/20/2016 15:28:10:  Epoch[12 of 25]-Minibatch[ 431- 440, 40.09%]: CE.SM = 1.06310120 * 10240; Err = 0.34765625 * 10240; time = 0.0400s; samplesPerSecond = 256211.4
12/20/2016 15:28:10:  Epoch[12 of 25]-Minibatch[ 441- 450, 41.00%]: CE.SM = 1.05367126 * 10240; Err = 0.34199219 * 10240; time = 0.0401s; samplesPerSecond = 255049.9
12/20/2016 15:28:11:  Epoch[12 of 25]-Minibatch[ 451- 460, 41.91%]: CE.SM = 1.07370605 * 10240; Err = 0.34824219 * 10240; time = 0.0400s; samplesPerSecond = 256172.9
12/20/2016 15:28:11:  Epoch[12 of 25]-Minibatch[ 461- 470, 42.82%]: CE.SM = 1.07284546 * 10240; Err = 0.34902344 * 10240; time = 0.0401s; samplesPerSecond = 255412.6
12/20/2016 15:28:11:  Epoch[12 of 25]-Minibatch[ 471- 480, 43.74%]: CE.SM = 1.06645813 * 10240; Err = 0.35156250 * 10240; time = 0.0404s; samplesPerSecond = 253540.7
12/20/2016 15:28:11:  Epoch[12 of 25]-Minibatch[ 481- 490, 44.65%]: CE.SM = 1.07221680 * 10240; Err = 0.34482422 * 10240; time = 0.0410s; samplesPerSecond = 249871.9
12/20/2016 15:28:11:  Epoch[12 of 25]-Minibatch[ 491- 500, 45.56%]: CE.SM = 1.07247925 * 10240; Err = 0.34628906 * 10240; time = 0.0408s; samplesPerSecond = 251146.6
12/20/2016 15:28:11:  Epoch[12 of 25]-Minibatch[ 501- 510, 46.47%]: CE.SM = 1.07966309 * 10240; Err = 0.34541016 * 10240; time = 0.0403s; samplesPerSecond = 254271.0
12/20/2016 15:28:11:  Epoch[12 of 25]-Minibatch[ 511- 520, 47.38%]: CE.SM = 1.06143799 * 10240; Err = 0.34921875 * 10240; time = 0.0410s; samplesPerSecond = 249780.5
12/20/2016 15:28:11:  Epoch[12 of 25]-Minibatch[ 521- 530, 48.29%]: CE.SM = 1.07542725 * 10240; Err = 0.34462891 * 10240; time = 0.0410s; samplesPerSecond = 249920.7
12/20/2016 15:28:11:  Epoch[12 of 25]-Minibatch[ 531- 540, 49.20%]: CE.SM = 1.06615601 * 10240; Err = 0.34404297 * 10240; time = 0.0401s; samplesPerSecond = 255527.3
12/20/2016 15:28:11:  Epoch[12 of 25]-Minibatch[ 541- 550, 50.11%]: CE.SM = 1.07817993 * 10240; Err = 0.34658203 * 10240; time = 0.0399s; samplesPerSecond = 256628.7
12/20/2016 15:28:11:  Epoch[12 of 25]-Minibatch[ 551- 560, 51.03%]: CE.SM = 1.05161743 * 10240; Err = 0.33574219 * 10240; time = 0.0400s; samplesPerSecond = 256172.9
12/20/2016 15:28:11:  Epoch[12 of 25]-Minibatch[ 561- 570, 51.94%]: CE.SM = 1.05651245 * 10240; Err = 0.33837891 * 10240; time = 0.0409s; samplesPerSecond = 250624.1
12/20/2016 15:28:11:  Epoch[12 of 25]-Minibatch[ 571- 580, 52.85%]: CE.SM = 1.06690063 * 10240; Err = 0.34541016 * 10240; time = 0.0406s; samplesPerSecond = 251937.5
12/20/2016 15:28:11:  Epoch[12 of 25]-Minibatch[ 581- 590, 53.76%]: CE.SM = 1.06219482 * 10240; Err = 0.35019531 * 10240; time = 0.0405s; samplesPerSecond = 252814.5
12/20/2016 15:28:11:  Epoch[12 of 25]-Minibatch[ 591- 600, 54.67%]: CE.SM = 1.06577759 * 10240; Err = 0.34492187 * 10240; time = 0.0399s; samplesPerSecond = 256886.3
12/20/2016 15:28:11:  Epoch[12 of 25]-Minibatch[ 601- 610, 55.58%]: CE.SM = 1.06442261 * 10240; Err = 0.34384766 * 10240; time = 0.0409s; samplesPerSecond = 250483.1
12/20/2016 15:28:11:  Epoch[12 of 25]-Minibatch[ 611- 620, 56.49%]: CE.SM = 1.06412964 * 10240; Err = 0.34951172 * 10240; time = 0.0398s; samplesPerSecond = 257551.7
12/20/2016 15:28:11:  Epoch[12 of 25]-Minibatch[ 621- 630, 57.40%]: CE.SM = 1.04780273 * 10240; Err = 0.34179688 * 10240; time = 0.0406s; samplesPerSecond = 252409.5
12/20/2016 15:28:11:  Epoch[12 of 25]-Minibatch[ 631- 640, 58.31%]: CE.SM = 1.04975586 * 10240; Err = 0.33662109 * 10240; time = 0.0409s; samplesPerSecond = 250415.7
12/20/2016 15:28:11:  Epoch[12 of 25]-Minibatch[ 641- 650, 59.23%]: CE.SM = 1.06049805 * 10240; Err = 0.35058594 * 10240; time = 0.0407s; samplesPerSecond = 251393.2
12/20/2016 15:28:11:  Epoch[12 of 25]-Minibatch[ 651- 660, 60.14%]: CE.SM = 1.04561157 * 10240; Err = 0.33945313 * 10240; time = 0.0409s; samplesPerSecond = 250556.7
12/20/2016 15:28:11:  Epoch[12 of 25]-Minibatch[ 661- 670, 61.05%]: CE.SM = 1.05741577 * 10240; Err = 0.34345703 * 10240; time = 0.0396s; samplesPerSecond = 258416.2
12/20/2016 15:28:11:  Epoch[12 of 25]-Minibatch[ 671- 680, 61.96%]: CE.SM = 1.07796021 * 10240; Err = 0.34941406 * 10240; time = 0.0401s; samplesPerSecond = 255234.3
12/20/2016 15:28:11:  Epoch[12 of 25]-Minibatch[ 681- 690, 62.87%]: CE.SM = 1.06064453 * 10240; Err = 0.34462891 * 10240; time = 0.0400s; samplesPerSecond = 255884.9
12/20/2016 15:28:11:  Epoch[12 of 25]-Minibatch[ 691- 700, 63.78%]: CE.SM = 1.06778564 * 10240; Err = 0.34619141 * 10240; time = 0.0402s; samplesPerSecond = 254618.7
12/20/2016 15:28:12:  Epoch[12 of 25]-Minibatch[ 701- 710, 64.69%]: CE.SM = 1.07796631 * 10240; Err = 0.35302734 * 10240; time = 0.0395s; samplesPerSecond = 258997.9
12/20/2016 15:28:12:  Epoch[12 of 25]-Minibatch[ 711- 720, 65.60%]: CE.SM = 1.05890503 * 10240; Err = 0.34072266 * 10240; time = 0.0399s; samplesPerSecond = 256480.9
12/20/2016 15:28:12:  Epoch[12 of 25]-Minibatch[ 721- 730, 66.51%]: CE.SM = 1.06159058 * 10240; Err = 0.34179688 * 10240; time = 0.0401s; samplesPerSecond = 255540.0
12/20/2016 15:28:12:  Epoch[12 of 25]-Minibatch[ 731- 740, 67.43%]: CE.SM = 1.07587891 * 10240; Err = 0.34316406 * 10240; time = 0.0402s; samplesPerSecond = 255037.2
12/20/2016 15:28:12:  Epoch[12 of 25]-Minibatch[ 741- 750, 68.34%]: CE.SM = 1.09038696 * 10240; Err = 0.34570312 * 10240; time = 0.0406s; samplesPerSecond = 252397.0
12/20/2016 15:28:12:  Epoch[12 of 25]-Minibatch[ 751- 760, 69.25%]: CE.SM = 1.07778320 * 10240; Err = 0.35126953 * 10240; time = 0.0400s; samplesPerSecond = 255808.1
12/20/2016 15:28:12:  Epoch[12 of 25]-Minibatch[ 761- 770, 70.16%]: CE.SM = 1.04916382 * 10240; Err = 0.34277344 * 10240; time = 0.0406s; samplesPerSecond = 252247.8
12/20/2016 15:28:12:  Epoch[12 of 25]-Minibatch[ 771- 780, 71.07%]: CE.SM = 1.06668091 * 10240; Err = 0.34658203 * 10240; time = 0.0399s; samplesPerSecond = 256474.5
12/20/2016 15:28:12:  Epoch[12 of 25]-Minibatch[ 781- 790, 71.98%]: CE.SM = 1.05971680 * 10240; Err = 0.33964844 * 10240; time = 0.0403s; samplesPerSecond = 254182.6
12/20/2016 15:28:12:  Epoch[12 of 25]-Minibatch[ 791- 800, 72.89%]: CE.SM = 1.05987549 * 10240; Err = 0.34384766 * 10240; time = 0.0398s; samplesPerSecond = 257577.7
12/20/2016 15:28:12:  Epoch[12 of 25]-Minibatch[ 801- 810, 73.80%]: CE.SM = 1.08850098 * 10240; Err = 0.35136719 * 10240; time = 0.0406s; samplesPerSecond = 252216.7
12/20/2016 15:28:12:  Epoch[12 of 25]-Minibatch[ 811- 820, 74.72%]: CE.SM = 1.07214966 * 10240; Err = 0.34853516 * 10240; time = 0.0406s; samplesPerSecond = 251949.9
12/20/2016 15:28:12:  Epoch[12 of 25]-Minibatch[ 821- 830, 75.63%]: CE.SM = 1.06380615 * 10240; Err = 0.34414062 * 10240; time = 0.0413s; samplesPerSecond = 247875.9
12/20/2016 15:28:12:  Epoch[12 of 25]-Minibatch[ 831- 840, 76.54%]: CE.SM = 1.06734009 * 10240; Err = 0.34765625 * 10240; time = 0.0407s; samplesPerSecond = 251399.4
12/20/2016 15:28:12:  Epoch[12 of 25]-Minibatch[ 841- 850, 77.45%]: CE.SM = 1.06422729 * 10240; Err = 0.34980469 * 10240; time = 0.0409s; samplesPerSecond = 250250.5
12/20/2016 15:28:12:  Epoch[12 of 25]-Minibatch[ 851- 860, 78.36%]: CE.SM = 1.06320801 * 10240; Err = 0.34433594 * 10240; time = 0.0409s; samplesPerSecond = 250470.9
12/20/2016 15:28:12:  Epoch[12 of 25]-Minibatch[ 861- 870, 79.27%]: CE.SM = 1.07850952 * 10240; Err = 0.34638672 * 10240; time = 0.0405s; samplesPerSecond = 252608.7
12/20/2016 15:28:12:  Epoch[12 of 25]-Minibatch[ 871- 880, 80.18%]: CE.SM = 1.07395020 * 10240; Err = 0.34765625 * 10240; time = 0.0412s; samplesPerSecond = 248815.5
12/20/2016 15:28:12:  Epoch[12 of 25]-Minibatch[ 881- 890, 81.09%]: CE.SM = 1.09028931 * 10240; Err = 0.34433594 * 10240; time = 0.0406s; samplesPerSecond = 252421.9
12/20/2016 15:28:12:  Epoch[12 of 25]-Minibatch[ 891- 900, 82.00%]: CE.SM = 1.07391968 * 10240; Err = 0.34648438 * 10240; time = 0.0411s; samplesPerSecond = 249045.4
12/20/2016 15:28:12:  Epoch[12 of 25]-Minibatch[ 901- 910, 82.92%]: CE.SM = 1.07014160 * 10240; Err = 0.34628906 * 10240; time = 0.0409s; samplesPerSecond = 250483.1
12/20/2016 15:28:12:  Epoch[12 of 25]-Minibatch[ 911- 920, 83.83%]: CE.SM = 1.06394043 * 10240; Err = 0.33769531 * 10240; time = 0.0407s; samplesPerSecond = 251498.2
12/20/2016 15:28:12:  Epoch[12 of 25]-Minibatch[ 921- 930, 84.74%]: CE.SM = 1.07266846 * 10240; Err = 0.34453125 * 10240; time = 0.0407s; samplesPerSecond = 251374.7
12/20/2016 15:28:12:  Epoch[12 of 25]-Minibatch[ 931- 940, 85.65%]: CE.SM = 1.07185059 * 10240; Err = 0.34765625 * 10240; time = 0.0400s; samplesPerSecond = 256057.6
12/20/2016 15:28:12:  Epoch[12 of 25]-Minibatch[ 941- 950, 86.56%]: CE.SM = 1.06099243 * 10240; Err = 0.34726563 * 10240; time = 0.0421s; samplesPerSecond = 243016.8
12/20/2016 15:28:13:  Epoch[12 of 25]-Minibatch[ 951- 960, 87.47%]: CE.SM = 1.07150269 * 10240; Err = 0.35107422 * 10240; time = 0.0402s; samplesPerSecond = 254789.7
12/20/2016 15:28:13:  Epoch[12 of 25]-Minibatch[ 961- 970, 88.38%]: CE.SM = 1.07643433 * 10240; Err = 0.34638672 * 10240; time = 0.0401s; samplesPerSecond = 255457.2
12/20/2016 15:28:13:  Epoch[12 of 25]-Minibatch[ 971- 980, 89.29%]: CE.SM = 1.06950684 * 10240; Err = 0.34550781 * 10240; time = 0.0398s; samplesPerSecond = 257105.6
12/20/2016 15:28:13:  Epoch[12 of 25]-Minibatch[ 981- 990, 90.21%]: CE.SM = 1.06688232 * 10240; Err = 0.34863281 * 10240; time = 0.0414s; samplesPerSecond = 247498.4
12/20/2016 15:28:13:  Epoch[12 of 25]-Minibatch[ 991-1000, 91.12%]: CE.SM = 1.07202148 * 10240; Err = 0.34667969 * 10240; time = 0.0405s; samplesPerSecond = 253133.3
12/20/2016 15:28:13:  Epoch[12 of 25]-Minibatch[1001-1010, 92.03%]: CE.SM = 1.07912598 * 10240; Err = 0.34833984 * 10240; time = 0.0497s; samplesPerSecond = 206023.8
12/20/2016 15:28:13:  Epoch[12 of 25]-Minibatch[1011-1020, 92.94%]: CE.SM = 1.05659180 * 10240; Err = 0.33974609 * 10240; time = 0.0477s; samplesPerSecond = 214643.6
12/20/2016 15:28:13:  Epoch[12 of 25]-Minibatch[1021-1030, 93.85%]: CE.SM = 1.04733887 * 10240; Err = 0.34277344 * 10240; time = 0.0459s; samplesPerSecond = 223020.8
12/20/2016 15:28:13:  Epoch[12 of 25]-Minibatch[1031-1040, 94.76%]: CE.SM = 1.07202148 * 10240; Err = 0.34580078 * 10240; time = 0.0412s; samplesPerSecond = 248712.7
12/20/2016 15:28:13:  Epoch[12 of 25]-Minibatch[1041-1050, 95.67%]: CE.SM = 1.05037842 * 10240; Err = 0.34384766 * 10240; time = 0.0437s; samplesPerSecond = 234566.5
12/20/2016 15:28:13:  Epoch[12 of 25]-Minibatch[1051-1060, 96.58%]: CE.SM = 1.07286377 * 10240; Err = 0.34833984 * 10240; time = 0.0410s; samplesPerSecond = 249969.5
12/20/2016 15:28:13:  Epoch[12 of 25]-Minibatch[1061-1070, 97.49%]: CE.SM = 1.06190186 * 10240; Err = 0.33896484 * 10240; time = 0.0416s; samplesPerSecond = 246378.9
12/20/2016 15:28:13:  Epoch[12 of 25]-Minibatch[1071-1080, 98.41%]: CE.SM = 1.06666260 * 10240; Err = 0.34414062 * 10240; time = 0.0413s; samplesPerSecond = 247773.9
12/20/2016 15:28:13:  Epoch[12 of 25]-Minibatch[1081-1090, 99.32%]: CE.SM = 1.09097900 * 10240; Err = 0.35869141 * 10240; time = 0.0406s; samplesPerSecond = 251949.9
12/20/2016 15:28:13: Finished Epoch[12 of 25]: [Training] CE.SM = 1.06581091 * 1124823; Err = 0.34550947 * 1124823; totalSamplesSeen = 13497876; learningRatePerSample = 0.003125; epochTime=4.69612s
12/20/2016 15:28:13: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn.12'

12/20/2016 15:28:13: Starting Epoch 13: learning rate per sample = 0.003125  effective momentum = 0.900000  momentum as time constant = 9719.0 samples

12/20/2016 15:28:13: Starting minibatch loop.
12/20/2016 15:28:13:  Epoch[13 of 25]-Minibatch[   1-  10, 0.91%]: CE.SM = 1.04875135 * 10240; Err = 0.33945313 * 10240; time = 0.0446s; samplesPerSecond = 229462.6
12/20/2016 15:28:13:  Epoch[13 of 25]-Minibatch[  11-  20, 1.82%]: CE.SM = 1.03939734 * 10240; Err = 0.33466797 * 10240; time = 0.0413s; samplesPerSecond = 248176.2
12/20/2016 15:28:14:  Epoch[13 of 25]-Minibatch[  21-  30, 2.73%]: CE.SM = 1.02011547 * 10240; Err = 0.33759766 * 10240; time = 0.0404s; samplesPerSecond = 253628.6
12/20/2016 15:28:14:  Epoch[13 of 25]-Minibatch[  31-  40, 3.64%]: CE.SM = 1.04209824 * 10240; Err = 0.33808594 * 10240; time = 0.0410s; samplesPerSecond = 249963.4
12/20/2016 15:28:14:  Epoch[13 of 25]-Minibatch[  41-  50, 4.56%]: CE.SM = 1.02967987 * 10240; Err = 0.33564453 * 10240; time = 0.0403s; samplesPerSecond = 254245.7
12/20/2016 15:28:14:  Epoch[13 of 25]-Minibatch[  51-  60, 5.47%]: CE.SM = 1.06318703 * 10240; Err = 0.34326172 * 10240; time = 0.0404s; samplesPerSecond = 253189.6
12/20/2016 15:28:14:  Epoch[13 of 25]-Minibatch[  61-  70, 6.38%]: CE.SM = 1.03744164 * 10240; Err = 0.33505859 * 10240; time = 0.0407s; samplesPerSecond = 251590.9
12/20/2016 15:28:14:  Epoch[13 of 25]-Minibatch[  71-  80, 7.29%]: CE.SM = 1.03744888 * 10240; Err = 0.33427734 * 10240; time = 0.0409s; samplesPerSecond = 250146.6
12/20/2016 15:28:14:  Epoch[13 of 25]-Minibatch[  81-  90, 8.20%]: CE.SM = 1.03545532 * 10240; Err = 0.33632812 * 10240; time = 0.0402s; samplesPerSecond = 254758.1
12/20/2016 15:28:14:  Epoch[13 of 25]-Minibatch[  91- 100, 9.11%]: CE.SM = 1.04748306 * 10240; Err = 0.34199219 * 10240; time = 0.0394s; samplesPerSecond = 259779.8
12/20/2016 15:28:14:  Epoch[13 of 25]-Minibatch[ 101- 110, 10.02%]: CE.SM = 1.02921143 * 10240; Err = 0.33330078 * 10240; time = 0.0405s; samplesPerSecond = 252964.4
12/20/2016 15:28:14:  Epoch[13 of 25]-Minibatch[ 111- 120, 10.93%]: CE.SM = 1.03821335 * 10240; Err = 0.34052734 * 10240; time = 0.0410s; samplesPerSecond = 249896.3
12/20/2016 15:28:14:  Epoch[13 of 25]-Minibatch[ 121- 130, 11.85%]: CE.SM = 1.03185730 * 10240; Err = 0.33457031 * 10240; time = 0.0409s; samplesPerSecond = 250587.3
12/20/2016 15:28:14:  Epoch[13 of 25]-Minibatch[ 131- 140, 12.76%]: CE.SM = 1.04805298 * 10240; Err = 0.33945313 * 10240; time = 0.0405s; samplesPerSecond = 252976.9
12/20/2016 15:28:14:  Epoch[13 of 25]-Minibatch[ 141- 150, 13.67%]: CE.SM = 1.02227783 * 10240; Err = 0.33125000 * 10240; time = 0.0400s; samplesPerSecond = 256179.3
12/20/2016 15:28:14:  Epoch[13 of 25]-Minibatch[ 151- 160, 14.58%]: CE.SM = 1.05559387 * 10240; Err = 0.34277344 * 10240; time = 0.0400s; samplesPerSecond = 255852.9
12/20/2016 15:28:14:  Epoch[13 of 25]-Minibatch[ 161- 170, 15.49%]: CE.SM = 1.06054688 * 10240; Err = 0.34677734 * 10240; time = 0.0410s; samplesPerSecond = 249945.1
12/20/2016 15:28:14:  Epoch[13 of 25]-Minibatch[ 171- 180, 16.40%]: CE.SM = 1.05206757 * 10240; Err = 0.34306641 * 10240; time = 0.0462s; samplesPerSecond = 221726.6
12/20/2016 15:28:14:  Epoch[13 of 25]-Minibatch[ 181- 190, 17.31%]: CE.SM = 1.04426117 * 10240; Err = 0.34238281 * 10240; time = 0.0414s; samplesPerSecond = 247211.6
12/20/2016 15:28:14:  Epoch[13 of 25]-Minibatch[ 191- 200, 18.22%]: CE.SM = 1.04319305 * 10240; Err = 0.33867188 * 10240; time = 0.0401s; samplesPerSecond = 255291.6
12/20/2016 15:28:14:  Epoch[13 of 25]-Minibatch[ 201- 210, 19.13%]: CE.SM = 1.05350952 * 10240; Err = 0.34228516 * 10240; time = 0.0395s; samplesPerSecond = 258971.7
12/20/2016 15:28:14:  Epoch[13 of 25]-Minibatch[ 211- 220, 20.05%]: CE.SM = 1.04410553 * 10240; Err = 0.34277344 * 10240; time = 0.0404s; samplesPerSecond = 253321.1
12/20/2016 15:28:14:  Epoch[13 of 25]-Minibatch[ 221- 230, 20.96%]: CE.SM = 1.05463409 * 10240; Err = 0.34101562 * 10240; time = 0.0399s; samplesPerSecond = 256500.2
12/20/2016 15:28:14:  Epoch[13 of 25]-Minibatch[ 231- 240, 21.87%]: CE.SM = 1.05737762 * 10240; Err = 0.35039063 * 10240; time = 0.0398s; samplesPerSecond = 257564.7
12/20/2016 15:28:14:  Epoch[13 of 25]-Minibatch[ 241- 250, 22.78%]: CE.SM = 1.04286652 * 10240; Err = 0.34150391 * 10240; time = 0.0405s; samplesPerSecond = 252720.9
12/20/2016 15:28:14:  Epoch[13 of 25]-Minibatch[ 251- 260, 23.69%]: CE.SM = 1.02565613 * 10240; Err = 0.33740234 * 10240; time = 0.0402s; samplesPerSecond = 254447.9
12/20/2016 15:28:14:  Epoch[13 of 25]-Minibatch[ 261- 270, 24.60%]: CE.SM = 1.04403076 * 10240; Err = 0.33505859 * 10240; time = 0.0411s; samplesPerSecond = 249160.5
12/20/2016 15:28:15:  Epoch[13 of 25]-Minibatch[ 271- 280, 25.51%]: CE.SM = 1.03589172 * 10240; Err = 0.34208984 * 10240; time = 0.0401s; samplesPerSecond = 255520.9
12/20/2016 15:28:15:  Epoch[13 of 25]-Minibatch[ 281- 290, 26.42%]: CE.SM = 1.04151306 * 10240; Err = 0.33769531 * 10240; time = 0.0398s; samplesPerSecond = 257383.4
12/20/2016 15:28:15:  Epoch[13 of 25]-Minibatch[ 291- 300, 27.33%]: CE.SM = 1.05623779 * 10240; Err = 0.34785156 * 10240; time = 0.0408s; samplesPerSecond = 250802.1
12/20/2016 15:28:15:  Epoch[13 of 25]-Minibatch[ 301- 310, 28.25%]: CE.SM = 1.02295227 * 10240; Err = 0.33212891 * 10240; time = 0.0401s; samplesPerSecond = 255151.6
12/20/2016 15:28:15:  Epoch[13 of 25]-Minibatch[ 311- 320, 29.16%]: CE.SM = 1.02006531 * 10240; Err = 0.33369141 * 10240; time = 0.0403s; samplesPerSecond = 254283.6
12/20/2016 15:28:15:  Epoch[13 of 25]-Minibatch[ 321- 330, 30.07%]: CE.SM = 1.03404846 * 10240; Err = 0.33593750 * 10240; time = 0.0411s; samplesPerSecond = 248990.9
12/20/2016 15:28:15:  Epoch[13 of 25]-Minibatch[ 331- 340, 30.98%]: CE.SM = 1.04704895 * 10240; Err = 0.33681641 * 10240; time = 0.0404s; samplesPerSecond = 253471.6
12/20/2016 15:28:15:  Epoch[13 of 25]-Minibatch[ 341- 350, 31.89%]: CE.SM = 1.05622864 * 10240; Err = 0.33457031 * 10240; time = 0.0399s; samplesPerSecond = 256346.1
12/20/2016 15:28:15:  Epoch[13 of 25]-Minibatch[ 351- 360, 32.80%]: CE.SM = 1.04801331 * 10240; Err = 0.34052734 * 10240; time = 0.0405s; samplesPerSecond = 252758.4
12/20/2016 15:28:15:  Epoch[13 of 25]-Minibatch[ 361- 370, 33.71%]: CE.SM = 1.03168640 * 10240; Err = 0.32841797 * 10240; time = 0.0404s; samplesPerSecond = 253616.0
12/20/2016 15:28:15:  Epoch[13 of 25]-Minibatch[ 371- 380, 34.62%]: CE.SM = 1.04610901 * 10240; Err = 0.33437500 * 10240; time = 0.0409s; samplesPerSecond = 250073.3
12/20/2016 15:28:15:  Epoch[13 of 25]-Minibatch[ 381- 390, 35.54%]: CE.SM = 1.02187195 * 10240; Err = 0.33398438 * 10240; time = 0.0404s; samplesPerSecond = 253754.3
12/20/2016 15:28:15:  Epoch[13 of 25]-Minibatch[ 391- 400, 36.45%]: CE.SM = 1.05013123 * 10240; Err = 0.34267578 * 10240; time = 0.0398s; samplesPerSecond = 257099.1
12/20/2016 15:28:15:  Epoch[13 of 25]-Minibatch[ 401- 410, 37.36%]: CE.SM = 1.03924255 * 10240; Err = 0.33818359 * 10240; time = 0.0403s; samplesPerSecond = 253905.3
12/20/2016 15:28:15:  Epoch[13 of 25]-Minibatch[ 411- 420, 38.27%]: CE.SM = 1.05722656 * 10240; Err = 0.34433594 * 10240; time = 0.0464s; samplesPerSecond = 220656.4
12/20/2016 15:28:15:  Epoch[13 of 25]-Minibatch[ 421- 430, 39.18%]: CE.SM = 1.04529114 * 10240; Err = 0.33857422 * 10240; time = 0.0426s; samplesPerSecond = 240511.1
12/20/2016 15:28:15:  Epoch[13 of 25]-Minibatch[ 431- 440, 40.09%]: CE.SM = 1.05514221 * 10240; Err = 0.34042969 * 10240; time = 0.0421s; samplesPerSecond = 243374.9
12/20/2016 15:28:15:  Epoch[13 of 25]-Minibatch[ 441- 450, 41.00%]: CE.SM = 1.03959045 * 10240; Err = 0.33613281 * 10240; time = 0.0409s; samplesPerSecond = 250403.5
12/20/2016 15:28:15:  Epoch[13 of 25]-Minibatch[ 451- 460, 41.91%]: CE.SM = 1.04461060 * 10240; Err = 0.34052734 * 10240; time = 0.0403s; samplesPerSecond = 254277.3
12/20/2016 15:28:15:  Epoch[13 of 25]-Minibatch[ 461- 470, 42.82%]: CE.SM = 1.03631897 * 10240; Err = 0.34228516 * 10240; time = 0.0413s; samplesPerSecond = 247797.9
12/20/2016 15:28:15:  Epoch[13 of 25]-Minibatch[ 471- 480, 43.74%]: CE.SM = 1.02595825 * 10240; Err = 0.33896484 * 10240; time = 0.0403s; samplesPerSecond = 253949.4
12/20/2016 15:28:15:  Epoch[13 of 25]-Minibatch[ 481- 490, 44.65%]: CE.SM = 1.05377502 * 10240; Err = 0.34365234 * 10240; time = 0.0396s; samplesPerSecond = 258703.5
12/20/2016 15:28:15:  Epoch[13 of 25]-Minibatch[ 491- 500, 45.56%]: CE.SM = 1.05429382 * 10240; Err = 0.34609375 * 10240; time = 0.0405s; samplesPerSecond = 252534.0
12/20/2016 15:28:15:  Epoch[13 of 25]-Minibatch[ 501- 510, 46.47%]: CE.SM = 1.01763916 * 10240; Err = 0.33037109 * 10240; time = 0.0400s; samplesPerSecond = 255731.5
12/20/2016 15:28:16:  Epoch[13 of 25]-Minibatch[ 511- 520, 47.38%]: CE.SM = 1.04519043 * 10240; Err = 0.33847656 * 10240; time = 0.0399s; samplesPerSecond = 256751.0
12/20/2016 15:28:16:  Epoch[13 of 25]-Minibatch[ 521- 530, 48.29%]: CE.SM = 1.04318237 * 10240; Err = 0.33955078 * 10240; time = 0.0411s; samplesPerSecond = 249069.6
12/20/2016 15:28:16:  Epoch[13 of 25]-Minibatch[ 531- 540, 49.20%]: CE.SM = 1.03227539 * 10240; Err = 0.33730469 * 10240; time = 0.0397s; samplesPerSecond = 258162.1
12/20/2016 15:28:16:  Epoch[13 of 25]-Minibatch[ 541- 550, 50.11%]: CE.SM = 1.04078369 * 10240; Err = 0.34091797 * 10240; time = 0.0412s; samplesPerSecond = 248338.7
12/20/2016 15:28:16:  Epoch[13 of 25]-Minibatch[ 551- 560, 51.03%]: CE.SM = 1.04467773 * 10240; Err = 0.33574219 * 10240; time = 0.0408s; samplesPerSecond = 251214.4
12/20/2016 15:28:16:  Epoch[13 of 25]-Minibatch[ 561- 570, 51.94%]: CE.SM = 1.05328979 * 10240; Err = 0.33847656 * 10240; time = 0.0406s; samplesPerSecond = 251931.3
12/20/2016 15:28:16:  Epoch[13 of 25]-Minibatch[ 571- 580, 52.85%]: CE.SM = 1.06043091 * 10240; Err = 0.34287109 * 10240; time = 0.0408s; samplesPerSecond = 250894.3
12/20/2016 15:28:16:  Epoch[13 of 25]-Minibatch[ 581- 590, 53.76%]: CE.SM = 1.05490112 * 10240; Err = 0.34082031 * 10240; time = 0.0398s; samplesPerSecond = 257435.2
12/20/2016 15:28:16:  Epoch[13 of 25]-Minibatch[ 591- 600, 54.67%]: CE.SM = 1.03976440 * 10240; Err = 0.34130859 * 10240; time = 0.0393s; samplesPerSecond = 260248.6
12/20/2016 15:28:16:  Epoch[13 of 25]-Minibatch[ 601- 610, 55.58%]: CE.SM = 1.03526001 * 10240; Err = 0.34130859 * 10240; time = 0.0396s; samplesPerSecond = 258631.6
12/20/2016 15:28:16:  Epoch[13 of 25]-Minibatch[ 611- 620, 56.49%]: CE.SM = 1.03577881 * 10240; Err = 0.34003906 * 10240; time = 0.0395s; samplesPerSecond = 259339.0
12/20/2016 15:28:16:  Epoch[13 of 25]-Minibatch[ 621- 630, 57.40%]: CE.SM = 1.06530151 * 10240; Err = 0.34609375 * 10240; time = 0.0393s; samplesPerSecond = 260294.9
12/20/2016 15:28:16:  Epoch[13 of 25]-Minibatch[ 631- 640, 58.31%]: CE.SM = 1.02883911 * 10240; Err = 0.34013672 * 10240; time = 0.0393s; samplesPerSecond = 260248.6
12/20/2016 15:28:16:  Epoch[13 of 25]-Minibatch[ 641- 650, 59.23%]: CE.SM = 1.04440308 * 10240; Err = 0.33603516 * 10240; time = 0.0396s; samplesPerSecond = 258481.4
12/20/2016 15:28:16:  Epoch[13 of 25]-Minibatch[ 651- 660, 60.14%]: CE.SM = 1.04428101 * 10240; Err = 0.33935547 * 10240; time = 0.0397s; samplesPerSecond = 257616.5
12/20/2016 15:28:16:  Epoch[13 of 25]-Minibatch[ 661- 670, 61.05%]: CE.SM = 1.04537354 * 10240; Err = 0.34199219 * 10240; time = 0.0393s; samplesPerSecond = 260745.6
12/20/2016 15:28:16:  Epoch[13 of 25]-Minibatch[ 671- 680, 61.96%]: CE.SM = 1.04030762 * 10240; Err = 0.33896484 * 10240; time = 0.0395s; samplesPerSecond = 259142.1
12/20/2016 15:28:16:  Epoch[13 of 25]-Minibatch[ 681- 690, 62.87%]: CE.SM = 1.05941772 * 10240; Err = 0.35175781 * 10240; time = 0.0394s; samplesPerSecond = 260103.1
12/20/2016 15:28:16:  Epoch[13 of 25]-Minibatch[ 691- 700, 63.78%]: CE.SM = 1.06369629 * 10240; Err = 0.34306641 * 10240; time = 0.0393s; samplesPerSecond = 260374.3
12/20/2016 15:28:16:  Epoch[13 of 25]-Minibatch[ 701- 710, 64.69%]: CE.SM = 1.05856934 * 10240; Err = 0.34472656 * 10240; time = 0.0395s; samplesPerSecond = 259004.5
12/20/2016 15:28:16:  Epoch[13 of 25]-Minibatch[ 711- 720, 65.60%]: CE.SM = 1.05756836 * 10240; Err = 0.34511719 * 10240; time = 0.0393s; samplesPerSecond = 260314.7
12/20/2016 15:28:16:  Epoch[13 of 25]-Minibatch[ 721- 730, 66.51%]: CE.SM = 1.03159790 * 10240; Err = 0.33222656 * 10240; time = 0.0396s; samplesPerSecond = 258631.6
12/20/2016 15:28:16:  Epoch[13 of 25]-Minibatch[ 731- 740, 67.43%]: CE.SM = 1.03023682 * 10240; Err = 0.33857422 * 10240; time = 0.0394s; samplesPerSecond = 260149.4
12/20/2016 15:28:16:  Epoch[13 of 25]-Minibatch[ 741- 750, 68.34%]: CE.SM = 1.04367065 * 10240; Err = 0.33886719 * 10240; time = 0.0396s; samplesPerSecond = 258775.4
12/20/2016 15:28:16:  Epoch[13 of 25]-Minibatch[ 751- 760, 69.25%]: CE.SM = 1.04042969 * 10240; Err = 0.34238281 * 10240; time = 0.0395s; samplesPerSecond = 258997.9
12/20/2016 15:28:17:  Epoch[13 of 25]-Minibatch[ 761- 770, 70.16%]: CE.SM = 1.04213867 * 10240; Err = 0.33847656 * 10240; time = 0.0397s; samplesPerSecond = 257876.0
12/20/2016 15:28:17:  Epoch[13 of 25]-Minibatch[ 771- 780, 71.07%]: CE.SM = 1.02920532 * 10240; Err = 0.33466797 * 10240; time = 0.0403s; samplesPerSecond = 253911.6
12/20/2016 15:28:17:  Epoch[13 of 25]-Minibatch[ 781- 790, 71.98%]: CE.SM = 1.03015137 * 10240; Err = 0.33701172 * 10240; time = 0.0400s; samplesPerSecond = 255846.5
12/20/2016 15:28:17:  Epoch[13 of 25]-Minibatch[ 791- 800, 72.89%]: CE.SM = 1.03036499 * 10240; Err = 0.33740234 * 10240; time = 0.0430s; samplesPerSecond = 238106.3
12/20/2016 15:28:17:  Epoch[13 of 25]-Minibatch[ 801- 810, 73.80%]: CE.SM = 1.03665161 * 10240; Err = 0.33984375 * 10240; time = 0.0399s; samplesPerSecond = 256725.2
12/20/2016 15:28:17:  Epoch[13 of 25]-Minibatch[ 811- 820, 74.72%]: CE.SM = 1.03489990 * 10240; Err = 0.33867188 * 10240; time = 0.0400s; samplesPerSecond = 255776.2
12/20/2016 15:28:17:  Epoch[13 of 25]-Minibatch[ 821- 830, 75.63%]: CE.SM = 1.06307983 * 10240; Err = 0.34296875 * 10240; time = 0.0399s; samplesPerSecond = 256487.3
12/20/2016 15:28:17:  Epoch[13 of 25]-Minibatch[ 831- 840, 76.54%]: CE.SM = 1.05029907 * 10240; Err = 0.33916016 * 10240; time = 0.0418s; samplesPerSecond = 245128.5
12/20/2016 15:28:17:  Epoch[13 of 25]-Minibatch[ 841- 850, 77.45%]: CE.SM = 1.04860229 * 10240; Err = 0.34335938 * 10240; time = 0.0399s; samplesPerSecond = 256873.4
12/20/2016 15:28:17:  Epoch[13 of 25]-Minibatch[ 851- 860, 78.36%]: CE.SM = 1.03699951 * 10240; Err = 0.33652344 * 10240; time = 0.0400s; samplesPerSecond = 255725.1
12/20/2016 15:28:17:  Epoch[13 of 25]-Minibatch[ 861- 870, 79.27%]: CE.SM = 1.05615845 * 10240; Err = 0.34404297 * 10240; time = 0.0400s; samplesPerSecond = 255948.8
12/20/2016 15:28:17:  Epoch[13 of 25]-Minibatch[ 871- 880, 80.18%]: CE.SM = 1.03006592 * 10240; Err = 0.33154297 * 10240; time = 0.0399s; samplesPerSecond = 256918.5
12/20/2016 15:28:17:  Epoch[13 of 25]-Minibatch[ 881- 890, 81.09%]: CE.SM = 1.04755249 * 10240; Err = 0.34042969 * 10240; time = 0.0400s; samplesPerSecond = 256307.6
12/20/2016 15:28:17:  Epoch[13 of 25]-Minibatch[ 891- 900, 82.00%]: CE.SM = 1.05226440 * 10240; Err = 0.34101562 * 10240; time = 0.0408s; samplesPerSecond = 251214.4
12/20/2016 15:28:17:  Epoch[13 of 25]-Minibatch[ 901- 910, 82.92%]: CE.SM = 1.03900757 * 10240; Err = 0.33701172 * 10240; time = 0.0402s; samplesPerSecond = 254840.5
12/20/2016 15:28:17:  Epoch[13 of 25]-Minibatch[ 911- 920, 83.83%]: CE.SM = 1.02318115 * 10240; Err = 0.33339844 * 10240; time = 0.0405s; samplesPerSecond = 253095.7
12/20/2016 15:28:17:  Epoch[13 of 25]-Minibatch[ 921- 930, 84.74%]: CE.SM = 1.05533447 * 10240; Err = 0.33945313 * 10240; time = 0.0428s; samplesPerSecond = 239006.6
12/20/2016 15:28:17:  Epoch[13 of 25]-Minibatch[ 931- 940, 85.65%]: CE.SM = 1.06942749 * 10240; Err = 0.34746094 * 10240; time = 0.0405s; samplesPerSecond = 252777.1
12/20/2016 15:28:17:  Epoch[13 of 25]-Minibatch[ 941- 950, 86.56%]: CE.SM = 1.04302979 * 10240; Err = 0.33681641 * 10240; time = 0.0407s; samplesPerSecond = 251492.0
12/20/2016 15:28:17:  Epoch[13 of 25]-Minibatch[ 951- 960, 87.47%]: CE.SM = 1.04989014 * 10240; Err = 0.34326172 * 10240; time = 0.0411s; samplesPerSecond = 249300.1
12/20/2016 15:28:17:  Epoch[13 of 25]-Minibatch[ 961- 970, 88.38%]: CE.SM = 1.06031494 * 10240; Err = 0.34736328 * 10240; time = 0.0402s; samplesPerSecond = 254663.0
12/20/2016 15:28:17:  Epoch[13 of 25]-Minibatch[ 971- 980, 89.29%]: CE.SM = 1.04013672 * 10240; Err = 0.33242187 * 10240; time = 0.0400s; samplesPerSecond = 255712.3
12/20/2016 15:28:17:  Epoch[13 of 25]-Minibatch[ 981- 990, 90.21%]: CE.SM = 1.07616577 * 10240; Err = 0.34687500 * 10240; time = 0.0407s; samplesPerSecond = 251288.3
12/20/2016 15:28:17:  Epoch[13 of 25]-Minibatch[ 991-1000, 91.12%]: CE.SM = 1.04653320 * 10240; Err = 0.33662109 * 10240; time = 0.0407s; samplesPerSecond = 251757.9
12/20/2016 15:28:17:  Epoch[13 of 25]-Minibatch[1001-1010, 92.03%]: CE.SM = 1.04088135 * 10240; Err = 0.33818359 * 10240; time = 0.0401s; samplesPerSecond = 255463.5
12/20/2016 15:28:18:  Epoch[13 of 25]-Minibatch[1011-1020, 92.94%]: CE.SM = 1.02924805 * 10240; Err = 0.33847656 * 10240; time = 0.0396s; samplesPerSecond = 258651.2
12/20/2016 15:28:18:  Epoch[13 of 25]-Minibatch[1021-1030, 93.85%]: CE.SM = 1.06711426 * 10240; Err = 0.35136719 * 10240; time = 0.0404s; samplesPerSecond = 253434.0
12/20/2016 15:28:18:  Epoch[13 of 25]-Minibatch[1031-1040, 94.76%]: CE.SM = 1.05181885 * 10240; Err = 0.34228516 * 10240; time = 0.0404s; samplesPerSecond = 253452.8
12/20/2016 15:28:18:  Epoch[13 of 25]-Minibatch[1041-1050, 95.67%]: CE.SM = 1.05152588 * 10240; Err = 0.34316406 * 10240; time = 0.0410s; samplesPerSecond = 249737.8
12/20/2016 15:28:18:  Epoch[13 of 25]-Minibatch[1051-1060, 96.58%]: CE.SM = 1.06781006 * 10240; Err = 0.34228516 * 10240; time = 0.0408s; samplesPerSecond = 250980.4
12/20/2016 15:28:18:  Epoch[13 of 25]-Minibatch[1061-1070, 97.49%]: CE.SM = 1.06998291 * 10240; Err = 0.33935547 * 10240; time = 0.0408s; samplesPerSecond = 250968.1
12/20/2016 15:28:18:  Epoch[13 of 25]-Minibatch[1071-1080, 98.41%]: CE.SM = 1.06507568 * 10240; Err = 0.34501953 * 10240; time = 0.0411s; samplesPerSecond = 248954.6
12/20/2016 15:28:18:  Epoch[13 of 25]-Minibatch[1081-1090, 99.32%]: CE.SM = 1.02835693 * 10240; Err = 0.33066406 * 10240; time = 0.0408s; samplesPerSecond = 251122.0
12/20/2016 15:28:18: Finished Epoch[13 of 25]: [Training] CE.SM = 1.04446833 * 1124823; Err = 0.33961788 * 1124823; totalSamplesSeen = 14622699; learningRatePerSample = 0.003125; epochTime=4.64036s
12/20/2016 15:28:18: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn.13'

12/20/2016 15:28:18: Starting Epoch 14: learning rate per sample = 0.003125  effective momentum = 0.900000  momentum as time constant = 9719.0 samples

12/20/2016 15:28:18: Starting minibatch loop.
12/20/2016 15:28:18:  Epoch[14 of 25]-Minibatch[   1-  10, 0.91%]: CE.SM = 1.00852814 * 10240; Err = 0.33251953 * 10240; time = 0.0454s; samplesPerSecond = 225491.1
12/20/2016 15:28:18:  Epoch[14 of 25]-Minibatch[  11-  20, 1.82%]: CE.SM = 1.01850967 * 10240; Err = 0.33300781 * 10240; time = 0.0413s; samplesPerSecond = 247995.9
12/20/2016 15:28:18:  Epoch[14 of 25]-Minibatch[  21-  30, 2.73%]: CE.SM = 1.00919514 * 10240; Err = 0.32724609 * 10240; time = 0.0412s; samplesPerSecond = 248525.6
12/20/2016 15:28:18:  Epoch[14 of 25]-Minibatch[  31-  40, 3.64%]: CE.SM = 0.99943409 * 10240; Err = 0.32529297 * 10240; time = 0.0412s; samplesPerSecond = 248694.6
12/20/2016 15:28:18:  Epoch[14 of 25]-Minibatch[  41-  50, 4.56%]: CE.SM = 1.00776291 * 10240; Err = 0.32851562 * 10240; time = 0.0403s; samplesPerSecond = 254006.1
12/20/2016 15:28:18:  Epoch[14 of 25]-Minibatch[  51-  60, 5.47%]: CE.SM = 1.01162033 * 10240; Err = 0.32890625 * 10240; time = 0.0405s; samplesPerSecond = 252608.7
12/20/2016 15:28:18:  Epoch[14 of 25]-Minibatch[  61-  70, 6.38%]: CE.SM = 1.00002632 * 10240; Err = 0.32246094 * 10240; time = 0.0406s; samplesPerSecond = 252390.8
12/20/2016 15:28:18:  Epoch[14 of 25]-Minibatch[  71-  80, 7.29%]: CE.SM = 1.01017456 * 10240; Err = 0.32792969 * 10240; time = 0.0400s; samplesPerSecond = 256115.3
12/20/2016 15:28:18:  Epoch[14 of 25]-Minibatch[  81-  90, 8.20%]: CE.SM = 1.00853119 * 10240; Err = 0.32783203 * 10240; time = 0.0402s; samplesPerSecond = 254492.1
12/20/2016 15:28:19:  Epoch[14 of 25]-Minibatch[  91- 100, 9.11%]: CE.SM = 0.99767151 * 10240; Err = 0.32607422 * 10240; time = 0.0402s; samplesPerSecond = 254656.7
12/20/2016 15:28:19:  Epoch[14 of 25]-Minibatch[ 101- 110, 10.02%]: CE.SM = 1.03881760 * 10240; Err = 0.33964844 * 10240; time = 0.0412s; samplesPerSecond = 248356.8
12/20/2016 15:28:19:  Epoch[14 of 25]-Minibatch[ 111- 120, 10.93%]: CE.SM = 1.02500534 * 10240; Err = 0.33056641 * 10240; time = 0.0445s; samplesPerSecond = 230303.8
12/20/2016 15:28:19:  Epoch[14 of 25]-Minibatch[ 121- 130, 11.85%]: CE.SM = 1.00317993 * 10240; Err = 0.33232422 * 10240; time = 0.0412s; samplesPerSecond = 248579.9
12/20/2016 15:28:19:  Epoch[14 of 25]-Minibatch[ 131- 140, 12.76%]: CE.SM = 1.02154388 * 10240; Err = 0.33359375 * 10240; time = 0.0406s; samplesPerSecond = 252365.9
12/20/2016 15:28:19:  Epoch[14 of 25]-Minibatch[ 141- 150, 13.67%]: CE.SM = 1.01135406 * 10240; Err = 0.33046875 * 10240; time = 0.0407s; samplesPerSecond = 251671.3
12/20/2016 15:28:19:  Epoch[14 of 25]-Minibatch[ 151- 160, 14.58%]: CE.SM = 1.01840515 * 10240; Err = 0.33369141 * 10240; time = 0.0404s; samplesPerSecond = 253634.9
12/20/2016 15:28:19:  Epoch[14 of 25]-Minibatch[ 161- 170, 15.49%]: CE.SM = 1.03035736 * 10240; Err = 0.33593750 * 10240; time = 0.0403s; samplesPerSecond = 253829.8
12/20/2016 15:28:19:  Epoch[14 of 25]-Minibatch[ 171- 180, 16.40%]: CE.SM = 1.02494049 * 10240; Err = 0.33496094 * 10240; time = 0.0447s; samplesPerSecond = 229041.8
12/20/2016 15:28:19:  Epoch[14 of 25]-Minibatch[ 181- 190, 17.31%]: CE.SM = 1.03384705 * 10240; Err = 0.33798828 * 10240; time = 0.0420s; samplesPerSecond = 243826.9
12/20/2016 15:28:19:  Epoch[14 of 25]-Minibatch[ 191- 200, 18.22%]: CE.SM = 1.01306610 * 10240; Err = 0.32880859 * 10240; time = 0.0408s; samplesPerSecond = 250986.5
12/20/2016 15:28:19:  Epoch[14 of 25]-Minibatch[ 201- 210, 19.13%]: CE.SM = 1.00383148 * 10240; Err = 0.32841797 * 10240; time = 0.0407s; samplesPerSecond = 251726.9
12/20/2016 15:28:19:  Epoch[14 of 25]-Minibatch[ 211- 220, 20.05%]: CE.SM = 1.00163727 * 10240; Err = 0.33603516 * 10240; time = 0.0402s; samplesPerSecond = 254644.0
12/20/2016 15:28:19:  Epoch[14 of 25]-Minibatch[ 221- 230, 20.96%]: CE.SM = 1.00738068 * 10240; Err = 0.33095703 * 10240; time = 0.0403s; samplesPerSecond = 253785.7
12/20/2016 15:28:19:  Epoch[14 of 25]-Minibatch[ 231- 240, 21.87%]: CE.SM = 0.99992065 * 10240; Err = 0.32714844 * 10240; time = 0.0410s; samplesPerSecond = 249652.6
12/20/2016 15:28:19:  Epoch[14 of 25]-Minibatch[ 241- 250, 22.78%]: CE.SM = 1.03378143 * 10240; Err = 0.33994141 * 10240; time = 0.0408s; samplesPerSecond = 250808.3
12/20/2016 15:28:19:  Epoch[14 of 25]-Minibatch[ 251- 260, 23.69%]: CE.SM = 1.03786926 * 10240; Err = 0.34238281 * 10240; time = 0.0402s; samplesPerSecond = 254498.5
12/20/2016 15:28:19:  Epoch[14 of 25]-Minibatch[ 261- 270, 24.60%]: CE.SM = 1.03044128 * 10240; Err = 0.33085938 * 10240; time = 0.0413s; samplesPerSecond = 247672.0
12/20/2016 15:28:19:  Epoch[14 of 25]-Minibatch[ 271- 280, 25.51%]: CE.SM = 1.00390015 * 10240; Err = 0.32519531 * 10240; time = 0.0403s; samplesPerSecond = 254081.7
12/20/2016 15:28:19:  Epoch[14 of 25]-Minibatch[ 281- 290, 26.42%]: CE.SM = 1.02746582 * 10240; Err = 0.33193359 * 10240; time = 0.0403s; samplesPerSecond = 253880.1
12/20/2016 15:28:19:  Epoch[14 of 25]-Minibatch[ 291- 300, 27.33%]: CE.SM = 1.01415405 * 10240; Err = 0.32880859 * 10240; time = 0.0404s; samplesPerSecond = 253716.6
12/20/2016 15:28:19:  Epoch[14 of 25]-Minibatch[ 301- 310, 28.25%]: CE.SM = 1.03742981 * 10240; Err = 0.34121094 * 10240; time = 0.0399s; samplesPerSecond = 256365.3
12/20/2016 15:28:19:  Epoch[14 of 25]-Minibatch[ 311- 320, 29.16%]: CE.SM = 1.02086487 * 10240; Err = 0.34033203 * 10240; time = 0.0400s; samplesPerSecond = 255884.9
12/20/2016 15:28:19:  Epoch[14 of 25]-Minibatch[ 321- 330, 30.07%]: CE.SM = 1.01775208 * 10240; Err = 0.33027344 * 10240; time = 0.0409s; samplesPerSecond = 250532.1
12/20/2016 15:28:20:  Epoch[14 of 25]-Minibatch[ 331- 340, 30.98%]: CE.SM = 1.03001404 * 10240; Err = 0.33564453 * 10240; time = 0.0397s; samplesPerSecond = 257999.5
12/20/2016 15:28:20:  Epoch[14 of 25]-Minibatch[ 341- 350, 31.89%]: CE.SM = 1.01799622 * 10240; Err = 0.33144531 * 10240; time = 0.0406s; samplesPerSecond = 252024.3
12/20/2016 15:28:20:  Epoch[14 of 25]-Minibatch[ 351- 360, 32.80%]: CE.SM = 1.04888611 * 10240; Err = 0.33652344 * 10240; time = 0.0405s; samplesPerSecond = 252814.5
12/20/2016 15:28:20:  Epoch[14 of 25]-Minibatch[ 361- 370, 33.71%]: CE.SM = 1.00744324 * 10240; Err = 0.32587891 * 10240; time = 0.0407s; samplesPerSecond = 251424.1
12/20/2016 15:28:20:  Epoch[14 of 25]-Minibatch[ 371- 380, 34.62%]: CE.SM = 1.00884705 * 10240; Err = 0.32773438 * 10240; time = 0.0411s; samplesPerSecond = 248984.9
12/20/2016 15:28:20:  Epoch[14 of 25]-Minibatch[ 381- 390, 35.54%]: CE.SM = 1.00649719 * 10240; Err = 0.33056641 * 10240; time = 0.0403s; samplesPerSecond = 254334.1
12/20/2016 15:28:20:  Epoch[14 of 25]-Minibatch[ 391- 400, 36.45%]: CE.SM = 1.02186279 * 10240; Err = 0.33525391 * 10240; time = 0.0412s; samplesPerSecond = 248579.9
12/20/2016 15:28:20:  Epoch[14 of 25]-Minibatch[ 401- 410, 37.36%]: CE.SM = 0.99064026 * 10240; Err = 0.32744141 * 10240; time = 0.0404s; samplesPerSecond = 253760.6
12/20/2016 15:28:20:  Epoch[14 of 25]-Minibatch[ 411- 420, 38.27%]: CE.SM = 1.01809082 * 10240; Err = 0.32744141 * 10240; time = 0.0399s; samplesPerSecond = 256828.3
12/20/2016 15:28:20:  Epoch[14 of 25]-Minibatch[ 421- 430, 39.18%]: CE.SM = 1.02988281 * 10240; Err = 0.33310547 * 10240; time = 0.0410s; samplesPerSecond = 249725.6
12/20/2016 15:28:20:  Epoch[14 of 25]-Minibatch[ 431- 440, 40.09%]: CE.SM = 1.02728882 * 10240; Err = 0.33359375 * 10240; time = 0.0410s; samplesPerSecond = 249829.2
12/20/2016 15:28:20:  Epoch[14 of 25]-Minibatch[ 441- 450, 41.00%]: CE.SM = 1.01832275 * 10240; Err = 0.33447266 * 10240; time = 0.0399s; samplesPerSecond = 256339.7
12/20/2016 15:28:20:  Epoch[14 of 25]-Minibatch[ 451- 460, 41.91%]: CE.SM = 1.02327881 * 10240; Err = 0.34003906 * 10240; time = 0.0404s; samplesPerSecond = 253346.2
12/20/2016 15:28:20:  Epoch[14 of 25]-Minibatch[ 461- 470, 42.82%]: CE.SM = 1.00322571 * 10240; Err = 0.32988281 * 10240; time = 0.0404s; samplesPerSecond = 253402.6
12/20/2016 15:28:20:  Epoch[14 of 25]-Minibatch[ 471- 480, 43.74%]: CE.SM = 1.01701660 * 10240; Err = 0.33203125 * 10240; time = 0.0408s; samplesPerSecond = 251091.2
12/20/2016 15:28:20:  Epoch[14 of 25]-Minibatch[ 481- 490, 44.65%]: CE.SM = 1.03197937 * 10240; Err = 0.33583984 * 10240; time = 0.0410s; samplesPerSecond = 249537.0
12/20/2016 15:28:20:  Epoch[14 of 25]-Minibatch[ 491- 500, 45.56%]: CE.SM = 1.03689270 * 10240; Err = 0.33593750 * 10240; time = 0.0400s; samplesPerSecond = 255705.9
12/20/2016 15:28:20:  Epoch[14 of 25]-Minibatch[ 501- 510, 46.47%]: CE.SM = 1.02754517 * 10240; Err = 0.33271484 * 10240; time = 0.0404s; samplesPerSecond = 253572.0
12/20/2016 15:28:20:  Epoch[14 of 25]-Minibatch[ 511- 520, 47.38%]: CE.SM = 1.02729492 * 10240; Err = 0.33544922 * 10240; time = 0.0406s; samplesPerSecond = 251974.7
12/20/2016 15:28:20:  Epoch[14 of 25]-Minibatch[ 521- 530, 48.29%]: CE.SM = 1.01984253 * 10240; Err = 0.33398438 * 10240; time = 0.0407s; samplesPerSecond = 251881.7
12/20/2016 15:28:20:  Epoch[14 of 25]-Minibatch[ 531- 540, 49.20%]: CE.SM = 1.04127808 * 10240; Err = 0.33398438 * 10240; time = 0.0437s; samplesPerSecond = 234164.2
12/20/2016 15:28:20:  Epoch[14 of 25]-Minibatch[ 541- 550, 50.11%]: CE.SM = 1.03184204 * 10240; Err = 0.33750000 * 10240; time = 0.0449s; samplesPerSecond = 228291.2
12/20/2016 15:28:20:  Epoch[14 of 25]-Minibatch[ 551- 560, 51.03%]: CE.SM = 1.04546509 * 10240; Err = 0.33310547 * 10240; time = 0.0394s; samplesPerSecond = 259839.1
12/20/2016 15:28:20:  Epoch[14 of 25]-Minibatch[ 561- 570, 51.94%]: CE.SM = 1.04448853 * 10240; Err = 0.34218750 * 10240; time = 0.0447s; samplesPerSecond = 229113.5
12/20/2016 15:28:20:  Epoch[14 of 25]-Minibatch[ 571- 580, 52.85%]: CE.SM = 1.03841553 * 10240; Err = 0.33710937 * 10240; time = 0.0437s; samplesPerSecond = 234153.5
12/20/2016 15:28:21:  Epoch[14 of 25]-Minibatch[ 581- 590, 53.76%]: CE.SM = 1.02224731 * 10240; Err = 0.33369141 * 10240; time = 0.0435s; samplesPerSecond = 235575.6
12/20/2016 15:28:21:  Epoch[14 of 25]-Minibatch[ 591- 600, 54.67%]: CE.SM = 1.03167114 * 10240; Err = 0.33886719 * 10240; time = 0.0512s; samplesPerSecond = 199929.7
12/20/2016 15:28:21:  Epoch[14 of 25]-Minibatch[ 601- 610, 55.58%]: CE.SM = 1.03799438 * 10240; Err = 0.33476563 * 10240; time = 0.0413s; samplesPerSecond = 247779.9
12/20/2016 15:28:21:  Epoch[14 of 25]-Minibatch[ 611- 620, 56.49%]: CE.SM = 1.04260254 * 10240; Err = 0.34248047 * 10240; time = 0.0404s; samplesPerSecond = 253490.4
12/20/2016 15:28:21:  Epoch[14 of 25]-Minibatch[ 621- 630, 57.40%]: CE.SM = 1.02069702 * 10240; Err = 0.33349609 * 10240; time = 0.0405s; samplesPerSecond = 252802.1
12/20/2016 15:28:21:  Epoch[14 of 25]-Minibatch[ 631- 640, 58.31%]: CE.SM = 1.01481934 * 10240; Err = 0.33271484 * 10240; time = 0.0460s; samplesPerSecond = 222783.0
12/20/2016 15:28:21:  Epoch[14 of 25]-Minibatch[ 641- 650, 59.23%]: CE.SM = 1.02542114 * 10240; Err = 0.33417969 * 10240; time = 0.0418s; samplesPerSecond = 244730.2
12/20/2016 15:28:21:  Epoch[14 of 25]-Minibatch[ 651- 660, 60.14%]: CE.SM = 1.03566284 * 10240; Err = 0.33691406 * 10240; time = 0.0420s; samplesPerSecond = 243908.2
12/20/2016 15:28:21:  Epoch[14 of 25]-Minibatch[ 661- 670, 61.05%]: CE.SM = 1.02672119 * 10240; Err = 0.33691406 * 10240; time = 0.0400s; samplesPerSecond = 255980.8
12/20/2016 15:28:21:  Epoch[14 of 25]-Minibatch[ 671- 680, 61.96%]: CE.SM = 1.01707764 * 10240; Err = 0.32998047 * 10240; time = 0.0406s; samplesPerSecond = 252073.9
12/20/2016 15:28:21:  Epoch[14 of 25]-Minibatch[ 681- 690, 62.87%]: CE.SM = 1.01884155 * 10240; Err = 0.32910156 * 10240; time = 0.0402s; samplesPerSecond = 254606.0
12/20/2016 15:28:21:  Epoch[14 of 25]-Minibatch[ 691- 700, 63.78%]: CE.SM = 1.04170532 * 10240; Err = 0.33916016 * 10240; time = 0.0407s; samplesPerSecond = 251683.6
12/20/2016 15:28:21:  Epoch[14 of 25]-Minibatch[ 701- 710, 64.69%]: CE.SM = 1.01190186 * 10240; Err = 0.33271484 * 10240; time = 0.0404s; samplesPerSecond = 253333.7
12/20/2016 15:28:21:  Epoch[14 of 25]-Minibatch[ 711- 720, 65.60%]: CE.SM = 1.03419800 * 10240; Err = 0.34003906 * 10240; time = 0.0401s; samplesPerSecond = 255635.7
12/20/2016 15:28:21:  Epoch[14 of 25]-Minibatch[ 721- 730, 66.51%]: CE.SM = 1.02478638 * 10240; Err = 0.32998047 * 10240; time = 0.0409s; samplesPerSecond = 250618.0
12/20/2016 15:28:21:  Epoch[14 of 25]-Minibatch[ 731- 740, 67.43%]: CE.SM = 0.99585571 * 10240; Err = 0.32675781 * 10240; time = 0.0405s; samplesPerSecond = 253120.8
12/20/2016 15:28:21:  Epoch[14 of 25]-Minibatch[ 741- 750, 68.34%]: CE.SM = 1.02780762 * 10240; Err = 0.32861328 * 10240; time = 0.0412s; samplesPerSecond = 248730.8
12/20/2016 15:28:21:  Epoch[14 of 25]-Minibatch[ 751- 760, 69.25%]: CE.SM = 1.03981934 * 10240; Err = 0.34082031 * 10240; time = 0.0402s; samplesPerSecond = 255030.9
12/20/2016 15:28:21:  Epoch[14 of 25]-Minibatch[ 761- 770, 70.16%]: CE.SM = 1.03504028 * 10240; Err = 0.34052734 * 10240; time = 0.0406s; samplesPerSecond = 251937.5
12/20/2016 15:28:21:  Epoch[14 of 25]-Minibatch[ 771- 780, 71.07%]: CE.SM = 1.03124390 * 10240; Err = 0.33554688 * 10240; time = 0.0414s; samplesPerSecond = 247384.8
12/20/2016 15:28:21:  Epoch[14 of 25]-Minibatch[ 781- 790, 71.98%]: CE.SM = 1.03909302 * 10240; Err = 0.33818359 * 10240; time = 0.0407s; samplesPerSecond = 251547.6
12/20/2016 15:28:21:  Epoch[14 of 25]-Minibatch[ 791- 800, 72.89%]: CE.SM = 1.03090820 * 10240; Err = 0.33681641 * 10240; time = 0.0404s; samplesPerSecond = 253339.9
12/20/2016 15:28:21:  Epoch[14 of 25]-Minibatch[ 801- 810, 73.80%]: CE.SM = 1.02778320 * 10240; Err = 0.33925781 * 10240; time = 0.0410s; samplesPerSecond = 249500.5
12/20/2016 15:28:21:  Epoch[14 of 25]-Minibatch[ 811- 820, 74.72%]: CE.SM = 1.02680054 * 10240; Err = 0.33339844 * 10240; time = 0.0398s; samplesPerSecond = 257286.4
12/20/2016 15:28:22:  Epoch[14 of 25]-Minibatch[ 821- 830, 75.63%]: CE.SM = 1.02650146 * 10240; Err = 0.33613281 * 10240; time = 0.0403s; samplesPerSecond = 253911.6
12/20/2016 15:28:22:  Epoch[14 of 25]-Minibatch[ 831- 840, 76.54%]: CE.SM = 1.03098755 * 10240; Err = 0.33603516 * 10240; time = 0.0410s; samplesPerSecond = 249616.1
12/20/2016 15:28:22:  Epoch[14 of 25]-Minibatch[ 841- 850, 77.45%]: CE.SM = 1.02899780 * 10240; Err = 0.33154297 * 10240; time = 0.0420s; samplesPerSecond = 243618.1
12/20/2016 15:28:22:  Epoch[14 of 25]-Minibatch[ 851- 860, 78.36%]: CE.SM = 1.01939697 * 10240; Err = 0.33681641 * 10240; time = 0.0407s; samplesPerSecond = 251819.8
12/20/2016 15:28:22:  Epoch[14 of 25]-Minibatch[ 861- 870, 79.27%]: CE.SM = 1.01268311 * 10240; Err = 0.32744141 * 10240; time = 0.0399s; samplesPerSecond = 256628.7
12/20/2016 15:28:22:  Epoch[14 of 25]-Minibatch[ 871- 880, 80.18%]: CE.SM = 1.02803345 * 10240; Err = 0.33613281 * 10240; time = 0.0402s; samplesPerSecond = 254973.7
12/20/2016 15:28:22:  Epoch[14 of 25]-Minibatch[ 881- 890, 81.09%]: CE.SM = 1.02637939 * 10240; Err = 0.33320312 * 10240; time = 0.0399s; samplesPerSecond = 256892.7
12/20/2016 15:28:22:  Epoch[14 of 25]-Minibatch[ 891- 900, 82.00%]: CE.SM = 1.01644897 * 10240; Err = 0.33212891 * 10240; time = 0.0398s; samplesPerSecond = 257124.9
12/20/2016 15:28:22:  Epoch[14 of 25]-Minibatch[ 901- 910, 82.92%]: CE.SM = 1.04458618 * 10240; Err = 0.33994141 * 10240; time = 0.0410s; samplesPerSecond = 249464.0
12/20/2016 15:28:22:  Epoch[14 of 25]-Minibatch[ 911- 920, 83.83%]: CE.SM = 1.02128296 * 10240; Err = 0.33291016 * 10240; time = 0.0399s; samplesPerSecond = 256802.5
12/20/2016 15:28:22:  Epoch[14 of 25]-Minibatch[ 921- 930, 84.74%]: CE.SM = 1.04809570 * 10240; Err = 0.33603516 * 10240; time = 0.0403s; samplesPerSecond = 253993.5
12/20/2016 15:28:22:  Epoch[14 of 25]-Minibatch[ 931- 940, 85.65%]: CE.SM = 1.03884888 * 10240; Err = 0.33808594 * 10240; time = 0.0408s; samplesPerSecond = 251220.5
12/20/2016 15:28:22:  Epoch[14 of 25]-Minibatch[ 941- 950, 86.56%]: CE.SM = 1.04652100 * 10240; Err = 0.33837891 * 10240; time = 0.0407s; samplesPerSecond = 251424.1
12/20/2016 15:28:22:  Epoch[14 of 25]-Minibatch[ 951- 960, 87.47%]: CE.SM = 1.03746948 * 10240; Err = 0.33125000 * 10240; time = 0.0400s; samplesPerSecond = 255846.5
12/20/2016 15:28:22:  Epoch[14 of 25]-Minibatch[ 961- 970, 88.38%]: CE.SM = 1.03901367 * 10240; Err = 0.33701172 * 10240; time = 0.0401s; samplesPerSecond = 255183.4
12/20/2016 15:28:22:  Epoch[14 of 25]-Minibatch[ 971- 980, 89.29%]: CE.SM = 1.03485107 * 10240; Err = 0.34033203 * 10240; time = 0.0443s; samplesPerSecond = 231219.1
12/20/2016 15:28:22:  Epoch[14 of 25]-Minibatch[ 981- 990, 90.21%]: CE.SM = 1.01763916 * 10240; Err = 0.33427734 * 10240; time = 0.0402s; samplesPerSecond = 254745.4
12/20/2016 15:28:22:  Epoch[14 of 25]-Minibatch[ 991-1000, 91.12%]: CE.SM = 1.03403320 * 10240; Err = 0.33330078 * 10240; time = 0.0403s; samplesPerSecond = 254378.3
12/20/2016 15:28:22:  Epoch[14 of 25]-Minibatch[1001-1010, 92.03%]: CE.SM = 1.04884644 * 10240; Err = 0.33886719 * 10240; time = 0.0399s; samplesPerSecond = 256937.8
12/20/2016 15:28:22:  Epoch[14 of 25]-Minibatch[1011-1020, 92.94%]: CE.SM = 1.02645264 * 10240; Err = 0.33544922 * 10240; time = 0.0406s; samplesPerSecond = 252117.4
12/20/2016 15:28:22:  Epoch[14 of 25]-Minibatch[1021-1030, 93.85%]: CE.SM = 1.03624268 * 10240; Err = 0.33505859 * 10240; time = 0.0402s; samplesPerSecond = 254682.0
12/20/2016 15:28:22:  Epoch[14 of 25]-Minibatch[1031-1040, 94.76%]: CE.SM = 1.04022217 * 10240; Err = 0.34228516 * 10240; time = 0.0400s; samplesPerSecond = 256121.7
12/20/2016 15:28:22:  Epoch[14 of 25]-Minibatch[1041-1050, 95.67%]: CE.SM = 1.03172607 * 10240; Err = 0.33837891 * 10240; time = 0.0407s; samplesPerSecond = 251597.1
12/20/2016 15:28:22:  Epoch[14 of 25]-Minibatch[1051-1060, 96.58%]: CE.SM = 1.04587402 * 10240; Err = 0.33857422 * 10240; time = 0.0398s; samplesPerSecond = 257015.2
12/20/2016 15:28:23:  Epoch[14 of 25]-Minibatch[1061-1070, 97.49%]: CE.SM = 1.02377930 * 10240; Err = 0.33662109 * 10240; time = 0.0403s; samplesPerSecond = 253905.3
12/20/2016 15:28:23:  Epoch[14 of 25]-Minibatch[1071-1080, 98.41%]: CE.SM = 1.01791992 * 10240; Err = 0.32958984 * 10240; time = 0.0399s; samplesPerSecond = 256333.2
12/20/2016 15:28:23:  Epoch[14 of 25]-Minibatch[1081-1090, 99.32%]: CE.SM = 1.02556152 * 10240; Err = 0.33427734 * 10240; time = 0.0410s; samplesPerSecond = 249981.7
12/20/2016 15:28:23: Finished Epoch[14 of 25]: [Training] CE.SM = 1.02427026 * 1124823; Err = 0.33385697 * 1124823; totalSamplesSeen = 15747522; learningRatePerSample = 0.003125; epochTime=4.69827s
12/20/2016 15:28:23: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn.14'

12/20/2016 15:28:23: Starting Epoch 15: learning rate per sample = 0.003125  effective momentum = 0.900000  momentum as time constant = 9719.0 samples

12/20/2016 15:28:23: Starting minibatch loop.
12/20/2016 15:28:23:  Epoch[15 of 25]-Minibatch[   1-  10, 0.91%]: CE.SM = 0.99149876 * 10240; Err = 0.32373047 * 10240; time = 0.0447s; samplesPerSecond = 228980.3
12/20/2016 15:28:23:  Epoch[15 of 25]-Minibatch[  11-  20, 1.82%]: CE.SM = 1.00630436 * 10240; Err = 0.32666016 * 10240; time = 0.0409s; samplesPerSecond = 250244.4
12/20/2016 15:28:23:  Epoch[15 of 25]-Minibatch[  21-  30, 2.73%]: CE.SM = 1.00654297 * 10240; Err = 0.32578125 * 10240; time = 0.0404s; samplesPerSecond = 253697.7
12/20/2016 15:28:23:  Epoch[15 of 25]-Minibatch[  31-  40, 3.64%]: CE.SM = 1.00494690 * 10240; Err = 0.32998047 * 10240; time = 0.0403s; samplesPerSecond = 254296.2
12/20/2016 15:28:23:  Epoch[15 of 25]-Minibatch[  41-  50, 4.56%]: CE.SM = 1.00456123 * 10240; Err = 0.32490234 * 10240; time = 0.0405s; samplesPerSecond = 252870.7
12/20/2016 15:28:23:  Epoch[15 of 25]-Minibatch[  51-  60, 5.47%]: CE.SM = 0.99241638 * 10240; Err = 0.32187500 * 10240; time = 0.0401s; samplesPerSecond = 255399.8
12/20/2016 15:28:23:  Epoch[15 of 25]-Minibatch[  61-  70, 6.38%]: CE.SM = 0.98301392 * 10240; Err = 0.32802734 * 10240; time = 0.0408s; samplesPerSecond = 250882.0
12/20/2016 15:28:23:  Epoch[15 of 25]-Minibatch[  71-  80, 7.29%]: CE.SM = 0.99612122 * 10240; Err = 0.32548828 * 10240; time = 0.0403s; samplesPerSecond = 254359.4
12/20/2016 15:28:23:  Epoch[15 of 25]-Minibatch[  81-  90, 8.20%]: CE.SM = 0.99092484 * 10240; Err = 0.32539062 * 10240; time = 0.0403s; samplesPerSecond = 254378.3
12/20/2016 15:28:23:  Epoch[15 of 25]-Minibatch[  91- 100, 9.11%]: CE.SM = 0.97182159 * 10240; Err = 0.31923828 * 10240; time = 0.0408s; samplesPerSecond = 250685.5
12/20/2016 15:28:23:  Epoch[15 of 25]-Minibatch[ 101- 110, 10.02%]: CE.SM = 0.99532623 * 10240; Err = 0.31943359 * 10240; time = 0.0409s; samplesPerSecond = 250360.6
12/20/2016 15:28:23:  Epoch[15 of 25]-Minibatch[ 111- 120, 10.93%]: CE.SM = 0.98752747 * 10240; Err = 0.32500000 * 10240; time = 0.0403s; samplesPerSecond = 253810.9
12/20/2016 15:28:23:  Epoch[15 of 25]-Minibatch[ 121- 130, 11.85%]: CE.SM = 1.00560303 * 10240; Err = 0.32177734 * 10240; time = 0.0401s; samplesPerSecond = 255221.6
12/20/2016 15:28:23:  Epoch[15 of 25]-Minibatch[ 131- 140, 12.76%]: CE.SM = 0.99192810 * 10240; Err = 0.32480469 * 10240; time = 0.0408s; samplesPerSecond = 251269.8
12/20/2016 15:28:23:  Epoch[15 of 25]-Minibatch[ 141- 150, 13.67%]: CE.SM = 0.99968109 * 10240; Err = 0.32753906 * 10240; time = 0.0398s; samplesPerSecond = 257002.3
12/20/2016 15:28:24:  Epoch[15 of 25]-Minibatch[ 151- 160, 14.58%]: CE.SM = 0.96886292 * 10240; Err = 0.31445312 * 10240; time = 0.0423s; samplesPerSecond = 242303.8
12/20/2016 15:28:24:  Epoch[15 of 25]-Minibatch[ 161- 170, 15.49%]: CE.SM = 0.99850464 * 10240; Err = 0.32861328 * 10240; time = 0.0426s; samplesPerSecond = 240161.4
12/20/2016 15:28:24:  Epoch[15 of 25]-Minibatch[ 171- 180, 16.40%]: CE.SM = 0.99571228 * 10240; Err = 0.32460937 * 10240; time = 0.0402s; samplesPerSecond = 254846.8
12/20/2016 15:28:24:  Epoch[15 of 25]-Minibatch[ 181- 190, 17.31%]: CE.SM = 1.00394745 * 10240; Err = 0.32753906 * 10240; time = 0.0432s; samplesPerSecond = 237157.8
12/20/2016 15:28:24:  Epoch[15 of 25]-Minibatch[ 191- 200, 18.22%]: CE.SM = 1.00802002 * 10240; Err = 0.32773438 * 10240; time = 0.0416s; samplesPerSecond = 246047.4
12/20/2016 15:28:24:  Epoch[15 of 25]-Minibatch[ 201- 210, 19.13%]: CE.SM = 1.00590515 * 10240; Err = 0.32734375 * 10240; time = 0.0397s; samplesPerSecond = 257765.7
12/20/2016 15:28:24:  Epoch[15 of 25]-Minibatch[ 211- 220, 20.05%]: CE.SM = 1.00864868 * 10240; Err = 0.33066406 * 10240; time = 0.0404s; samplesPerSecond = 253741.7
12/20/2016 15:28:24:  Epoch[15 of 25]-Minibatch[ 221- 230, 20.96%]: CE.SM = 0.99833374 * 10240; Err = 0.32900391 * 10240; time = 0.0401s; samplesPerSecond = 255559.2
12/20/2016 15:28:24:  Epoch[15 of 25]-Minibatch[ 231- 240, 21.87%]: CE.SM = 0.99663696 * 10240; Err = 0.33027344 * 10240; time = 0.0404s; samplesPerSecond = 253528.1
12/20/2016 15:28:24:  Epoch[15 of 25]-Minibatch[ 241- 250, 22.78%]: CE.SM = 0.99343872 * 10240; Err = 0.32509766 * 10240; time = 0.0403s; samplesPerSecond = 253854.9
12/20/2016 15:28:24:  Epoch[15 of 25]-Minibatch[ 251- 260, 23.69%]: CE.SM = 0.98840942 * 10240; Err = 0.32304688 * 10240; time = 0.0407s; samplesPerSecond = 251665.1
12/20/2016 15:28:24:  Epoch[15 of 25]-Minibatch[ 261- 270, 24.60%]: CE.SM = 1.00335083 * 10240; Err = 0.32714844 * 10240; time = 0.0404s; samplesPerSecond = 253415.2
12/20/2016 15:28:24:  Epoch[15 of 25]-Minibatch[ 271- 280, 25.51%]: CE.SM = 0.99136047 * 10240; Err = 0.32509766 * 10240; time = 0.0410s; samplesPerSecond = 249737.8
12/20/2016 15:28:24:  Epoch[15 of 25]-Minibatch[ 281- 290, 26.42%]: CE.SM = 0.99117737 * 10240; Err = 0.32402344 * 10240; time = 0.0406s; samplesPerSecond = 252191.9
12/20/2016 15:28:24:  Epoch[15 of 25]-Minibatch[ 291- 300, 27.33%]: CE.SM = 0.99434509 * 10240; Err = 0.32285156 * 10240; time = 0.0408s; samplesPerSecond = 250937.3
12/20/2016 15:28:24:  Epoch[15 of 25]-Minibatch[ 301- 310, 28.25%]: CE.SM = 1.02209167 * 10240; Err = 0.33261719 * 10240; time = 0.0399s; samplesPerSecond = 256641.6
12/20/2016 15:28:24:  Epoch[15 of 25]-Minibatch[ 311- 320, 29.16%]: CE.SM = 1.01520691 * 10240; Err = 0.32841797 * 10240; time = 0.0406s; samplesPerSecond = 252235.4
12/20/2016 15:28:24:  Epoch[15 of 25]-Minibatch[ 321- 330, 30.07%]: CE.SM = 0.99738770 * 10240; Err = 0.32734375 * 10240; time = 0.0399s; samplesPerSecond = 256500.2
12/20/2016 15:28:24:  Epoch[15 of 25]-Minibatch[ 331- 340, 30.98%]: CE.SM = 0.99541321 * 10240; Err = 0.32617188 * 10240; time = 0.0400s; samplesPerSecond = 255820.9
12/20/2016 15:28:24:  Epoch[15 of 25]-Minibatch[ 341- 350, 31.89%]: CE.SM = 0.99474792 * 10240; Err = 0.33037109 * 10240; time = 0.0400s; samplesPerSecond = 255872.1
12/20/2016 15:28:24:  Epoch[15 of 25]-Minibatch[ 351- 360, 32.80%]: CE.SM = 1.02146912 * 10240; Err = 0.33222656 * 10240; time = 0.0400s; samplesPerSecond = 255859.3
12/20/2016 15:28:24:  Epoch[15 of 25]-Minibatch[ 361- 370, 33.71%]: CE.SM = 1.02563171 * 10240; Err = 0.34042969 * 10240; time = 0.0399s; samplesPerSecond = 256905.6
12/20/2016 15:28:24:  Epoch[15 of 25]-Minibatch[ 371- 380, 34.62%]: CE.SM = 1.00981445 * 10240; Err = 0.32783203 * 10240; time = 0.0401s; samplesPerSecond = 255355.2
12/20/2016 15:28:24:  Epoch[15 of 25]-Minibatch[ 381- 390, 35.54%]: CE.SM = 0.99905396 * 10240; Err = 0.32138672 * 10240; time = 0.0399s; samplesPerSecond = 256892.7
12/20/2016 15:28:25:  Epoch[15 of 25]-Minibatch[ 391- 400, 36.45%]: CE.SM = 0.99369202 * 10240; Err = 0.32265625 * 10240; time = 0.0396s; samplesPerSecond = 258690.4
12/20/2016 15:28:25:  Epoch[15 of 25]-Minibatch[ 401- 410, 37.36%]: CE.SM = 1.00304260 * 10240; Err = 0.32822266 * 10240; time = 0.0401s; samplesPerSecond = 255546.4
12/20/2016 15:28:25:  Epoch[15 of 25]-Minibatch[ 411- 420, 38.27%]: CE.SM = 1.02773438 * 10240; Err = 0.33173828 * 10240; time = 0.0403s; samplesPerSecond = 253987.2
12/20/2016 15:28:25:  Epoch[15 of 25]-Minibatch[ 421- 430, 39.18%]: CE.SM = 1.00973511 * 10240; Err = 0.32587891 * 10240; time = 0.0405s; samplesPerSecond = 252608.7
12/20/2016 15:28:25:  Epoch[15 of 25]-Minibatch[ 431- 440, 40.09%]: CE.SM = 1.00690308 * 10240; Err = 0.32871094 * 10240; time = 0.0402s; samplesPerSecond = 254568.1
12/20/2016 15:28:25:  Epoch[15 of 25]-Minibatch[ 441- 450, 41.00%]: CE.SM = 1.00464478 * 10240; Err = 0.33125000 * 10240; time = 0.0404s; samplesPerSecond = 253220.9
12/20/2016 15:28:25:  Epoch[15 of 25]-Minibatch[ 451- 460, 41.91%]: CE.SM = 0.99775391 * 10240; Err = 0.32851562 * 10240; time = 0.0407s; samplesPerSecond = 251442.6
12/20/2016 15:28:25:  Epoch[15 of 25]-Minibatch[ 461- 470, 42.82%]: CE.SM = 1.02959595 * 10240; Err = 0.33955078 * 10240; time = 0.0404s; samplesPerSecond = 253540.7
12/20/2016 15:28:25:  Epoch[15 of 25]-Minibatch[ 471- 480, 43.74%]: CE.SM = 1.01241455 * 10240; Err = 0.33017578 * 10240; time = 0.0403s; samplesPerSecond = 254031.3
12/20/2016 15:28:25:  Epoch[15 of 25]-Minibatch[ 481- 490, 44.65%]: CE.SM = 1.00711670 * 10240; Err = 0.32324219 * 10240; time = 0.0405s; samplesPerSecond = 252902.0
12/20/2016 15:28:25:  Epoch[15 of 25]-Minibatch[ 491- 500, 45.56%]: CE.SM = 1.00159302 * 10240; Err = 0.32998047 * 10240; time = 0.0407s; samplesPerSecond = 251696.0
12/20/2016 15:28:25:  Epoch[15 of 25]-Minibatch[ 501- 510, 46.47%]: CE.SM = 1.01038208 * 10240; Err = 0.33564453 * 10240; time = 0.0404s; samplesPerSecond = 253521.8
12/20/2016 15:28:25:  Epoch[15 of 25]-Minibatch[ 511- 520, 47.38%]: CE.SM = 1.01987305 * 10240; Err = 0.33398438 * 10240; time = 0.0405s; samplesPerSecond = 252739.7
12/20/2016 15:28:25:  Epoch[15 of 25]-Minibatch[ 521- 530, 48.29%]: CE.SM = 0.96953125 * 10240; Err = 0.32021484 * 10240; time = 0.0409s; samplesPerSecond = 250244.4
12/20/2016 15:28:25:  Epoch[15 of 25]-Minibatch[ 531- 540, 49.20%]: CE.SM = 1.00943604 * 10240; Err = 0.32802734 * 10240; time = 0.0412s; samplesPerSecond = 248573.9
12/20/2016 15:28:25:  Epoch[15 of 25]-Minibatch[ 541- 550, 50.11%]: CE.SM = 0.98834839 * 10240; Err = 0.32177734 * 10240; time = 0.0403s; samplesPerSecond = 254207.8
12/20/2016 15:28:25:  Epoch[15 of 25]-Minibatch[ 551- 560, 51.03%]: CE.SM = 1.01497803 * 10240; Err = 0.33056641 * 10240; time = 0.0406s; samplesPerSecond = 252036.7
12/20/2016 15:28:25:  Epoch[15 of 25]-Minibatch[ 561- 570, 51.94%]: CE.SM = 0.97204590 * 10240; Err = 0.31992188 * 10240; time = 0.0402s; samplesPerSecond = 255030.9
12/20/2016 15:28:25:  Epoch[15 of 25]-Minibatch[ 571- 580, 52.85%]: CE.SM = 1.03496704 * 10240; Err = 0.33603516 * 10240; time = 0.0401s; samplesPerSecond = 255642.1
12/20/2016 15:28:25:  Epoch[15 of 25]-Minibatch[ 581- 590, 53.76%]: CE.SM = 1.03123169 * 10240; Err = 0.33583984 * 10240; time = 0.0406s; samplesPerSecond = 252291.3
12/20/2016 15:28:25:  Epoch[15 of 25]-Minibatch[ 591- 600, 54.67%]: CE.SM = 0.99764404 * 10240; Err = 0.32880859 * 10240; time = 0.0400s; samplesPerSecond = 255680.4
12/20/2016 15:28:25:  Epoch[15 of 25]-Minibatch[ 601- 610, 55.58%]: CE.SM = 1.00866089 * 10240; Err = 0.33134766 * 10240; time = 0.0401s; samplesPerSecond = 255215.2
12/20/2016 15:28:25:  Epoch[15 of 25]-Minibatch[ 611- 620, 56.49%]: CE.SM = 1.00814209 * 10240; Err = 0.33203125 * 10240; time = 0.0403s; samplesPerSecond = 254220.5
12/20/2016 15:28:25:  Epoch[15 of 25]-Minibatch[ 621- 630, 57.40%]: CE.SM = 1.00563965 * 10240; Err = 0.33115234 * 10240; time = 0.0406s; samplesPerSecond = 252204.3
12/20/2016 15:28:25:  Epoch[15 of 25]-Minibatch[ 631- 640, 58.31%]: CE.SM = 1.01726074 * 10240; Err = 0.32939453 * 10240; time = 0.0428s; samplesPerSecond = 239012.2
12/20/2016 15:28:26:  Epoch[15 of 25]-Minibatch[ 641- 650, 59.23%]: CE.SM = 1.00595093 * 10240; Err = 0.33154297 * 10240; time = 0.0477s; samplesPerSecond = 214823.7
12/20/2016 15:28:26:  Epoch[15 of 25]-Minibatch[ 651- 660, 60.14%]: CE.SM = 1.00419922 * 10240; Err = 0.32763672 * 10240; time = 0.0495s; samplesPerSecond = 206969.0
12/20/2016 15:28:26:  Epoch[15 of 25]-Minibatch[ 661- 670, 61.05%]: CE.SM = 0.99892578 * 10240; Err = 0.33017578 * 10240; time = 0.0413s; samplesPerSecond = 248140.2
12/20/2016 15:28:26:  Epoch[15 of 25]-Minibatch[ 671- 680, 61.96%]: CE.SM = 1.01533203 * 10240; Err = 0.32958984 * 10240; time = 0.0408s; samplesPerSecond = 250851.3
12/20/2016 15:28:26:  Epoch[15 of 25]-Minibatch[ 681- 690, 62.87%]: CE.SM = 0.98330688 * 10240; Err = 0.31914063 * 10240; time = 0.0394s; samplesPerSecond = 259832.5
12/20/2016 15:28:26:  Epoch[15 of 25]-Minibatch[ 691- 700, 63.78%]: CE.SM = 1.01292114 * 10240; Err = 0.32871094 * 10240; time = 0.0437s; samplesPerSecond = 234357.1
12/20/2016 15:28:26:  Epoch[15 of 25]-Minibatch[ 701- 710, 64.69%]: CE.SM = 1.02352905 * 10240; Err = 0.33193359 * 10240; time = 0.0428s; samplesPerSecond = 239442.5
12/20/2016 15:28:26:  Epoch[15 of 25]-Minibatch[ 711- 720, 65.60%]: CE.SM = 1.02783813 * 10240; Err = 0.33486328 * 10240; time = 0.0429s; samplesPerSecond = 238477.8
12/20/2016 15:28:26:  Epoch[15 of 25]-Minibatch[ 721- 730, 66.51%]: CE.SM = 0.98634644 * 10240; Err = 0.32695313 * 10240; time = 0.0416s; samplesPerSecond = 245929.2
12/20/2016 15:28:26:  Epoch[15 of 25]-Minibatch[ 731- 740, 67.43%]: CE.SM = 1.01493530 * 10240; Err = 0.32548828 * 10240; time = 0.0409s; samplesPerSecond = 250348.4
12/20/2016 15:28:26:  Epoch[15 of 25]-Minibatch[ 741- 750, 68.34%]: CE.SM = 1.03643799 * 10240; Err = 0.33583984 * 10240; time = 0.0412s; samplesPerSecond = 248489.4
12/20/2016 15:28:26:  Epoch[15 of 25]-Minibatch[ 751- 760, 69.25%]: CE.SM = 0.99200439 * 10240; Err = 0.32695313 * 10240; time = 0.0407s; samplesPerSecond = 251640.3
12/20/2016 15:28:26:  Epoch[15 of 25]-Minibatch[ 761- 770, 70.16%]: CE.SM = 1.02038574 * 10240; Err = 0.33066406 * 10240; time = 0.0410s; samplesPerSecond = 249597.8
12/20/2016 15:28:26:  Epoch[15 of 25]-Minibatch[ 771- 780, 71.07%]: CE.SM = 1.00277710 * 10240; Err = 0.33251953 * 10240; time = 0.0395s; samplesPerSecond = 258978.2
12/20/2016 15:28:26:  Epoch[15 of 25]-Minibatch[ 781- 790, 71.98%]: CE.SM = 0.98495483 * 10240; Err = 0.32343750 * 10240; time = 0.0410s; samplesPerSecond = 249817.0
12/20/2016 15:28:26:  Epoch[15 of 25]-Minibatch[ 791- 800, 72.89%]: CE.SM = 1.01039429 * 10240; Err = 0.32636719 * 10240; time = 0.0405s; samplesPerSecond = 253001.9
12/20/2016 15:28:26:  Epoch[15 of 25]-Minibatch[ 801- 810, 73.80%]: CE.SM = 1.03341064 * 10240; Err = 0.33320312 * 10240; time = 0.0406s; samplesPerSecond = 252316.2
12/20/2016 15:28:26:  Epoch[15 of 25]-Minibatch[ 811- 820, 74.72%]: CE.SM = 1.01683960 * 10240; Err = 0.33017578 * 10240; time = 0.0397s; samplesPerSecond = 257811.1
12/20/2016 15:28:26:  Epoch[15 of 25]-Minibatch[ 821- 830, 75.63%]: CE.SM = 1.01015015 * 10240; Err = 0.32880859 * 10240; time = 0.0410s; samplesPerSecond = 249579.6
12/20/2016 15:28:26:  Epoch[15 of 25]-Minibatch[ 831- 840, 76.54%]: CE.SM = 1.01224976 * 10240; Err = 0.33154297 * 10240; time = 0.0401s; samplesPerSecond = 255654.9
12/20/2016 15:28:26:  Epoch[15 of 25]-Minibatch[ 841- 850, 77.45%]: CE.SM = 1.00236816 * 10240; Err = 0.32578125 * 10240; time = 0.0403s; samplesPerSecond = 253842.3
12/20/2016 15:28:26:  Epoch[15 of 25]-Minibatch[ 851- 860, 78.36%]: CE.SM = 1.02648926 * 10240; Err = 0.33720703 * 10240; time = 0.0407s; samplesPerSecond = 251566.1
12/20/2016 15:28:26:  Epoch[15 of 25]-Minibatch[ 861- 870, 79.27%]: CE.SM = 1.01269531 * 10240; Err = 0.33222656 * 10240; time = 0.0464s; samplesPerSecond = 220547.1
12/20/2016 15:28:26:  Epoch[15 of 25]-Minibatch[ 871- 880, 80.18%]: CE.SM = 0.98779297 * 10240; Err = 0.31943359 * 10240; time = 0.0406s; samplesPerSecond = 252421.9
12/20/2016 15:28:27:  Epoch[15 of 25]-Minibatch[ 881- 890, 81.09%]: CE.SM = 1.00421143 * 10240; Err = 0.32900391 * 10240; time = 0.0413s; samplesPerSecond = 247720.0
12/20/2016 15:28:27:  Epoch[15 of 25]-Minibatch[ 891- 900, 82.00%]: CE.SM = 1.02113647 * 10240; Err = 0.33339844 * 10240; time = 0.0423s; samplesPerSecond = 242177.7
12/20/2016 15:28:27:  Epoch[15 of 25]-Minibatch[ 901- 910, 82.92%]: CE.SM = 1.00871582 * 10240; Err = 0.32744141 * 10240; time = 0.0417s; samplesPerSecond = 245334.1
12/20/2016 15:28:27:  Epoch[15 of 25]-Minibatch[ 911- 920, 83.83%]: CE.SM = 0.98062744 * 10240; Err = 0.32275391 * 10240; time = 0.0406s; samplesPerSecond = 252428.1
12/20/2016 15:28:27:  Epoch[15 of 25]-Minibatch[ 921- 930, 84.74%]: CE.SM = 1.02538452 * 10240; Err = 0.33173828 * 10240; time = 0.0399s; samplesPerSecond = 256615.9
12/20/2016 15:28:27:  Epoch[15 of 25]-Minibatch[ 931- 940, 85.65%]: CE.SM = 1.02030029 * 10240; Err = 0.33564453 * 10240; time = 0.0405s; samplesPerSecond = 252914.4
12/20/2016 15:28:27:  Epoch[15 of 25]-Minibatch[ 941- 950, 86.56%]: CE.SM = 1.00197754 * 10240; Err = 0.33164063 * 10240; time = 0.0398s; samplesPerSecond = 257538.8
12/20/2016 15:28:27:  Epoch[15 of 25]-Minibatch[ 951- 960, 87.47%]: CE.SM = 0.99797363 * 10240; Err = 0.32392578 * 10240; time = 0.0403s; samplesPerSecond = 253823.5
12/20/2016 15:28:27:  Epoch[15 of 25]-Minibatch[ 961- 970, 88.38%]: CE.SM = 1.01544800 * 10240; Err = 0.33583984 * 10240; time = 0.0415s; samplesPerSecond = 246830.3
12/20/2016 15:28:27:  Epoch[15 of 25]-Minibatch[ 971- 980, 89.29%]: CE.SM = 1.01843262 * 10240; Err = 0.33437500 * 10240; time = 0.0403s; samplesPerSecond = 254081.7
12/20/2016 15:28:27:  Epoch[15 of 25]-Minibatch[ 981- 990, 90.21%]: CE.SM = 0.99081421 * 10240; Err = 0.32421875 * 10240; time = 0.0395s; samplesPerSecond = 258991.4
12/20/2016 15:28:27:  Epoch[15 of 25]-Minibatch[ 991-1000, 91.12%]: CE.SM = 0.99042969 * 10240; Err = 0.32929687 * 10240; time = 0.0409s; samplesPerSecond = 250134.3
12/20/2016 15:28:27:  Epoch[15 of 25]-Minibatch[1001-1010, 92.03%]: CE.SM = 1.01426392 * 10240; Err = 0.32978516 * 10240; time = 0.0410s; samplesPerSecond = 249884.1
12/20/2016 15:28:27:  Epoch[15 of 25]-Minibatch[1011-1020, 92.94%]: CE.SM = 0.99699097 * 10240; Err = 0.32958984 * 10240; time = 0.0432s; samplesPerSecond = 237080.9
12/20/2016 15:28:27:  Epoch[15 of 25]-Minibatch[1021-1030, 93.85%]: CE.SM = 0.99622803 * 10240; Err = 0.32568359 * 10240; time = 0.0407s; samplesPerSecond = 251863.1
12/20/2016 15:28:27:  Epoch[15 of 25]-Minibatch[1031-1040, 94.76%]: CE.SM = 1.01579590 * 10240; Err = 0.33710937 * 10240; time = 0.0402s; samplesPerSecond = 254891.2
12/20/2016 15:28:27:  Epoch[15 of 25]-Minibatch[1041-1050, 95.67%]: CE.SM = 1.02604980 * 10240; Err = 0.33476563 * 10240; time = 0.0407s; samplesPerSecond = 251584.7
12/20/2016 15:28:27:  Epoch[15 of 25]-Minibatch[1051-1060, 96.58%]: CE.SM = 1.00106201 * 10240; Err = 0.33007812 * 10240; time = 0.0415s; samplesPerSecond = 246545.0
12/20/2016 15:28:27:  Epoch[15 of 25]-Minibatch[1061-1070, 97.49%]: CE.SM = 0.97840576 * 10240; Err = 0.31689453 * 10240; time = 0.0411s; samplesPerSecond = 249051.5
12/20/2016 15:28:27:  Epoch[15 of 25]-Minibatch[1071-1080, 98.41%]: CE.SM = 1.01240234 * 10240; Err = 0.32724609 * 10240; time = 0.0397s; samplesPerSecond = 257610.1
12/20/2016 15:28:27:  Epoch[15 of 25]-Minibatch[1081-1090, 99.32%]: CE.SM = 1.01557617 * 10240; Err = 0.33095703 * 10240; time = 0.0408s; samplesPerSecond = 250900.4
12/20/2016 15:28:27: Finished Epoch[15 of 25]: [Training] CE.SM = 1.00453927 * 1124823; Err = 0.32829076 * 1124823; totalSamplesSeen = 16872345; learningRatePerSample = 0.003125; epochTime=4.6851s
12/20/2016 15:28:27: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn.15'

12/20/2016 15:28:27: Starting Epoch 16: learning rate per sample = 0.000078  effective momentum = 0.900000  momentum as time constant = 9719.0 samples

12/20/2016 15:28:28: Starting minibatch loop.
12/20/2016 15:28:28:  Epoch[16 of 25]-Minibatch[   1-  10, 0.91%]: CE.SM = 1.07960358 * 10240; Err = 0.35673828 * 10240; time = 0.0528s; samplesPerSecond = 193994.5
12/20/2016 15:28:28:  Epoch[16 of 25]-Minibatch[  11-  20, 1.82%]: CE.SM = 1.16351948 * 10240; Err = 0.37890625 * 10240; time = 0.0413s; samplesPerSecond = 248050.0
12/20/2016 15:28:28:  Epoch[16 of 25]-Minibatch[  21-  30, 2.73%]: CE.SM = 1.14358425 * 10240; Err = 0.37089844 * 10240; time = 0.0416s; samplesPerSecond = 245858.3
12/20/2016 15:28:28:  Epoch[16 of 25]-Minibatch[  31-  40, 3.64%]: CE.SM = 1.04223480 * 10240; Err = 0.34316406 * 10240; time = 0.0412s; samplesPerSecond = 248839.6
12/20/2016 15:28:28:  Epoch[16 of 25]-Minibatch[  41-  50, 4.56%]: CE.SM = 1.00647392 * 10240; Err = 0.33339844 * 10240; time = 0.0436s; samplesPerSecond = 234700.9
12/20/2016 15:28:28:  Epoch[16 of 25]-Minibatch[  51-  60, 5.47%]: CE.SM = 1.01250725 * 10240; Err = 0.33017578 * 10240; time = 0.0403s; samplesPerSecond = 253798.3
12/20/2016 15:28:28:  Epoch[16 of 25]-Minibatch[  61-  70, 6.38%]: CE.SM = 1.00059814 * 10240; Err = 0.32246094 * 10240; time = 0.0404s; samplesPerSecond = 253704.0
12/20/2016 15:28:28:  Epoch[16 of 25]-Minibatch[  71-  80, 7.29%]: CE.SM = 0.97044144 * 10240; Err = 0.31884766 * 10240; time = 0.0440s; samplesPerSecond = 232806.6
12/20/2016 15:28:28:  Epoch[16 of 25]-Minibatch[  81-  90, 8.20%]: CE.SM = 0.96881332 * 10240; Err = 0.31777344 * 10240; time = 0.0402s; samplesPerSecond = 254549.1
12/20/2016 15:28:28:  Epoch[16 of 25]-Minibatch[  91- 100, 9.11%]: CE.SM = 0.94943085 * 10240; Err = 0.30781250 * 10240; time = 0.0400s; samplesPerSecond = 256288.3
12/20/2016 15:28:28:  Epoch[16 of 25]-Minibatch[ 101- 110, 10.02%]: CE.SM = 0.93818741 * 10240; Err = 0.30751953 * 10240; time = 0.0404s; samplesPerSecond = 253358.7
12/20/2016 15:28:28:  Epoch[16 of 25]-Minibatch[ 111- 120, 10.93%]: CE.SM = 0.93485107 * 10240; Err = 0.30761719 * 10240; time = 0.0412s; samplesPerSecond = 248507.5
12/20/2016 15:28:28:  Epoch[16 of 25]-Minibatch[ 121- 130, 11.85%]: CE.SM = 0.94505081 * 10240; Err = 0.30800781 * 10240; time = 0.0407s; samplesPerSecond = 251875.5
12/20/2016 15:28:28:  Epoch[16 of 25]-Minibatch[ 131- 140, 12.76%]: CE.SM = 0.96141510 * 10240; Err = 0.31201172 * 10240; time = 0.0409s; samplesPerSecond = 250317.8
12/20/2016 15:28:28:  Epoch[16 of 25]-Minibatch[ 141- 150, 13.67%]: CE.SM = 0.95787811 * 10240; Err = 0.31132813 * 10240; time = 0.0403s; samplesPerSecond = 253880.1
12/20/2016 15:28:28:  Epoch[16 of 25]-Minibatch[ 151- 160, 14.58%]: CE.SM = 0.93242035 * 10240; Err = 0.30751953 * 10240; time = 0.0412s; samplesPerSecond = 248567.8
12/20/2016 15:28:28:  Epoch[16 of 25]-Minibatch[ 161- 170, 15.49%]: CE.SM = 0.92131805 * 10240; Err = 0.29951172 * 10240; time = 0.0412s; samplesPerSecond = 248773.1
12/20/2016 15:28:28:  Epoch[16 of 25]-Minibatch[ 171- 180, 16.40%]: CE.SM = 0.91912842 * 10240; Err = 0.29863281 * 10240; time = 0.0420s; samplesPerSecond = 243856.0
12/20/2016 15:28:28:  Epoch[16 of 25]-Minibatch[ 181- 190, 17.31%]: CE.SM = 0.95112762 * 10240; Err = 0.30761719 * 10240; time = 0.0410s; samplesPerSecond = 250012.2
12/20/2016 15:28:28:  Epoch[16 of 25]-Minibatch[ 191- 200, 18.22%]: CE.SM = 0.94266205 * 10240; Err = 0.31005859 * 10240; time = 0.0403s; samplesPerSecond = 254106.9
12/20/2016 15:28:29:  Epoch[16 of 25]-Minibatch[ 201- 210, 19.13%]: CE.SM = 0.93181763 * 10240; Err = 0.30595703 * 10240; time = 0.0401s; samplesPerSecond = 255654.9
12/20/2016 15:28:29:  Epoch[16 of 25]-Minibatch[ 211- 220, 20.05%]: CE.SM = 0.93152161 * 10240; Err = 0.30527344 * 10240; time = 0.0410s; samplesPerSecond = 249963.4
12/20/2016 15:28:29:  Epoch[16 of 25]-Minibatch[ 221- 230, 20.96%]: CE.SM = 0.92882843 * 10240; Err = 0.30292969 * 10240; time = 0.0404s; samplesPerSecond = 253578.3
12/20/2016 15:28:29:  Epoch[16 of 25]-Minibatch[ 231- 240, 21.87%]: CE.SM = 0.94536133 * 10240; Err = 0.30107422 * 10240; time = 0.0405s; samplesPerSecond = 252552.7
12/20/2016 15:28:29:  Epoch[16 of 25]-Minibatch[ 241- 250, 22.78%]: CE.SM = 0.93947754 * 10240; Err = 0.30507812 * 10240; time = 0.0409s; samplesPerSecond = 250140.5
12/20/2016 15:28:29:  Epoch[16 of 25]-Minibatch[ 251- 260, 23.69%]: CE.SM = 0.92796021 * 10240; Err = 0.30332031 * 10240; time = 0.0404s; samplesPerSecond = 253691.4
12/20/2016 15:28:29:  Epoch[16 of 25]-Minibatch[ 261- 270, 24.60%]: CE.SM = 0.91209717 * 10240; Err = 0.30146484 * 10240; time = 0.0420s; samplesPerSecond = 243542.8
12/20/2016 15:28:29:  Epoch[16 of 25]-Minibatch[ 271- 280, 25.51%]: CE.SM = 0.93193359 * 10240; Err = 0.29921875 * 10240; time = 0.0411s; samplesPerSecond = 249451.9
12/20/2016 15:28:29:  Epoch[16 of 25]-Minibatch[ 281- 290, 26.42%]: CE.SM = 0.92875671 * 10240; Err = 0.30195312 * 10240; time = 0.0410s; samplesPerSecond = 249518.8
12/20/2016 15:28:29:  Epoch[16 of 25]-Minibatch[ 291- 300, 27.33%]: CE.SM = 0.93595276 * 10240; Err = 0.30439453 * 10240; time = 0.0412s; samplesPerSecond = 248386.9
12/20/2016 15:28:29:  Epoch[16 of 25]-Minibatch[ 301- 310, 28.25%]: CE.SM = 0.92866211 * 10240; Err = 0.30390625 * 10240; time = 0.0399s; samplesPerSecond = 256738.1
12/20/2016 15:28:29:  Epoch[16 of 25]-Minibatch[ 311- 320, 29.16%]: CE.SM = 0.90826721 * 10240; Err = 0.29345703 * 10240; time = 0.0399s; samplesPerSecond = 256622.3
12/20/2016 15:28:29:  Epoch[16 of 25]-Minibatch[ 321- 330, 30.07%]: CE.SM = 0.92055359 * 10240; Err = 0.29814453 * 10240; time = 0.0405s; samplesPerSecond = 252908.2
12/20/2016 15:28:29:  Epoch[16 of 25]-Minibatch[ 331- 340, 30.98%]: CE.SM = 0.91578674 * 10240; Err = 0.29951172 * 10240; time = 0.0405s; samplesPerSecond = 252895.7
12/20/2016 15:28:29:  Epoch[16 of 25]-Minibatch[ 341- 350, 31.89%]: CE.SM = 0.93758850 * 10240; Err = 0.30527344 * 10240; time = 0.0398s; samplesPerSecond = 257364.0
12/20/2016 15:28:29:  Epoch[16 of 25]-Minibatch[ 351- 360, 32.80%]: CE.SM = 0.92815247 * 10240; Err = 0.30029297 * 10240; time = 0.0421s; samplesPerSecond = 243490.7
12/20/2016 15:28:29:  Epoch[16 of 25]-Minibatch[ 361- 370, 33.71%]: CE.SM = 0.91559143 * 10240; Err = 0.29882812 * 10240; time = 0.0401s; samplesPerSecond = 255355.2
12/20/2016 15:28:29:  Epoch[16 of 25]-Minibatch[ 371- 380, 34.62%]: CE.SM = 0.92392273 * 10240; Err = 0.30371094 * 10240; time = 0.0408s; samplesPerSecond = 250703.9
12/20/2016 15:28:29:  Epoch[16 of 25]-Minibatch[ 381- 390, 35.54%]: CE.SM = 0.94250183 * 10240; Err = 0.31044922 * 10240; time = 0.0395s; samplesPerSecond = 259444.1
12/20/2016 15:28:29:  Epoch[16 of 25]-Minibatch[ 391- 400, 36.45%]: CE.SM = 0.91598511 * 10240; Err = 0.29794922 * 10240; time = 0.0398s; samplesPerSecond = 257422.3
12/20/2016 15:28:29:  Epoch[16 of 25]-Minibatch[ 401- 410, 37.36%]: CE.SM = 0.91358643 * 10240; Err = 0.29755859 * 10240; time = 0.0403s; samplesPerSecond = 254176.3
12/20/2016 15:28:29:  Epoch[16 of 25]-Minibatch[ 411- 420, 38.27%]: CE.SM = 0.93222351 * 10240; Err = 0.30419922 * 10240; time = 0.0401s; samplesPerSecond = 255425.3
12/20/2016 15:28:29:  Epoch[16 of 25]-Minibatch[ 421- 430, 39.18%]: CE.SM = 0.91119995 * 10240; Err = 0.29726562 * 10240; time = 0.0401s; samplesPerSecond = 255113.5
12/20/2016 15:28:29:  Epoch[16 of 25]-Minibatch[ 431- 440, 40.09%]: CE.SM = 0.92104797 * 10240; Err = 0.30117187 * 10240; time = 0.0404s; samplesPerSecond = 253383.8
12/20/2016 15:28:30:  Epoch[16 of 25]-Minibatch[ 441- 450, 41.00%]: CE.SM = 0.93521423 * 10240; Err = 0.30712891 * 10240; time = 0.0421s; samplesPerSecond = 243415.4
12/20/2016 15:28:30:  Epoch[16 of 25]-Minibatch[ 451- 460, 41.91%]: CE.SM = 0.91877136 * 10240; Err = 0.30302734 * 10240; time = 0.0405s; samplesPerSecond = 252808.3
12/20/2016 15:28:30:  Epoch[16 of 25]-Minibatch[ 461- 470, 42.82%]: CE.SM = 0.90205994 * 10240; Err = 0.29560547 * 10240; time = 0.0447s; samplesPerSecond = 229154.5
12/20/2016 15:28:30:  Epoch[16 of 25]-Minibatch[ 471- 480, 43.74%]: CE.SM = 0.94646301 * 10240; Err = 0.30878906 * 10240; time = 0.0405s; samplesPerSecond = 252558.9
12/20/2016 15:28:30:  Epoch[16 of 25]-Minibatch[ 481- 490, 44.65%]: CE.SM = 0.91907349 * 10240; Err = 0.30156250 * 10240; time = 0.0406s; samplesPerSecond = 252378.4
12/20/2016 15:28:30:  Epoch[16 of 25]-Minibatch[ 491- 500, 45.56%]: CE.SM = 0.92705994 * 10240; Err = 0.30058594 * 10240; time = 0.0435s; samplesPerSecond = 235326.6
12/20/2016 15:28:30:  Epoch[16 of 25]-Minibatch[ 501- 510, 46.47%]: CE.SM = 0.91704712 * 10240; Err = 0.29873047 * 10240; time = 0.0442s; samplesPerSecond = 231915.6
12/20/2016 15:28:30:  Epoch[16 of 25]-Minibatch[ 511- 520, 47.38%]: CE.SM = 0.92648621 * 10240; Err = 0.30195312 * 10240; time = 0.0405s; samplesPerSecond = 253033.2
12/20/2016 15:28:30:  Epoch[16 of 25]-Minibatch[ 521- 530, 48.29%]: CE.SM = 0.92183228 * 10240; Err = 0.30478516 * 10240; time = 0.0414s; samplesPerSecond = 247146.0
12/20/2016 15:28:30:  Epoch[16 of 25]-Minibatch[ 531- 540, 49.20%]: CE.SM = 0.93182068 * 10240; Err = 0.30615234 * 10240; time = 0.0409s; samplesPerSecond = 250128.2
12/20/2016 15:28:30:  Epoch[16 of 25]-Minibatch[ 541- 550, 50.11%]: CE.SM = 0.91538696 * 10240; Err = 0.29521484 * 10240; time = 0.0403s; samplesPerSecond = 254384.7
12/20/2016 15:28:30:  Epoch[16 of 25]-Minibatch[ 551- 560, 51.03%]: CE.SM = 0.89108887 * 10240; Err = 0.29150391 * 10240; time = 0.0409s; samplesPerSecond = 250067.2
12/20/2016 15:28:30:  Epoch[16 of 25]-Minibatch[ 561- 570, 51.94%]: CE.SM = 0.91610718 * 10240; Err = 0.29853516 * 10240; time = 0.0412s; samplesPerSecond = 248356.8
12/20/2016 15:28:30:  Epoch[16 of 25]-Minibatch[ 571- 580, 52.85%]: CE.SM = 0.90562134 * 10240; Err = 0.29462891 * 10240; time = 0.0409s; samplesPerSecond = 250128.2
12/20/2016 15:28:30:  Epoch[16 of 25]-Minibatch[ 581- 590, 53.76%]: CE.SM = 0.89802856 * 10240; Err = 0.29560547 * 10240; time = 0.0406s; samplesPerSecond = 252198.1
12/20/2016 15:28:30:  Epoch[16 of 25]-Minibatch[ 591- 600, 54.67%]: CE.SM = 0.90948486 * 10240; Err = 0.29833984 * 10240; time = 0.0400s; samplesPerSecond = 255833.7
12/20/2016 15:28:30:  Epoch[16 of 25]-Minibatch[ 601- 610, 55.58%]: CE.SM = 0.90927734 * 10240; Err = 0.29775391 * 10240; time = 0.0401s; samplesPerSecond = 255317.0
12/20/2016 15:28:30:  Epoch[16 of 25]-Minibatch[ 611- 620, 56.49%]: CE.SM = 0.88792725 * 10240; Err = 0.28789063 * 10240; time = 0.0417s; samplesPerSecond = 245752.1
12/20/2016 15:28:30:  Epoch[16 of 25]-Minibatch[ 621- 630, 57.40%]: CE.SM = 0.91231079 * 10240; Err = 0.30039063 * 10240; time = 0.0405s; samplesPerSecond = 253045.7
12/20/2016 15:28:30:  Epoch[16 of 25]-Minibatch[ 631- 640, 58.31%]: CE.SM = 0.91224976 * 10240; Err = 0.29277344 * 10240; time = 0.0399s; samplesPerSecond = 256738.1
12/20/2016 15:28:30:  Epoch[16 of 25]-Minibatch[ 641- 650, 59.23%]: CE.SM = 0.88911133 * 10240; Err = 0.29208984 * 10240; time = 0.0398s; samplesPerSecond = 257519.4
12/20/2016 15:28:30:  Epoch[16 of 25]-Minibatch[ 651- 660, 60.14%]: CE.SM = 0.91264038 * 10240; Err = 0.29111328 * 10240; time = 0.0414s; samplesPerSecond = 247468.5
12/20/2016 15:28:30:  Epoch[16 of 25]-Minibatch[ 661- 670, 61.05%]: CE.SM = 0.88978882 * 10240; Err = 0.29033203 * 10240; time = 0.0406s; samplesPerSecond = 252266.5
12/20/2016 15:28:30:  Epoch[16 of 25]-Minibatch[ 671- 680, 61.96%]: CE.SM = 0.89997559 * 10240; Err = 0.29531250 * 10240; time = 0.0404s; samplesPerSecond = 253697.7
12/20/2016 15:28:30:  Epoch[16 of 25]-Minibatch[ 681- 690, 62.87%]: CE.SM = 0.92284546 * 10240; Err = 0.29501953 * 10240; time = 0.0405s; samplesPerSecond = 252602.5
12/20/2016 15:28:31:  Epoch[16 of 25]-Minibatch[ 691- 700, 63.78%]: CE.SM = 0.90628662 * 10240; Err = 0.29541016 * 10240; time = 0.0399s; samplesPerSecond = 256847.6
12/20/2016 15:28:31:  Epoch[16 of 25]-Minibatch[ 701- 710, 64.69%]: CE.SM = 0.92286987 * 10240; Err = 0.29775391 * 10240; time = 0.0415s; samplesPerSecond = 246479.7
12/20/2016 15:28:31:  Epoch[16 of 25]-Minibatch[ 711- 720, 65.60%]: CE.SM = 0.89057617 * 10240; Err = 0.28974609 * 10240; time = 0.0403s; samplesPerSecond = 254327.8
12/20/2016 15:28:31:  Epoch[16 of 25]-Minibatch[ 721- 730, 66.51%]: CE.SM = 0.90949097 * 10240; Err = 0.29658203 * 10240; time = 0.0402s; samplesPerSecond = 254416.3
12/20/2016 15:28:31:  Epoch[16 of 25]-Minibatch[ 731- 740, 67.43%]: CE.SM = 0.92474365 * 10240; Err = 0.29882812 * 10240; time = 0.0409s; samplesPerSecond = 250311.7
12/20/2016 15:28:31:  Epoch[16 of 25]-Minibatch[ 741- 750, 68.34%]: CE.SM = 0.88953247 * 10240; Err = 0.28886719 * 10240; time = 0.0397s; samplesPerSecond = 257746.2
12/20/2016 15:28:31:  Epoch[16 of 25]-Minibatch[ 751- 760, 69.25%]: CE.SM = 0.91535034 * 10240; Err = 0.29931641 * 10240; time = 0.0403s; samplesPerSecond = 254391.0
12/20/2016 15:28:31:  Epoch[16 of 25]-Minibatch[ 761- 770, 70.16%]: CE.SM = 0.90737915 * 10240; Err = 0.29355469 * 10240; time = 0.0408s; samplesPerSecond = 251085.0
12/20/2016 15:28:31:  Epoch[16 of 25]-Minibatch[ 771- 780, 71.07%]: CE.SM = 0.90374756 * 10240; Err = 0.30087891 * 10240; time = 0.0410s; samplesPerSecond = 249792.7
12/20/2016 15:28:31:  Epoch[16 of 25]-Minibatch[ 781- 790, 71.98%]: CE.SM = 0.91340332 * 10240; Err = 0.29677734 * 10240; time = 0.0414s; samplesPerSecond = 247313.1
12/20/2016 15:28:31:  Epoch[16 of 25]-Minibatch[ 791- 800, 72.89%]: CE.SM = 0.90559692 * 10240; Err = 0.29541016 * 10240; time = 0.0442s; samplesPerSecond = 231689.9
12/20/2016 15:28:31:  Epoch[16 of 25]-Minibatch[ 801- 810, 73.80%]: CE.SM = 0.90640259 * 10240; Err = 0.29423828 * 10240; time = 0.0402s; samplesPerSecond = 254454.2
12/20/2016 15:28:31:  Epoch[16 of 25]-Minibatch[ 811- 820, 74.72%]: CE.SM = 0.89691772 * 10240; Err = 0.29482422 * 10240; time = 0.0404s; samplesPerSecond = 253685.1
12/20/2016 15:28:31:  Epoch[16 of 25]-Minibatch[ 821- 830, 75.63%]: CE.SM = 0.90128174 * 10240; Err = 0.29804687 * 10240; time = 0.0413s; samplesPerSecond = 248074.0
12/20/2016 15:28:31:  Epoch[16 of 25]-Minibatch[ 831- 840, 76.54%]: CE.SM = 0.89745483 * 10240; Err = 0.29462891 * 10240; time = 0.0409s; samplesPerSecond = 250605.7
12/20/2016 15:28:31:  Epoch[16 of 25]-Minibatch[ 841- 850, 77.45%]: CE.SM = 0.91715698 * 10240; Err = 0.30058594 * 10240; time = 0.0390s; samplesPerSecond = 262328.7
12/20/2016 15:28:31:  Epoch[16 of 25]-Minibatch[ 851- 860, 78.36%]: CE.SM = 0.90385132 * 10240; Err = 0.29101562 * 10240; time = 0.0397s; samplesPerSecond = 257694.3
12/20/2016 15:28:31:  Epoch[16 of 25]-Minibatch[ 861- 870, 79.27%]: CE.SM = 0.88482666 * 10240; Err = 0.28847656 * 10240; time = 0.0404s; samplesPerSecond = 253760.6
12/20/2016 15:28:31:  Epoch[16 of 25]-Minibatch[ 871- 880, 80.18%]: CE.SM = 0.89381104 * 10240; Err = 0.29541016 * 10240; time = 0.0414s; samplesPerSecond = 247140.0
12/20/2016 15:28:31:  Epoch[16 of 25]-Minibatch[ 881- 890, 81.09%]: CE.SM = 0.89484253 * 10240; Err = 0.28730469 * 10240; time = 0.0435s; samplesPerSecond = 235423.9
12/20/2016 15:28:31:  Epoch[16 of 25]-Minibatch[ 891- 900, 82.00%]: CE.SM = 0.89036255 * 10240; Err = 0.28994141 * 10240; time = 0.0403s; samplesPerSecond = 254018.7
12/20/2016 15:28:31:  Epoch[16 of 25]-Minibatch[ 901- 910, 82.92%]: CE.SM = 0.90881958 * 10240; Err = 0.28964844 * 10240; time = 0.0413s; samplesPerSecond = 247833.9
12/20/2016 15:28:31:  Epoch[16 of 25]-Minibatch[ 911- 920, 83.83%]: CE.SM = 0.90809937 * 10240; Err = 0.29570313 * 10240; time = 0.0418s; samplesPerSecond = 245146.2
12/20/2016 15:28:31:  Epoch[16 of 25]-Minibatch[ 921- 930, 84.74%]: CE.SM = 0.91411743 * 10240; Err = 0.29609375 * 10240; time = 0.0409s; samplesPerSecond = 250323.9
12/20/2016 15:28:32:  Epoch[16 of 25]-Minibatch[ 931- 940, 85.65%]: CE.SM = 0.90811768 * 10240; Err = 0.29453125 * 10240; time = 0.0402s; samplesPerSecond = 254764.4
12/20/2016 15:28:32:  Epoch[16 of 25]-Minibatch[ 941- 950, 86.56%]: CE.SM = 0.89965820 * 10240; Err = 0.29345703 * 10240; time = 0.0435s; samplesPerSecond = 235332.0
12/20/2016 15:28:32:  Epoch[16 of 25]-Minibatch[ 951- 960, 87.47%]: CE.SM = 0.88569946 * 10240; Err = 0.29306641 * 10240; time = 0.0424s; samplesPerSecond = 241412.6
12/20/2016 15:28:32:  Epoch[16 of 25]-Minibatch[ 961- 970, 88.38%]: CE.SM = 0.91292725 * 10240; Err = 0.30273438 * 10240; time = 0.0445s; samplesPerSecond = 229890.2
12/20/2016 15:28:32:  Epoch[16 of 25]-Minibatch[ 971- 980, 89.29%]: CE.SM = 0.90972900 * 10240; Err = 0.29453125 * 10240; time = 0.0443s; samplesPerSecond = 231005.2
12/20/2016 15:28:32:  Epoch[16 of 25]-Minibatch[ 981- 990, 90.21%]: CE.SM = 0.92211304 * 10240; Err = 0.30087891 * 10240; time = 0.0408s; samplesPerSecond = 251005.0
12/20/2016 15:28:32:  Epoch[16 of 25]-Minibatch[ 991-1000, 91.12%]: CE.SM = 0.89945679 * 10240; Err = 0.29658203 * 10240; time = 0.0416s; samplesPerSecond = 246029.6
12/20/2016 15:28:32:  Epoch[16 of 25]-Minibatch[1001-1010, 92.03%]: CE.SM = 0.90399170 * 10240; Err = 0.29091797 * 10240; time = 0.0483s; samplesPerSecond = 212219.2
12/20/2016 15:28:32:  Epoch[16 of 25]-Minibatch[1011-1020, 92.94%]: CE.SM = 0.90747681 * 10240; Err = 0.29794922 * 10240; time = 0.0400s; samplesPerSecond = 255955.2
12/20/2016 15:28:32:  Epoch[16 of 25]-Minibatch[1021-1030, 93.85%]: CE.SM = 0.90913086 * 10240; Err = 0.29921875 * 10240; time = 0.0474s; samplesPerSecond = 216198.0
12/20/2016 15:28:32:  Epoch[16 of 25]-Minibatch[1031-1040, 94.76%]: CE.SM = 0.89845581 * 10240; Err = 0.29638672 * 10240; time = 0.0460s; samplesPerSecond = 222681.3
12/20/2016 15:28:32:  Epoch[16 of 25]-Minibatch[1041-1050, 95.67%]: CE.SM = 0.91730347 * 10240; Err = 0.29833984 * 10240; time = 0.0415s; samplesPerSecond = 246556.9
12/20/2016 15:28:32:  Epoch[16 of 25]-Minibatch[1051-1060, 96.58%]: CE.SM = 0.89798584 * 10240; Err = 0.29521484 * 10240; time = 0.0438s; samplesPerSecond = 233635.3
12/20/2016 15:28:32:  Epoch[16 of 25]-Minibatch[1061-1070, 97.49%]: CE.SM = 0.91313477 * 10240; Err = 0.30185547 * 10240; time = 0.0404s; samplesPerSecond = 253339.9
12/20/2016 15:28:32:  Epoch[16 of 25]-Minibatch[1071-1080, 98.41%]: CE.SM = 0.89807739 * 10240; Err = 0.29130859 * 10240; time = 0.0403s; samplesPerSecond = 254031.3
12/20/2016 15:28:32:  Epoch[16 of 25]-Minibatch[1081-1090, 99.32%]: CE.SM = 0.89791870 * 10240; Err = 0.29199219 * 10240; time = 0.0402s; samplesPerSecond = 254872.2
12/20/2016 15:28:32: Finished Epoch[16 of 25]: [Training] CE.SM = 0.92622178 * 1124823; Err = 0.30220221 * 1124823; totalSamplesSeen = 17997168; learningRatePerSample = 7.8124998e-05; epochTime=4.72917s
12/20/2016 15:28:32: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn.16'

12/20/2016 15:28:32: Starting Epoch 17: learning rate per sample = 0.000078  effective momentum = 0.900000  momentum as time constant = 9719.0 samples

12/20/2016 15:28:32: Starting minibatch loop.
12/20/2016 15:28:33:  Epoch[17 of 25]-Minibatch[   1-  10, 0.91%]: CE.SM = 0.90412092 * 10240; Err = 0.29003906 * 10240; time = 0.0458s; samplesPerSecond = 223678.5
12/20/2016 15:28:33:  Epoch[17 of 25]-Minibatch[  11-  20, 1.82%]: CE.SM = 0.90112982 * 10240; Err = 0.29199219 * 10240; time = 0.0411s; samplesPerSecond = 249385.1
12/20/2016 15:28:33:  Epoch[17 of 25]-Minibatch[  21-  30, 2.73%]: CE.SM = 0.91675682 * 10240; Err = 0.30146484 * 10240; time = 0.0413s; samplesPerSecond = 247899.9
12/20/2016 15:28:33:  Epoch[17 of 25]-Minibatch[  31-  40, 3.64%]: CE.SM = 0.90920506 * 10240; Err = 0.29423828 * 10240; time = 0.0403s; samplesPerSecond = 254100.6
12/20/2016 15:28:33:  Epoch[17 of 25]-Minibatch[  41-  50, 4.56%]: CE.SM = 0.91325760 * 10240; Err = 0.29687500 * 10240; time = 0.0399s; samplesPerSecond = 256879.8
12/20/2016 15:28:33:  Epoch[17 of 25]-Minibatch[  51-  60, 5.47%]: CE.SM = 0.89388924 * 10240; Err = 0.29423828 * 10240; time = 0.0399s; samplesPerSecond = 256866.9
12/20/2016 15:28:33:  Epoch[17 of 25]-Minibatch[  61-  70, 6.38%]: CE.SM = 0.90582275 * 10240; Err = 0.29228516 * 10240; time = 0.0401s; samplesPerSecond = 255177.1
12/20/2016 15:28:33:  Epoch[17 of 25]-Minibatch[  71-  80, 7.29%]: CE.SM = 0.89727592 * 10240; Err = 0.29189453 * 10240; time = 0.0400s; samplesPerSecond = 256160.1
12/20/2016 15:28:33:  Epoch[17 of 25]-Minibatch[  81-  90, 8.20%]: CE.SM = 0.90358810 * 10240; Err = 0.29599609 * 10240; time = 0.0456s; samplesPerSecond = 224448.2
12/20/2016 15:28:33:  Epoch[17 of 25]-Minibatch[  91- 100, 9.11%]: CE.SM = 0.88601685 * 10240; Err = 0.29277344 * 10240; time = 0.0417s; samplesPerSecond = 245622.5
12/20/2016 15:28:33:  Epoch[17 of 25]-Minibatch[ 101- 110, 10.02%]: CE.SM = 0.89641113 * 10240; Err = 0.28974609 * 10240; time = 0.0416s; samplesPerSecond = 246000.1
12/20/2016 15:28:33:  Epoch[17 of 25]-Minibatch[ 111- 120, 10.93%]: CE.SM = 0.88794632 * 10240; Err = 0.29052734 * 10240; time = 0.0407s; samplesPerSecond = 251621.8
12/20/2016 15:28:33:  Epoch[17 of 25]-Minibatch[ 121- 130, 11.85%]: CE.SM = 0.89933777 * 10240; Err = 0.28691406 * 10240; time = 0.0410s; samplesPerSecond = 249701.3
12/20/2016 15:28:33:  Epoch[17 of 25]-Minibatch[ 131- 140, 12.76%]: CE.SM = 0.88847351 * 10240; Err = 0.29384766 * 10240; time = 0.0406s; samplesPerSecond = 252260.2
12/20/2016 15:28:33:  Epoch[17 of 25]-Minibatch[ 141- 150, 13.67%]: CE.SM = 0.93061066 * 10240; Err = 0.30097656 * 10240; time = 0.0402s; samplesPerSecond = 254460.5
12/20/2016 15:28:33:  Epoch[17 of 25]-Minibatch[ 151- 160, 14.58%]: CE.SM = 0.89831238 * 10240; Err = 0.29492188 * 10240; time = 0.0415s; samplesPerSecond = 246824.3
12/20/2016 15:28:33:  Epoch[17 of 25]-Minibatch[ 161- 170, 15.49%]: CE.SM = 0.89291382 * 10240; Err = 0.29121094 * 10240; time = 0.0403s; samplesPerSecond = 254025.0
12/20/2016 15:28:33:  Epoch[17 of 25]-Minibatch[ 171- 180, 16.40%]: CE.SM = 0.92010651 * 10240; Err = 0.30527344 * 10240; time = 0.0409s; samplesPerSecond = 250562.8
12/20/2016 15:28:33:  Epoch[17 of 25]-Minibatch[ 181- 190, 17.31%]: CE.SM = 0.90031891 * 10240; Err = 0.29023437 * 10240; time = 0.0402s; samplesPerSecond = 255037.2
12/20/2016 15:28:33:  Epoch[17 of 25]-Minibatch[ 191- 200, 18.22%]: CE.SM = 0.89866486 * 10240; Err = 0.29628906 * 10240; time = 0.0407s; samplesPerSecond = 251362.4
12/20/2016 15:28:33:  Epoch[17 of 25]-Minibatch[ 201- 210, 19.13%]: CE.SM = 0.89811096 * 10240; Err = 0.29238281 * 10240; time = 0.0412s; samplesPerSecond = 248278.5
12/20/2016 15:28:33:  Epoch[17 of 25]-Minibatch[ 211- 220, 20.05%]: CE.SM = 0.89045715 * 10240; Err = 0.28935547 * 10240; time = 0.0402s; samplesPerSecond = 254916.6
12/20/2016 15:28:33:  Epoch[17 of 25]-Minibatch[ 221- 230, 20.96%]: CE.SM = 0.91043243 * 10240; Err = 0.29199219 * 10240; time = 0.0405s; samplesPerSecond = 253058.2
12/20/2016 15:28:33:  Epoch[17 of 25]-Minibatch[ 231- 240, 21.87%]: CE.SM = 0.91732178 * 10240; Err = 0.29755859 * 10240; time = 0.0397s; samplesPerSecond = 257785.2
12/20/2016 15:28:33:  Epoch[17 of 25]-Minibatch[ 241- 250, 22.78%]: CE.SM = 0.89291077 * 10240; Err = 0.29531250 * 10240; time = 0.0402s; samplesPerSecond = 254992.8
12/20/2016 15:28:34:  Epoch[17 of 25]-Minibatch[ 251- 260, 23.69%]: CE.SM = 0.90088501 * 10240; Err = 0.30205078 * 10240; time = 0.0405s; samplesPerSecond = 252939.4
12/20/2016 15:28:34:  Epoch[17 of 25]-Minibatch[ 261- 270, 24.60%]: CE.SM = 0.88849182 * 10240; Err = 0.28876953 * 10240; time = 0.0405s; samplesPerSecond = 252702.2
12/20/2016 15:28:34:  Epoch[17 of 25]-Minibatch[ 271- 280, 25.51%]: CE.SM = 0.91011200 * 10240; Err = 0.29726562 * 10240; time = 0.0403s; samplesPerSecond = 254296.2
12/20/2016 15:28:34:  Epoch[17 of 25]-Minibatch[ 281- 290, 26.42%]: CE.SM = 0.91981812 * 10240; Err = 0.29580078 * 10240; time = 0.0408s; samplesPerSecond = 251023.5
12/20/2016 15:28:34:  Epoch[17 of 25]-Minibatch[ 291- 300, 27.33%]: CE.SM = 0.89315796 * 10240; Err = 0.28271484 * 10240; time = 0.0401s; samplesPerSecond = 255654.9
12/20/2016 15:28:34:  Epoch[17 of 25]-Minibatch[ 301- 310, 28.25%]: CE.SM = 0.90255737 * 10240; Err = 0.30039063 * 10240; time = 0.0406s; samplesPerSecond = 252123.6
12/20/2016 15:28:34:  Epoch[17 of 25]-Minibatch[ 311- 320, 29.16%]: CE.SM = 0.88550720 * 10240; Err = 0.29072266 * 10240; time = 0.0410s; samplesPerSecond = 249652.6
12/20/2016 15:28:34:  Epoch[17 of 25]-Minibatch[ 321- 330, 30.07%]: CE.SM = 0.89993591 * 10240; Err = 0.29394531 * 10240; time = 0.0405s; samplesPerSecond = 252720.9
12/20/2016 15:28:34:  Epoch[17 of 25]-Minibatch[ 331- 340, 30.98%]: CE.SM = 0.88606873 * 10240; Err = 0.29013672 * 10240; time = 0.0400s; samplesPerSecond = 256000.0
12/20/2016 15:28:34:  Epoch[17 of 25]-Minibatch[ 341- 350, 31.89%]: CE.SM = 0.90436096 * 10240; Err = 0.28837891 * 10240; time = 0.0408s; samplesPerSecond = 250697.7
12/20/2016 15:28:34:  Epoch[17 of 25]-Minibatch[ 351- 360, 32.80%]: CE.SM = 0.88900146 * 10240; Err = 0.29345703 * 10240; time = 0.0405s; samplesPerSecond = 252534.0
12/20/2016 15:28:34:  Epoch[17 of 25]-Minibatch[ 361- 370, 33.71%]: CE.SM = 0.89709167 * 10240; Err = 0.28603516 * 10240; time = 0.0411s; samplesPerSecond = 249203.0
12/20/2016 15:28:34:  Epoch[17 of 25]-Minibatch[ 371- 380, 34.62%]: CE.SM = 0.89120483 * 10240; Err = 0.29013672 * 10240; time = 0.0403s; samplesPerSecond = 254113.2
12/20/2016 15:28:34:  Epoch[17 of 25]-Minibatch[ 381- 390, 35.54%]: CE.SM = 0.90176392 * 10240; Err = 0.29238281 * 10240; time = 0.0398s; samplesPerSecond = 257499.9
12/20/2016 15:28:34:  Epoch[17 of 25]-Minibatch[ 391- 400, 36.45%]: CE.SM = 0.89359741 * 10240; Err = 0.29140625 * 10240; time = 0.0402s; samplesPerSecond = 254618.7
12/20/2016 15:28:34:  Epoch[17 of 25]-Minibatch[ 401- 410, 37.36%]: CE.SM = 0.90227051 * 10240; Err = 0.29814453 * 10240; time = 0.0407s; samplesPerSecond = 251770.3
12/20/2016 15:28:34:  Epoch[17 of 25]-Minibatch[ 411- 420, 38.27%]: CE.SM = 0.88808594 * 10240; Err = 0.29062500 * 10240; time = 0.0413s; samplesPerSecond = 247702.0
12/20/2016 15:28:34:  Epoch[17 of 25]-Minibatch[ 421- 430, 39.18%]: CE.SM = 0.90125732 * 10240; Err = 0.29003906 * 10240; time = 0.0406s; samplesPerSecond = 252148.4
12/20/2016 15:28:34:  Epoch[17 of 25]-Minibatch[ 431- 440, 40.09%]: CE.SM = 0.88392944 * 10240; Err = 0.28613281 * 10240; time = 0.0400s; samplesPerSecond = 255948.8
12/20/2016 15:28:34:  Epoch[17 of 25]-Minibatch[ 441- 450, 41.00%]: CE.SM = 0.90635376 * 10240; Err = 0.29746094 * 10240; time = 0.0395s; samplesPerSecond = 259030.7
12/20/2016 15:28:34:  Epoch[17 of 25]-Minibatch[ 451- 460, 41.91%]: CE.SM = 0.90874329 * 10240; Err = 0.29257813 * 10240; time = 0.0401s; samplesPerSecond = 255132.5
12/20/2016 15:28:34:  Epoch[17 of 25]-Minibatch[ 461- 470, 42.82%]: CE.SM = 0.88788757 * 10240; Err = 0.28906250 * 10240; time = 0.0414s; samplesPerSecond = 247624.1
12/20/2016 15:28:34:  Epoch[17 of 25]-Minibatch[ 471- 480, 43.74%]: CE.SM = 0.87886353 * 10240; Err = 0.28476563 * 10240; time = 0.0399s; samplesPerSecond = 256545.2
12/20/2016 15:28:34:  Epoch[17 of 25]-Minibatch[ 481- 490, 44.65%]: CE.SM = 0.88734436 * 10240; Err = 0.29111328 * 10240; time = 0.0400s; samplesPerSecond = 256166.5
12/20/2016 15:28:35:  Epoch[17 of 25]-Minibatch[ 491- 500, 45.56%]: CE.SM = 0.88006287 * 10240; Err = 0.28681641 * 10240; time = 0.0400s; samplesPerSecond = 256160.1
12/20/2016 15:28:35:  Epoch[17 of 25]-Minibatch[ 501- 510, 46.47%]: CE.SM = 0.89539490 * 10240; Err = 0.29033203 * 10240; time = 0.0404s; samplesPerSecond = 253258.5
12/20/2016 15:28:35:  Epoch[17 of 25]-Minibatch[ 511- 520, 47.38%]: CE.SM = 0.89907837 * 10240; Err = 0.29814453 * 10240; time = 0.0404s; samplesPerSecond = 253383.8
12/20/2016 15:28:35:  Epoch[17 of 25]-Minibatch[ 521- 530, 48.29%]: CE.SM = 0.90302734 * 10240; Err = 0.29775391 * 10240; time = 0.0397s; samplesPerSecond = 257623.0
12/20/2016 15:28:35:  Epoch[17 of 25]-Minibatch[ 531- 540, 49.20%]: CE.SM = 0.88634949 * 10240; Err = 0.29228516 * 10240; time = 0.0396s; samplesPerSecond = 258435.8
12/20/2016 15:28:35:  Epoch[17 of 25]-Minibatch[ 541- 550, 50.11%]: CE.SM = 0.90551758 * 10240; Err = 0.29697266 * 10240; time = 0.0423s; samplesPerSecond = 242177.7
12/20/2016 15:28:35:  Epoch[17 of 25]-Minibatch[ 551- 560, 51.03%]: CE.SM = 0.90307922 * 10240; Err = 0.29648438 * 10240; time = 0.0398s; samplesPerSecond = 257357.6
12/20/2016 15:28:35:  Epoch[17 of 25]-Minibatch[ 561- 570, 51.94%]: CE.SM = 0.89735718 * 10240; Err = 0.29472656 * 10240; time = 0.0401s; samplesPerSecond = 255323.4
12/20/2016 15:28:35:  Epoch[17 of 25]-Minibatch[ 571- 580, 52.85%]: CE.SM = 0.88468018 * 10240; Err = 0.29072266 * 10240; time = 0.0400s; samplesPerSecond = 255705.9
12/20/2016 15:28:35:  Epoch[17 of 25]-Minibatch[ 581- 590, 53.76%]: CE.SM = 0.89934082 * 10240; Err = 0.29365234 * 10240; time = 0.0398s; samplesPerSecond = 257002.3
12/20/2016 15:28:35:  Epoch[17 of 25]-Minibatch[ 591- 600, 54.67%]: CE.SM = 0.89109497 * 10240; Err = 0.29296875 * 10240; time = 0.0402s; samplesPerSecond = 254846.8
12/20/2016 15:28:35:  Epoch[17 of 25]-Minibatch[ 601- 610, 55.58%]: CE.SM = 0.91264648 * 10240; Err = 0.29824219 * 10240; time = 0.0398s; samplesPerSecond = 257137.8
12/20/2016 15:28:35:  Epoch[17 of 25]-Minibatch[ 611- 620, 56.49%]: CE.SM = 0.89247437 * 10240; Err = 0.28964844 * 10240; time = 0.0415s; samplesPerSecond = 246949.3
12/20/2016 15:28:35:  Epoch[17 of 25]-Minibatch[ 621- 630, 57.40%]: CE.SM = 0.88226318 * 10240; Err = 0.29052734 * 10240; time = 0.0404s; samplesPerSecond = 253641.1
12/20/2016 15:28:35:  Epoch[17 of 25]-Minibatch[ 631- 640, 58.31%]: CE.SM = 0.89592896 * 10240; Err = 0.29462891 * 10240; time = 0.0400s; samplesPerSecond = 256115.3
12/20/2016 15:28:35:  Epoch[17 of 25]-Minibatch[ 641- 650, 59.23%]: CE.SM = 0.90838623 * 10240; Err = 0.29160156 * 10240; time = 0.0400s; samplesPerSecond = 255910.4
12/20/2016 15:28:35:  Epoch[17 of 25]-Minibatch[ 651- 660, 60.14%]: CE.SM = 0.88621826 * 10240; Err = 0.29384766 * 10240; time = 0.0399s; samplesPerSecond = 256513.0
12/20/2016 15:28:35:  Epoch[17 of 25]-Minibatch[ 661- 670, 61.05%]: CE.SM = 0.89883423 * 10240; Err = 0.29091797 * 10240; time = 0.0400s; samplesPerSecond = 256051.2
12/20/2016 15:28:35:  Epoch[17 of 25]-Minibatch[ 671- 680, 61.96%]: CE.SM = 0.89150391 * 10240; Err = 0.29130859 * 10240; time = 0.0402s; samplesPerSecond = 254669.4
12/20/2016 15:28:35:  Epoch[17 of 25]-Minibatch[ 681- 690, 62.87%]: CE.SM = 0.88930054 * 10240; Err = 0.28447266 * 10240; time = 0.0405s; samplesPerSecond = 253102.0
12/20/2016 15:28:35:  Epoch[17 of 25]-Minibatch[ 691- 700, 63.78%]: CE.SM = 0.91492920 * 10240; Err = 0.29160156 * 10240; time = 0.0399s; samplesPerSecond = 256493.8
12/20/2016 15:28:35:  Epoch[17 of 25]-Minibatch[ 701- 710, 64.69%]: CE.SM = 0.90905762 * 10240; Err = 0.29570313 * 10240; time = 0.0414s; samplesPerSecond = 247110.2
12/20/2016 15:28:35:  Epoch[17 of 25]-Minibatch[ 711- 720, 65.60%]: CE.SM = 0.87902222 * 10240; Err = 0.28427734 * 10240; time = 0.0403s; samplesPerSecond = 254403.6
12/20/2016 15:28:35:  Epoch[17 of 25]-Minibatch[ 721- 730, 66.51%]: CE.SM = 0.90328369 * 10240; Err = 0.29785156 * 10240; time = 0.0404s; samplesPerSecond = 253289.8
12/20/2016 15:28:35:  Epoch[17 of 25]-Minibatch[ 731- 740, 67.43%]: CE.SM = 0.87766113 * 10240; Err = 0.28486328 * 10240; time = 0.0406s; samplesPerSecond = 251943.7
12/20/2016 15:28:36:  Epoch[17 of 25]-Minibatch[ 741- 750, 68.34%]: CE.SM = 0.87254639 * 10240; Err = 0.28310547 * 10240; time = 0.0400s; samplesPerSecond = 255846.5
12/20/2016 15:28:36:  Epoch[17 of 25]-Minibatch[ 751- 760, 69.25%]: CE.SM = 0.89284058 * 10240; Err = 0.29189453 * 10240; time = 0.0399s; samplesPerSecond = 256603.0
12/20/2016 15:28:36:  Epoch[17 of 25]-Minibatch[ 761- 770, 70.16%]: CE.SM = 0.89855347 * 10240; Err = 0.29023437 * 10240; time = 0.0399s; samplesPerSecond = 256474.5
12/20/2016 15:28:36:  Epoch[17 of 25]-Minibatch[ 771- 780, 71.07%]: CE.SM = 0.89534912 * 10240; Err = 0.28710938 * 10240; time = 0.0429s; samplesPerSecond = 238655.7
12/20/2016 15:28:36:  Epoch[17 of 25]-Minibatch[ 781- 790, 71.98%]: CE.SM = 0.89276123 * 10240; Err = 0.29199219 * 10240; time = 0.0400s; samplesPerSecond = 255872.1
12/20/2016 15:28:36:  Epoch[17 of 25]-Minibatch[ 791- 800, 72.89%]: CE.SM = 0.87394409 * 10240; Err = 0.28710938 * 10240; time = 0.0415s; samplesPerSecond = 246604.4
12/20/2016 15:28:36:  Epoch[17 of 25]-Minibatch[ 801- 810, 73.80%]: CE.SM = 0.89252930 * 10240; Err = 0.29355469 * 10240; time = 0.0402s; samplesPerSecond = 254770.7
12/20/2016 15:28:36:  Epoch[17 of 25]-Minibatch[ 811- 820, 74.72%]: CE.SM = 0.89291992 * 10240; Err = 0.29433594 * 10240; time = 0.0400s; samplesPerSecond = 256089.6
12/20/2016 15:28:36:  Epoch[17 of 25]-Minibatch[ 821- 830, 75.63%]: CE.SM = 0.89168701 * 10240; Err = 0.28867188 * 10240; time = 0.0399s; samplesPerSecond = 256744.6
12/20/2016 15:28:36:  Epoch[17 of 25]-Minibatch[ 831- 840, 76.54%]: CE.SM = 0.89912109 * 10240; Err = 0.29130859 * 10240; time = 0.0405s; samplesPerSecond = 253083.2
12/20/2016 15:28:36:  Epoch[17 of 25]-Minibatch[ 841- 850, 77.45%]: CE.SM = 0.90093384 * 10240; Err = 0.29638672 * 10240; time = 0.0409s; samplesPerSecond = 250366.7
12/20/2016 15:28:36:  Epoch[17 of 25]-Minibatch[ 851- 860, 78.36%]: CE.SM = 0.91223145 * 10240; Err = 0.30000000 * 10240; time = 0.0404s; samplesPerSecond = 253446.5
12/20/2016 15:28:36:  Epoch[17 of 25]-Minibatch[ 861- 870, 79.27%]: CE.SM = 0.92183228 * 10240; Err = 0.30097656 * 10240; time = 0.0395s; samplesPerSecond = 259411.3
12/20/2016 15:28:36:  Epoch[17 of 25]-Minibatch[ 871- 880, 80.18%]: CE.SM = 0.89049072 * 10240; Err = 0.28925781 * 10240; time = 0.0395s; samplesPerSecond = 259325.9
12/20/2016 15:28:36:  Epoch[17 of 25]-Minibatch[ 881- 890, 81.09%]: CE.SM = 0.89038696 * 10240; Err = 0.29296875 * 10240; time = 0.0422s; samplesPerSecond = 242907.3
12/20/2016 15:28:36:  Epoch[17 of 25]-Minibatch[ 891- 900, 82.00%]: CE.SM = 0.90272827 * 10240; Err = 0.29355469 * 10240; time = 0.0398s; samplesPerSecond = 257480.5
12/20/2016 15:28:36:  Epoch[17 of 25]-Minibatch[ 901- 910, 82.92%]: CE.SM = 0.88849487 * 10240; Err = 0.29501953 * 10240; time = 0.0395s; samplesPerSecond = 259220.8
12/20/2016 15:28:36:  Epoch[17 of 25]-Minibatch[ 911- 920, 83.83%]: CE.SM = 0.91412964 * 10240; Err = 0.30126953 * 10240; time = 0.0396s; samplesPerSecond = 258664.2
12/20/2016 15:28:36:  Epoch[17 of 25]-Minibatch[ 921- 930, 84.74%]: CE.SM = 0.87891235 * 10240; Err = 0.29101562 * 10240; time = 0.0398s; samplesPerSecond = 257351.1
12/20/2016 15:28:36:  Epoch[17 of 25]-Minibatch[ 931- 940, 85.65%]: CE.SM = 0.90983276 * 10240; Err = 0.29501953 * 10240; time = 0.0406s; samplesPerSecond = 252428.1
12/20/2016 15:28:36:  Epoch[17 of 25]-Minibatch[ 941- 950, 86.56%]: CE.SM = 0.90461426 * 10240; Err = 0.29492188 * 10240; time = 0.0403s; samplesPerSecond = 253779.4
12/20/2016 15:28:36:  Epoch[17 of 25]-Minibatch[ 951- 960, 87.47%]: CE.SM = 0.90034180 * 10240; Err = 0.29794922 * 10240; time = 0.0404s; samplesPerSecond = 253189.6
12/20/2016 15:28:36:  Epoch[17 of 25]-Minibatch[ 961- 970, 88.38%]: CE.SM = 0.87952881 * 10240; Err = 0.29179688 * 10240; time = 0.0414s; samplesPerSecond = 247331.0
12/20/2016 15:28:36:  Epoch[17 of 25]-Minibatch[ 971- 980, 89.29%]: CE.SM = 0.87628174 * 10240; Err = 0.28818359 * 10240; time = 0.0393s; samplesPerSecond = 260407.4
12/20/2016 15:28:36:  Epoch[17 of 25]-Minibatch[ 981- 990, 90.21%]: CE.SM = 0.90747681 * 10240; Err = 0.29990234 * 10240; time = 0.0396s; samplesPerSecond = 258893.1
12/20/2016 15:28:37:  Epoch[17 of 25]-Minibatch[ 991-1000, 91.12%]: CE.SM = 0.88422241 * 10240; Err = 0.29199219 * 10240; time = 0.0397s; samplesPerSecond = 257941.0
12/20/2016 15:28:37:  Epoch[17 of 25]-Minibatch[1001-1010, 92.03%]: CE.SM = 0.89851074 * 10240; Err = 0.29199219 * 10240; time = 0.0404s; samplesPerSecond = 253622.3
12/20/2016 15:28:37:  Epoch[17 of 25]-Minibatch[1011-1020, 92.94%]: CE.SM = 0.90225830 * 10240; Err = 0.29316406 * 10240; time = 0.0408s; samplesPerSecond = 250759.1
12/20/2016 15:28:37:  Epoch[17 of 25]-Minibatch[1021-1030, 93.85%]: CE.SM = 0.88027954 * 10240; Err = 0.28359375 * 10240; time = 0.0411s; samplesPerSecond = 249312.2
12/20/2016 15:28:37:  Epoch[17 of 25]-Minibatch[1031-1040, 94.76%]: CE.SM = 0.88977051 * 10240; Err = 0.29052734 * 10240; time = 0.0404s; samplesPerSecond = 253484.2
12/20/2016 15:28:37:  Epoch[17 of 25]-Minibatch[1041-1050, 95.67%]: CE.SM = 0.91628418 * 10240; Err = 0.29814453 * 10240; time = 0.0400s; samplesPerSecond = 256076.8
12/20/2016 15:28:37:  Epoch[17 of 25]-Minibatch[1051-1060, 96.58%]: CE.SM = 0.89398804 * 10240; Err = 0.29130859 * 10240; time = 0.0400s; samplesPerSecond = 256320.4
12/20/2016 15:28:37:  Epoch[17 of 25]-Minibatch[1061-1070, 97.49%]: CE.SM = 0.89567871 * 10240; Err = 0.28857422 * 10240; time = 0.0408s; samplesPerSecond = 250912.7
12/20/2016 15:28:37:  Epoch[17 of 25]-Minibatch[1071-1080, 98.41%]: CE.SM = 0.88997803 * 10240; Err = 0.29375000 * 10240; time = 0.0400s; samplesPerSecond = 255974.4
12/20/2016 15:28:37:  Epoch[17 of 25]-Minibatch[1081-1090, 99.32%]: CE.SM = 0.89987183 * 10240; Err = 0.29267578 * 10240; time = 0.0399s; samplesPerSecond = 256660.9
12/20/2016 15:28:37: Finished Epoch[17 of 25]: [Training] CE.SM = 0.89700780 * 1124823; Err = 0.29266738 * 1124823; totalSamplesSeen = 19121991; learningRatePerSample = 7.8124998e-05; epochTime=4.63891s
12/20/2016 15:28:37: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn.17'

12/20/2016 15:28:37: Starting Epoch 18: learning rate per sample = 0.000078  effective momentum = 0.900000  momentum as time constant = 9719.0 samples

12/20/2016 15:28:37: Starting minibatch loop.
12/20/2016 15:28:37:  Epoch[18 of 25]-Minibatch[   1-  10, 0.91%]: CE.SM = 0.89383736 * 10240; Err = 0.29189453 * 10240; time = 0.0450s; samplesPerSecond = 227343.4
12/20/2016 15:28:37:  Epoch[18 of 25]-Minibatch[  11-  20, 1.82%]: CE.SM = 0.88127975 * 10240; Err = 0.28896484 * 10240; time = 0.0408s; samplesPerSecond = 251195.9
12/20/2016 15:28:37:  Epoch[18 of 25]-Minibatch[  21-  30, 2.73%]: CE.SM = 0.88243351 * 10240; Err = 0.28808594 * 10240; time = 0.0408s; samplesPerSecond = 250882.0
12/20/2016 15:28:37:  Epoch[18 of 25]-Minibatch[  31-  40, 3.64%]: CE.SM = 0.89826603 * 10240; Err = 0.28847656 * 10240; time = 0.0404s; samplesPerSecond = 253302.3
12/20/2016 15:28:37:  Epoch[18 of 25]-Minibatch[  41-  50, 4.56%]: CE.SM = 0.89412651 * 10240; Err = 0.28808594 * 10240; time = 0.0403s; samplesPerSecond = 253962.0
12/20/2016 15:28:37:  Epoch[18 of 25]-Minibatch[  51-  60, 5.47%]: CE.SM = 0.85543976 * 10240; Err = 0.28369141 * 10240; time = 0.0401s; samplesPerSecond = 255285.2
12/20/2016 15:28:37:  Epoch[18 of 25]-Minibatch[  61-  70, 6.38%]: CE.SM = 0.87813377 * 10240; Err = 0.28720703 * 10240; time = 0.0406s; samplesPerSecond = 252210.5
12/20/2016 15:28:38:  Epoch[18 of 25]-Minibatch[  71-  80, 7.29%]: CE.SM = 0.89251633 * 10240; Err = 0.29267578 * 10240; time = 0.0404s; samplesPerSecond = 253402.6
12/20/2016 15:28:38:  Epoch[18 of 25]-Minibatch[  81-  90, 8.20%]: CE.SM = 0.89709320 * 10240; Err = 0.29433594 * 10240; time = 0.0400s; samplesPerSecond = 256096.0
12/20/2016 15:28:38:  Epoch[18 of 25]-Minibatch[  91- 100, 9.11%]: CE.SM = 0.89258194 * 10240; Err = 0.28886719 * 10240; time = 0.0403s; samplesPerSecond = 253899.0
12/20/2016 15:28:38:  Epoch[18 of 25]-Minibatch[ 101- 110, 10.02%]: CE.SM = 0.91410446 * 10240; Err = 0.29902344 * 10240; time = 0.0400s; samplesPerSecond = 256076.8
12/20/2016 15:28:38:  Epoch[18 of 25]-Minibatch[ 111- 120, 10.93%]: CE.SM = 0.87619324 * 10240; Err = 0.29140625 * 10240; time = 0.0409s; samplesPerSecond = 250464.7
12/20/2016 15:28:38:  Epoch[18 of 25]-Minibatch[ 121- 130, 11.85%]: CE.SM = 0.89623337 * 10240; Err = 0.29150391 * 10240; time = 0.0429s; samplesPerSecond = 238928.6
12/20/2016 15:28:38:  Epoch[18 of 25]-Minibatch[ 131- 140, 12.76%]: CE.SM = 0.90151215 * 10240; Err = 0.29208984 * 10240; time = 0.0407s; samplesPerSecond = 251411.7
12/20/2016 15:28:38:  Epoch[18 of 25]-Minibatch[ 141- 150, 13.67%]: CE.SM = 0.88769455 * 10240; Err = 0.29443359 * 10240; time = 0.0409s; samplesPerSecond = 250177.1
12/20/2016 15:28:38:  Epoch[18 of 25]-Minibatch[ 151- 160, 14.58%]: CE.SM = 0.87801361 * 10240; Err = 0.28662109 * 10240; time = 0.0437s; samplesPerSecond = 234357.1
12/20/2016 15:28:38:  Epoch[18 of 25]-Minibatch[ 161- 170, 15.49%]: CE.SM = 0.88697662 * 10240; Err = 0.29560547 * 10240; time = 0.0411s; samplesPerSecond = 248869.9
12/20/2016 15:28:38:  Epoch[18 of 25]-Minibatch[ 171- 180, 16.40%]: CE.SM = 0.88288422 * 10240; Err = 0.28886719 * 10240; time = 0.0410s; samplesPerSecond = 249488.4
12/20/2016 15:28:38:  Epoch[18 of 25]-Minibatch[ 181- 190, 17.31%]: CE.SM = 0.90088959 * 10240; Err = 0.29306641 * 10240; time = 0.0404s; samplesPerSecond = 253704.0
12/20/2016 15:28:38:  Epoch[18 of 25]-Minibatch[ 191- 200, 18.22%]: CE.SM = 0.86827087 * 10240; Err = 0.28574219 * 10240; time = 0.0405s; samplesPerSecond = 252696.0
12/20/2016 15:28:38:  Epoch[18 of 25]-Minibatch[ 201- 210, 19.13%]: CE.SM = 0.87580414 * 10240; Err = 0.28583984 * 10240; time = 0.0442s; samplesPerSecond = 231412.4
12/20/2016 15:28:38:  Epoch[18 of 25]-Minibatch[ 211- 220, 20.05%]: CE.SM = 0.89096527 * 10240; Err = 0.28857422 * 10240; time = 0.0442s; samplesPerSecond = 231658.5
12/20/2016 15:28:38:  Epoch[18 of 25]-Minibatch[ 221- 230, 20.96%]: CE.SM = 0.88234558 * 10240; Err = 0.28349609 * 10240; time = 0.0472s; samplesPerSecond = 216880.2
12/20/2016 15:28:38:  Epoch[18 of 25]-Minibatch[ 231- 240, 21.87%]: CE.SM = 0.87076569 * 10240; Err = 0.28583984 * 10240; time = 0.0429s; samplesPerSecond = 238694.6
12/20/2016 15:28:38:  Epoch[18 of 25]-Minibatch[ 241- 250, 22.78%]: CE.SM = 0.91783295 * 10240; Err = 0.30380859 * 10240; time = 0.0496s; samplesPerSecond = 206430.8
12/20/2016 15:28:38:  Epoch[18 of 25]-Minibatch[ 251- 260, 23.69%]: CE.SM = 0.87869263 * 10240; Err = 0.28789063 * 10240; time = 0.0412s; samplesPerSecond = 248380.9
12/20/2016 15:28:38:  Epoch[18 of 25]-Minibatch[ 261- 270, 24.60%]: CE.SM = 0.91579285 * 10240; Err = 0.29667969 * 10240; time = 0.0407s; samplesPerSecond = 251356.2
12/20/2016 15:28:38:  Epoch[18 of 25]-Minibatch[ 271- 280, 25.51%]: CE.SM = 0.88540802 * 10240; Err = 0.28896484 * 10240; time = 0.0402s; samplesPerSecond = 254587.0
12/20/2016 15:28:38:  Epoch[18 of 25]-Minibatch[ 281- 290, 26.42%]: CE.SM = 0.88533783 * 10240; Err = 0.28583984 * 10240; time = 0.0406s; samplesPerSecond = 252235.4
12/20/2016 15:28:38:  Epoch[18 of 25]-Minibatch[ 291- 300, 27.33%]: CE.SM = 0.89354248 * 10240; Err = 0.29492188 * 10240; time = 0.0412s; samplesPerSecond = 248519.6
12/20/2016 15:28:38:  Epoch[18 of 25]-Minibatch[ 301- 310, 28.25%]: CE.SM = 0.89325562 * 10240; Err = 0.28642578 * 10240; time = 0.0405s; samplesPerSecond = 253102.0
12/20/2016 15:28:39:  Epoch[18 of 25]-Minibatch[ 311- 320, 29.16%]: CE.SM = 0.89050293 * 10240; Err = 0.29062500 * 10240; time = 0.0403s; samplesPerSecond = 253943.1
12/20/2016 15:28:39:  Epoch[18 of 25]-Minibatch[ 321- 330, 30.07%]: CE.SM = 0.90156555 * 10240; Err = 0.29091797 * 10240; time = 0.0408s; samplesPerSecond = 251269.8
12/20/2016 15:28:39:  Epoch[18 of 25]-Minibatch[ 331- 340, 30.98%]: CE.SM = 0.89214478 * 10240; Err = 0.29111328 * 10240; time = 0.0407s; samplesPerSecond = 251652.7
12/20/2016 15:28:39:  Epoch[18 of 25]-Minibatch[ 341- 350, 31.89%]: CE.SM = 0.89555359 * 10240; Err = 0.29140625 * 10240; time = 0.0428s; samplesPerSecond = 239280.3
12/20/2016 15:28:39:  Epoch[18 of 25]-Minibatch[ 351- 360, 32.80%]: CE.SM = 0.88199768 * 10240; Err = 0.29052734 * 10240; time = 0.0409s; samplesPerSecond = 250183.2
12/20/2016 15:28:39:  Epoch[18 of 25]-Minibatch[ 361- 370, 33.71%]: CE.SM = 0.90384827 * 10240; Err = 0.29482422 * 10240; time = 0.0414s; samplesPerSecond = 247426.7
12/20/2016 15:28:39:  Epoch[18 of 25]-Minibatch[ 371- 380, 34.62%]: CE.SM = 0.88686523 * 10240; Err = 0.28857422 * 10240; time = 0.0402s; samplesPerSecond = 254783.4
12/20/2016 15:28:39:  Epoch[18 of 25]-Minibatch[ 381- 390, 35.54%]: CE.SM = 0.89262085 * 10240; Err = 0.28789063 * 10240; time = 0.0451s; samplesPerSecond = 226980.5
12/20/2016 15:28:39:  Epoch[18 of 25]-Minibatch[ 391- 400, 36.45%]: CE.SM = 0.88292542 * 10240; Err = 0.28574219 * 10240; time = 0.0433s; samplesPerSecond = 236522.4
12/20/2016 15:28:39:  Epoch[18 of 25]-Minibatch[ 401- 410, 37.36%]: CE.SM = 0.88554687 * 10240; Err = 0.29218750 * 10240; time = 0.0436s; samplesPerSecond = 234883.9
12/20/2016 15:28:39:  Epoch[18 of 25]-Minibatch[ 411- 420, 38.27%]: CE.SM = 0.88801880 * 10240; Err = 0.28935547 * 10240; time = 0.0403s; samplesPerSecond = 254252.0
12/20/2016 15:28:39:  Epoch[18 of 25]-Minibatch[ 421- 430, 39.18%]: CE.SM = 0.91038818 * 10240; Err = 0.29599609 * 10240; time = 0.0405s; samplesPerSecond = 252758.4
12/20/2016 15:28:39:  Epoch[18 of 25]-Minibatch[ 431- 440, 40.09%]: CE.SM = 0.90216980 * 10240; Err = 0.28906250 * 10240; time = 0.0400s; samplesPerSecond = 256275.5
12/20/2016 15:28:39:  Epoch[18 of 25]-Minibatch[ 441- 450, 41.00%]: CE.SM = 0.90681763 * 10240; Err = 0.29599609 * 10240; time = 0.0402s; samplesPerSecond = 254625.0
12/20/2016 15:28:39:  Epoch[18 of 25]-Minibatch[ 451- 460, 41.91%]: CE.SM = 0.89718933 * 10240; Err = 0.29150391 * 10240; time = 0.0400s; samplesPerSecond = 255693.2
12/20/2016 15:28:39:  Epoch[18 of 25]-Minibatch[ 461- 470, 42.82%]: CE.SM = 0.86503906 * 10240; Err = 0.28203125 * 10240; time = 0.0407s; samplesPerSecond = 251881.7
12/20/2016 15:28:39:  Epoch[18 of 25]-Minibatch[ 471- 480, 43.74%]: CE.SM = 0.89010010 * 10240; Err = 0.29082031 * 10240; time = 0.0398s; samplesPerSecond = 256995.9
12/20/2016 15:28:39:  Epoch[18 of 25]-Minibatch[ 481- 490, 44.65%]: CE.SM = 0.89631042 * 10240; Err = 0.29042969 * 10240; time = 0.0397s; samplesPerSecond = 257954.0
12/20/2016 15:28:39:  Epoch[18 of 25]-Minibatch[ 491- 500, 45.56%]: CE.SM = 0.88273010 * 10240; Err = 0.29169922 * 10240; time = 0.0402s; samplesPerSecond = 254910.3
12/20/2016 15:28:39:  Epoch[18 of 25]-Minibatch[ 501- 510, 46.47%]: CE.SM = 0.90837708 * 10240; Err = 0.29521484 * 10240; time = 0.0398s; samplesPerSecond = 257124.9
12/20/2016 15:28:39:  Epoch[18 of 25]-Minibatch[ 511- 520, 47.38%]: CE.SM = 0.90657959 * 10240; Err = 0.29267578 * 10240; time = 0.0391s; samplesPerSecond = 261959.6
12/20/2016 15:28:39:  Epoch[18 of 25]-Minibatch[ 521- 530, 48.29%]: CE.SM = 0.89542542 * 10240; Err = 0.29326172 * 10240; time = 0.0399s; samplesPerSecond = 256346.1
12/20/2016 15:28:39:  Epoch[18 of 25]-Minibatch[ 531- 540, 49.20%]: CE.SM = 0.89123230 * 10240; Err = 0.29130859 * 10240; time = 0.0399s; samplesPerSecond = 256435.9
12/20/2016 15:28:39:  Epoch[18 of 25]-Minibatch[ 541- 550, 50.11%]: CE.SM = 0.90615540 * 10240; Err = 0.29687500 * 10240; time = 0.0404s; samplesPerSecond = 253616.0
12/20/2016 15:28:40:  Epoch[18 of 25]-Minibatch[ 551- 560, 51.03%]: CE.SM = 0.87980957 * 10240; Err = 0.28554687 * 10240; time = 0.0402s; samplesPerSecond = 254720.0
12/20/2016 15:28:40:  Epoch[18 of 25]-Minibatch[ 561- 570, 51.94%]: CE.SM = 0.87937927 * 10240; Err = 0.27998047 * 10240; time = 0.0399s; samplesPerSecond = 256603.0
12/20/2016 15:28:40:  Epoch[18 of 25]-Minibatch[ 571- 580, 52.85%]: CE.SM = 0.90408936 * 10240; Err = 0.29560547 * 10240; time = 0.0404s; samplesPerSecond = 253183.3
12/20/2016 15:28:40:  Epoch[18 of 25]-Minibatch[ 581- 590, 53.76%]: CE.SM = 0.88983154 * 10240; Err = 0.28652344 * 10240; time = 0.0415s; samplesPerSecond = 246598.4
12/20/2016 15:28:40:  Epoch[18 of 25]-Minibatch[ 591- 600, 54.67%]: CE.SM = 0.90495605 * 10240; Err = 0.29736328 * 10240; time = 0.0406s; samplesPerSecond = 251937.5
12/20/2016 15:28:40:  Epoch[18 of 25]-Minibatch[ 601- 610, 55.58%]: CE.SM = 0.87313232 * 10240; Err = 0.28593750 * 10240; time = 0.0405s; samplesPerSecond = 252733.4
12/20/2016 15:28:40:  Epoch[18 of 25]-Minibatch[ 611- 620, 56.49%]: CE.SM = 0.89249878 * 10240; Err = 0.28798828 * 10240; time = 0.0407s; samplesPerSecond = 251584.7
12/20/2016 15:28:40:  Epoch[18 of 25]-Minibatch[ 621- 630, 57.40%]: CE.SM = 0.87913208 * 10240; Err = 0.28417969 * 10240; time = 0.0401s; samplesPerSecond = 255043.6
12/20/2016 15:28:40:  Epoch[18 of 25]-Minibatch[ 631- 640, 58.31%]: CE.SM = 0.92365112 * 10240; Err = 0.29570313 * 10240; time = 0.0404s; samplesPerSecond = 253754.3
12/20/2016 15:28:40:  Epoch[18 of 25]-Minibatch[ 641- 650, 59.23%]: CE.SM = 0.89142456 * 10240; Err = 0.29580078 * 10240; time = 0.0405s; samplesPerSecond = 252714.7
12/20/2016 15:28:40:  Epoch[18 of 25]-Minibatch[ 651- 660, 60.14%]: CE.SM = 0.88873291 * 10240; Err = 0.28857422 * 10240; time = 0.0404s; samplesPerSecond = 253546.9
12/20/2016 15:28:40:  Epoch[18 of 25]-Minibatch[ 661- 670, 61.05%]: CE.SM = 0.89405518 * 10240; Err = 0.29101562 * 10240; time = 0.0404s; samplesPerSecond = 253239.7
12/20/2016 15:28:40:  Epoch[18 of 25]-Minibatch[ 671- 680, 61.96%]: CE.SM = 0.89899902 * 10240; Err = 0.29101562 * 10240; time = 0.0412s; samplesPerSecond = 248646.3
12/20/2016 15:28:40:  Epoch[18 of 25]-Minibatch[ 681- 690, 62.87%]: CE.SM = 0.89829102 * 10240; Err = 0.29482422 * 10240; time = 0.0406s; samplesPerSecond = 252154.6
12/20/2016 15:28:40:  Epoch[18 of 25]-Minibatch[ 691- 700, 63.78%]: CE.SM = 0.88671265 * 10240; Err = 0.29169922 * 10240; time = 0.0400s; samplesPerSecond = 256089.6
12/20/2016 15:28:40:  Epoch[18 of 25]-Minibatch[ 701- 710, 64.69%]: CE.SM = 0.89524536 * 10240; Err = 0.28955078 * 10240; time = 0.0409s; samplesPerSecond = 250293.3
12/20/2016 15:28:40:  Epoch[18 of 25]-Minibatch[ 711- 720, 65.60%]: CE.SM = 0.90271606 * 10240; Err = 0.29326172 * 10240; time = 0.0409s; samplesPerSecond = 250446.4
12/20/2016 15:28:40:  Epoch[18 of 25]-Minibatch[ 721- 730, 66.51%]: CE.SM = 0.87120361 * 10240; Err = 0.28847656 * 10240; time = 0.0413s; samplesPerSecond = 248032.0
12/20/2016 15:28:40:  Epoch[18 of 25]-Minibatch[ 731- 740, 67.43%]: CE.SM = 0.91523438 * 10240; Err = 0.29267578 * 10240; time = 0.0410s; samplesPerSecond = 249932.9
12/20/2016 15:28:40:  Epoch[18 of 25]-Minibatch[ 741- 750, 68.34%]: CE.SM = 0.90683594 * 10240; Err = 0.29472656 * 10240; time = 0.0407s; samplesPerSecond = 251764.1
12/20/2016 15:28:40:  Epoch[18 of 25]-Minibatch[ 751- 760, 69.25%]: CE.SM = 0.90033569 * 10240; Err = 0.29228516 * 10240; time = 0.0410s; samplesPerSecond = 249579.6
12/20/2016 15:28:40:  Epoch[18 of 25]-Minibatch[ 761- 770, 70.16%]: CE.SM = 0.90782471 * 10240; Err = 0.29560547 * 10240; time = 0.0400s; samplesPerSecond = 256269.1
12/20/2016 15:28:40:  Epoch[18 of 25]-Minibatch[ 771- 780, 71.07%]: CE.SM = 0.87052612 * 10240; Err = 0.28486328 * 10240; time = 0.0396s; samplesPerSecond = 258279.3
12/20/2016 15:28:40:  Epoch[18 of 25]-Minibatch[ 781- 790, 71.98%]: CE.SM = 0.87833252 * 10240; Err = 0.28857422 * 10240; time = 0.0394s; samplesPerSecond = 259707.3
12/20/2016 15:28:40:  Epoch[18 of 25]-Minibatch[ 791- 800, 72.89%]: CE.SM = 0.87550049 * 10240; Err = 0.28427734 * 10240; time = 0.0394s; samplesPerSecond = 259720.5
12/20/2016 15:28:41:  Epoch[18 of 25]-Minibatch[ 801- 810, 73.80%]: CE.SM = 0.88869629 * 10240; Err = 0.28876953 * 10240; time = 0.0399s; samplesPerSecond = 256429.5
12/20/2016 15:28:41:  Epoch[18 of 25]-Minibatch[ 811- 820, 74.72%]: CE.SM = 0.88814697 * 10240; Err = 0.29521484 * 10240; time = 0.0394s; samplesPerSecond = 259674.4
12/20/2016 15:28:41:  Epoch[18 of 25]-Minibatch[ 821- 830, 75.63%]: CE.SM = 0.90090332 * 10240; Err = 0.29824219 * 10240; time = 0.0398s; samplesPerSecond = 257260.6
12/20/2016 15:28:41:  Epoch[18 of 25]-Minibatch[ 831- 840, 76.54%]: CE.SM = 0.88994751 * 10240; Err = 0.28955078 * 10240; time = 0.0408s; samplesPerSecond = 250912.7
12/20/2016 15:28:41:  Epoch[18 of 25]-Minibatch[ 841- 850, 77.45%]: CE.SM = 0.88897705 * 10240; Err = 0.28886719 * 10240; time = 0.0405s; samplesPerSecond = 252527.7
12/20/2016 15:28:41:  Epoch[18 of 25]-Minibatch[ 851- 860, 78.36%]: CE.SM = 0.91168213 * 10240; Err = 0.29716797 * 10240; time = 0.0411s; samplesPerSecond = 249403.3
12/20/2016 15:28:41:  Epoch[18 of 25]-Minibatch[ 861- 870, 79.27%]: CE.SM = 0.87035522 * 10240; Err = 0.28183594 * 10240; time = 0.0406s; samplesPerSecond = 252260.2
12/20/2016 15:28:41:  Epoch[18 of 25]-Minibatch[ 871- 880, 80.18%]: CE.SM = 0.88597412 * 10240; Err = 0.28681641 * 10240; time = 0.0405s; samplesPerSecond = 252745.9
12/20/2016 15:28:41:  Epoch[18 of 25]-Minibatch[ 881- 890, 81.09%]: CE.SM = 0.88104248 * 10240; Err = 0.28867188 * 10240; time = 0.0401s; samplesPerSecond = 255431.7
12/20/2016 15:28:41:  Epoch[18 of 25]-Minibatch[ 891- 900, 82.00%]: CE.SM = 0.85686646 * 10240; Err = 0.28271484 * 10240; time = 0.0405s; samplesPerSecond = 252658.6
12/20/2016 15:28:41:  Epoch[18 of 25]-Minibatch[ 901- 910, 82.92%]: CE.SM = 0.87673340 * 10240; Err = 0.29169922 * 10240; time = 0.0413s; samplesPerSecond = 248001.9
12/20/2016 15:28:41:  Epoch[18 of 25]-Minibatch[ 911- 920, 83.83%]: CE.SM = 0.89444580 * 10240; Err = 0.29316406 * 10240; time = 0.0402s; samplesPerSecond = 254644.0
12/20/2016 15:28:41:  Epoch[18 of 25]-Minibatch[ 921- 930, 84.74%]: CE.SM = 0.87208252 * 10240; Err = 0.28173828 * 10240; time = 0.0406s; samplesPerSecond = 252042.9
12/20/2016 15:28:41:  Epoch[18 of 25]-Minibatch[ 931- 940, 85.65%]: CE.SM = 0.88341064 * 10240; Err = 0.28867188 * 10240; time = 0.0395s; samplesPerSecond = 258984.8
12/20/2016 15:28:41:  Epoch[18 of 25]-Minibatch[ 941- 950, 86.56%]: CE.SM = 0.90829468 * 10240; Err = 0.29736328 * 10240; time = 0.0397s; samplesPerSecond = 257765.7
12/20/2016 15:28:41:  Epoch[18 of 25]-Minibatch[ 951- 960, 87.47%]: CE.SM = 0.91029053 * 10240; Err = 0.29960938 * 10240; time = 0.0396s; samplesPerSecond = 258338.0
12/20/2016 15:28:41:  Epoch[18 of 25]-Minibatch[ 961- 970, 88.38%]: CE.SM = 0.88134155 * 10240; Err = 0.28769531 * 10240; time = 0.0393s; samplesPerSecond = 260321.3
12/20/2016 15:28:41:  Epoch[18 of 25]-Minibatch[ 971- 980, 89.29%]: CE.SM = 0.90134277 * 10240; Err = 0.29306641 * 10240; time = 0.0396s; samplesPerSecond = 258775.4
12/20/2016 15:28:41:  Epoch[18 of 25]-Minibatch[ 981- 990, 90.21%]: CE.SM = 0.89659424 * 10240; Err = 0.29130859 * 10240; time = 0.0399s; samplesPerSecond = 256615.9
12/20/2016 15:28:41:  Epoch[18 of 25]-Minibatch[ 991-1000, 91.12%]: CE.SM = 0.89070435 * 10240; Err = 0.28984375 * 10240; time = 0.0393s; samplesPerSecond = 260321.3
12/20/2016 15:28:41:  Epoch[18 of 25]-Minibatch[1001-1010, 92.03%]: CE.SM = 0.91090088 * 10240; Err = 0.30019531 * 10240; time = 0.0391s; samplesPerSecond = 261618.3
12/20/2016 15:28:41:  Epoch[18 of 25]-Minibatch[1011-1020, 92.94%]: CE.SM = 0.91301880 * 10240; Err = 0.29443359 * 10240; time = 0.0413s; samplesPerSecond = 248230.4
12/20/2016 15:28:41:  Epoch[18 of 25]-Minibatch[1021-1030, 93.85%]: CE.SM = 0.87901611 * 10240; Err = 0.28798828 * 10240; time = 0.0392s; samplesPerSecond = 261104.6
12/20/2016 15:28:41:  Epoch[18 of 25]-Minibatch[1031-1040, 94.76%]: CE.SM = 0.89003906 * 10240; Err = 0.28720703 * 10240; time = 0.0477s; samplesPerSecond = 214702.1
12/20/2016 15:28:41:  Epoch[18 of 25]-Minibatch[1041-1050, 95.67%]: CE.SM = 0.88964844 * 10240; Err = 0.28857422 * 10240; time = 0.0415s; samplesPerSecond = 246485.7
12/20/2016 15:28:42:  Epoch[18 of 25]-Minibatch[1051-1060, 96.58%]: CE.SM = 0.88574829 * 10240; Err = 0.28779297 * 10240; time = 0.0423s; samplesPerSecond = 241937.4
12/20/2016 15:28:42:  Epoch[18 of 25]-Minibatch[1061-1070, 97.49%]: CE.SM = 0.87697144 * 10240; Err = 0.28417969 * 10240; time = 0.0416s; samplesPerSecond = 245958.7
12/20/2016 15:28:42:  Epoch[18 of 25]-Minibatch[1071-1080, 98.41%]: CE.SM = 0.90316772 * 10240; Err = 0.29619141 * 10240; time = 0.0411s; samplesPerSecond = 249372.9
12/20/2016 15:28:42:  Epoch[18 of 25]-Minibatch[1081-1090, 99.32%]: CE.SM = 0.90186157 * 10240; Err = 0.29306641 * 10240; time = 0.0415s; samplesPerSecond = 246800.5
12/20/2016 15:28:42: Finished Epoch[18 of 25]: [Training] CE.SM = 0.89119966 * 1124823; Err = 0.29061906 * 1124823; totalSamplesSeen = 20246814; learningRatePerSample = 7.8124998e-05; epochTime=4.69038s
12/20/2016 15:28:42: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn.18'

12/20/2016 15:28:42: Starting Epoch 19: learning rate per sample = 0.000078  effective momentum = 0.900000  momentum as time constant = 9719.0 samples

12/20/2016 15:28:42: Starting minibatch loop.
12/20/2016 15:28:42:  Epoch[19 of 25]-Minibatch[   1-  10, 0.91%]: CE.SM = 0.87962294 * 10240; Err = 0.28632812 * 10240; time = 0.0460s; samplesPerSecond = 222487.8
12/20/2016 15:28:42:  Epoch[19 of 25]-Minibatch[  11-  20, 1.82%]: CE.SM = 0.89071550 * 10240; Err = 0.29472656 * 10240; time = 0.0445s; samplesPerSecond = 229998.7
12/20/2016 15:28:42:  Epoch[19 of 25]-Minibatch[  21-  30, 2.73%]: CE.SM = 0.88504391 * 10240; Err = 0.28916016 * 10240; time = 0.0418s; samplesPerSecond = 244964.4
12/20/2016 15:28:42:  Epoch[19 of 25]-Minibatch[  31-  40, 3.64%]: CE.SM = 0.89437389 * 10240; Err = 0.28886719 * 10240; time = 0.0410s; samplesPerSecond = 249993.9
12/20/2016 15:28:42:  Epoch[19 of 25]-Minibatch[  41-  50, 4.56%]: CE.SM = 0.90220795 * 10240; Err = 0.29736328 * 10240; time = 0.0409s; samplesPerSecond = 250158.8
12/20/2016 15:28:42:  Epoch[19 of 25]-Minibatch[  51-  60, 5.47%]: CE.SM = 0.87970276 * 10240; Err = 0.28623047 * 10240; time = 0.0410s; samplesPerSecond = 250024.4
12/20/2016 15:28:42:  Epoch[19 of 25]-Minibatch[  61-  70, 6.38%]: CE.SM = 0.90337944 * 10240; Err = 0.29667969 * 10240; time = 0.0414s; samplesPerSecond = 247594.2
12/20/2016 15:28:42:  Epoch[19 of 25]-Minibatch[  71-  80, 7.29%]: CE.SM = 0.87315063 * 10240; Err = 0.28681641 * 10240; time = 0.0409s; samplesPerSecond = 250428.0
12/20/2016 15:28:42:  Epoch[19 of 25]-Minibatch[  81-  90, 8.20%]: CE.SM = 0.89035721 * 10240; Err = 0.29062500 * 10240; time = 0.0415s; samplesPerSecond = 246723.2
12/20/2016 15:28:42:  Epoch[19 of 25]-Minibatch[  91- 100, 9.11%]: CE.SM = 0.90742874 * 10240; Err = 0.29296875 * 10240; time = 0.0399s; samplesPerSecond = 256686.6
12/20/2016 15:28:42:  Epoch[19 of 25]-Minibatch[ 101- 110, 10.02%]: CE.SM = 0.89921036 * 10240; Err = 0.29560547 * 10240; time = 0.0404s; samplesPerSecond = 253697.7
12/20/2016 15:28:42:  Epoch[19 of 25]-Minibatch[ 111- 120, 10.93%]: CE.SM = 0.87123871 * 10240; Err = 0.28417969 * 10240; time = 0.0404s; samplesPerSecond = 253741.7
12/20/2016 15:28:42:  Epoch[19 of 25]-Minibatch[ 121- 130, 11.85%]: CE.SM = 0.88871460 * 10240; Err = 0.29150391 * 10240; time = 0.0408s; samplesPerSecond = 250832.8
12/20/2016 15:28:43:  Epoch[19 of 25]-Minibatch[ 131- 140, 12.76%]: CE.SM = 0.87746887 * 10240; Err = 0.28886719 * 10240; time = 0.0398s; samplesPerSecond = 257047.5
12/20/2016 15:28:43:  Epoch[19 of 25]-Minibatch[ 141- 150, 13.67%]: CE.SM = 0.88139572 * 10240; Err = 0.29443359 * 10240; time = 0.0405s; samplesPerSecond = 253108.2
12/20/2016 15:28:43:  Epoch[19 of 25]-Minibatch[ 151- 160, 14.58%]: CE.SM = 0.86646881 * 10240; Err = 0.28613281 * 10240; time = 0.0403s; samplesPerSecond = 254346.7
12/20/2016 15:28:43:  Epoch[19 of 25]-Minibatch[ 161- 170, 15.49%]: CE.SM = 0.88643341 * 10240; Err = 0.28574219 * 10240; time = 0.0408s; samplesPerSecond = 250918.9
12/20/2016 15:28:43:  Epoch[19 of 25]-Minibatch[ 171- 180, 16.40%]: CE.SM = 0.86393127 * 10240; Err = 0.28183594 * 10240; time = 0.0415s; samplesPerSecond = 246711.3
12/20/2016 15:28:43:  Epoch[19 of 25]-Minibatch[ 181- 190, 17.31%]: CE.SM = 0.86647491 * 10240; Err = 0.27578125 * 10240; time = 0.0402s; samplesPerSecond = 254973.7
12/20/2016 15:28:43:  Epoch[19 of 25]-Minibatch[ 191- 200, 18.22%]: CE.SM = 0.89956207 * 10240; Err = 0.29541016 * 10240; time = 0.0399s; samplesPerSecond = 256815.4
12/20/2016 15:28:43:  Epoch[19 of 25]-Minibatch[ 201- 210, 19.13%]: CE.SM = 0.88895721 * 10240; Err = 0.29228516 * 10240; time = 0.0400s; samplesPerSecond = 256275.5
12/20/2016 15:28:43:  Epoch[19 of 25]-Minibatch[ 211- 220, 20.05%]: CE.SM = 0.88708954 * 10240; Err = 0.28535156 * 10240; time = 0.0404s; samplesPerSecond = 253521.8
12/20/2016 15:28:43:  Epoch[19 of 25]-Minibatch[ 221- 230, 20.96%]: CE.SM = 0.88720245 * 10240; Err = 0.28994141 * 10240; time = 0.0402s; samplesPerSecond = 254783.4
12/20/2016 15:28:43:  Epoch[19 of 25]-Minibatch[ 231- 240, 21.87%]: CE.SM = 0.88229523 * 10240; Err = 0.29238281 * 10240; time = 0.0402s; samplesPerSecond = 254435.2
12/20/2016 15:28:43:  Epoch[19 of 25]-Minibatch[ 241- 250, 22.78%]: CE.SM = 0.87356720 * 10240; Err = 0.27988281 * 10240; time = 0.0400s; samplesPerSecond = 256108.8
12/20/2016 15:28:43:  Epoch[19 of 25]-Minibatch[ 251- 260, 23.69%]: CE.SM = 0.90503387 * 10240; Err = 0.29501953 * 10240; time = 0.0397s; samplesPerSecond = 257759.2
12/20/2016 15:28:43:  Epoch[19 of 25]-Minibatch[ 261- 270, 24.60%]: CE.SM = 0.90319214 * 10240; Err = 0.29199219 * 10240; time = 0.0403s; samplesPerSecond = 254151.1
12/20/2016 15:28:43:  Epoch[19 of 25]-Minibatch[ 271- 280, 25.51%]: CE.SM = 0.88021545 * 10240; Err = 0.28115234 * 10240; time = 0.0406s; samplesPerSecond = 252111.2
12/20/2016 15:28:43:  Epoch[19 of 25]-Minibatch[ 281- 290, 26.42%]: CE.SM = 0.90535278 * 10240; Err = 0.28994141 * 10240; time = 0.0400s; samplesPerSecond = 256262.7
12/20/2016 15:28:43:  Epoch[19 of 25]-Minibatch[ 291- 300, 27.33%]: CE.SM = 0.88496399 * 10240; Err = 0.29072266 * 10240; time = 0.0402s; samplesPerSecond = 255005.5
12/20/2016 15:28:43:  Epoch[19 of 25]-Minibatch[ 301- 310, 28.25%]: CE.SM = 0.89364929 * 10240; Err = 0.29042969 * 10240; time = 0.0413s; samplesPerSecond = 248176.2
12/20/2016 15:28:43:  Epoch[19 of 25]-Minibatch[ 311- 320, 29.16%]: CE.SM = 0.89519043 * 10240; Err = 0.29042969 * 10240; time = 0.0411s; samplesPerSecond = 249015.1
12/20/2016 15:28:43:  Epoch[19 of 25]-Minibatch[ 321- 330, 30.07%]: CE.SM = 0.90115356 * 10240; Err = 0.29501953 * 10240; time = 0.0402s; samplesPerSecond = 255030.9
12/20/2016 15:28:43:  Epoch[19 of 25]-Minibatch[ 331- 340, 30.98%]: CE.SM = 0.89341736 * 10240; Err = 0.29277344 * 10240; time = 0.0407s; samplesPerSecond = 251455.0
12/20/2016 15:28:43:  Epoch[19 of 25]-Minibatch[ 341- 350, 31.89%]: CE.SM = 0.89242249 * 10240; Err = 0.28916016 * 10240; time = 0.0403s; samplesPerSecond = 254397.3
12/20/2016 15:28:43:  Epoch[19 of 25]-Minibatch[ 351- 360, 32.80%]: CE.SM = 0.88807373 * 10240; Err = 0.28369141 * 10240; time = 0.0399s; samplesPerSecond = 256378.2
12/20/2016 15:28:43:  Epoch[19 of 25]-Minibatch[ 361- 370, 33.71%]: CE.SM = 0.87855530 * 10240; Err = 0.28535156 * 10240; time = 0.0399s; samplesPerSecond = 256725.2
12/20/2016 15:28:44:  Epoch[19 of 25]-Minibatch[ 371- 380, 34.62%]: CE.SM = 0.87426758 * 10240; Err = 0.28652344 * 10240; time = 0.0397s; samplesPerSecond = 257999.5
12/20/2016 15:28:44:  Epoch[19 of 25]-Minibatch[ 381- 390, 35.54%]: CE.SM = 0.88787842 * 10240; Err = 0.29140625 * 10240; time = 0.0403s; samplesPerSecond = 254409.9
12/20/2016 15:28:44:  Epoch[19 of 25]-Minibatch[ 391- 400, 36.45%]: CE.SM = 0.88469238 * 10240; Err = 0.28808594 * 10240; time = 0.0398s; samplesPerSecond = 257383.4
12/20/2016 15:28:44:  Epoch[19 of 25]-Minibatch[ 401- 410, 37.36%]: CE.SM = 0.88736267 * 10240; Err = 0.29189453 * 10240; time = 0.0402s; samplesPerSecond = 254796.1
12/20/2016 15:28:44:  Epoch[19 of 25]-Minibatch[ 411- 420, 38.27%]: CE.SM = 0.89655151 * 10240; Err = 0.29511719 * 10240; time = 0.0405s; samplesPerSecond = 252733.4
12/20/2016 15:28:44:  Epoch[19 of 25]-Minibatch[ 421- 430, 39.18%]: CE.SM = 0.88424377 * 10240; Err = 0.28769531 * 10240; time = 0.0401s; samplesPerSecond = 255342.5
12/20/2016 15:28:44:  Epoch[19 of 25]-Minibatch[ 431- 440, 40.09%]: CE.SM = 0.88736572 * 10240; Err = 0.28662109 * 10240; time = 0.0405s; samplesPerSecond = 253008.2
12/20/2016 15:28:44:  Epoch[19 of 25]-Minibatch[ 441- 450, 41.00%]: CE.SM = 0.87603149 * 10240; Err = 0.28339844 * 10240; time = 0.0405s; samplesPerSecond = 252939.4
12/20/2016 15:28:44:  Epoch[19 of 25]-Minibatch[ 451- 460, 41.91%]: CE.SM = 0.88830872 * 10240; Err = 0.28837891 * 10240; time = 0.0399s; samplesPerSecond = 256358.9
12/20/2016 15:28:44:  Epoch[19 of 25]-Minibatch[ 461- 470, 42.82%]: CE.SM = 0.88743896 * 10240; Err = 0.29160156 * 10240; time = 0.0431s; samplesPerSecond = 237686.3
12/20/2016 15:28:44:  Epoch[19 of 25]-Minibatch[ 471- 480, 43.74%]: CE.SM = 0.86860046 * 10240; Err = 0.28691406 * 10240; time = 0.0402s; samplesPerSecond = 254530.1
12/20/2016 15:28:44:  Epoch[19 of 25]-Minibatch[ 481- 490, 44.65%]: CE.SM = 0.87680359 * 10240; Err = 0.28740234 * 10240; time = 0.0402s; samplesPerSecond = 254802.4
12/20/2016 15:28:44:  Epoch[19 of 25]-Minibatch[ 491- 500, 45.56%]: CE.SM = 0.89150391 * 10240; Err = 0.28955078 * 10240; time = 0.0401s; samplesPerSecond = 255425.3
12/20/2016 15:28:44:  Epoch[19 of 25]-Minibatch[ 501- 510, 46.47%]: CE.SM = 0.90589294 * 10240; Err = 0.29619141 * 10240; time = 0.0401s; samplesPerSecond = 255482.6
12/20/2016 15:28:44:  Epoch[19 of 25]-Minibatch[ 511- 520, 47.38%]: CE.SM = 0.85047607 * 10240; Err = 0.27929688 * 10240; time = 0.0400s; samplesPerSecond = 256012.8
12/20/2016 15:28:44:  Epoch[19 of 25]-Minibatch[ 521- 530, 48.29%]: CE.SM = 0.88740540 * 10240; Err = 0.29169922 * 10240; time = 0.0406s; samplesPerSecond = 252247.8
12/20/2016 15:28:44:  Epoch[19 of 25]-Minibatch[ 531- 540, 49.20%]: CE.SM = 0.88671265 * 10240; Err = 0.28681641 * 10240; time = 0.0408s; samplesPerSecond = 250710.0
12/20/2016 15:28:44:  Epoch[19 of 25]-Minibatch[ 541- 550, 50.11%]: CE.SM = 0.88181458 * 10240; Err = 0.28828125 * 10240; time = 0.0404s; samplesPerSecond = 253553.2
12/20/2016 15:28:44:  Epoch[19 of 25]-Minibatch[ 551- 560, 51.03%]: CE.SM = 0.86550293 * 10240; Err = 0.28632812 * 10240; time = 0.0398s; samplesPerSecond = 257066.8
12/20/2016 15:28:44:  Epoch[19 of 25]-Minibatch[ 561- 570, 51.94%]: CE.SM = 0.88218994 * 10240; Err = 0.28935547 * 10240; time = 0.0407s; samplesPerSecond = 251776.5
12/20/2016 15:28:44:  Epoch[19 of 25]-Minibatch[ 571- 580, 52.85%]: CE.SM = 0.88628235 * 10240; Err = 0.28896484 * 10240; time = 0.0398s; samplesPerSecond = 257383.4
12/20/2016 15:28:44:  Epoch[19 of 25]-Minibatch[ 581- 590, 53.76%]: CE.SM = 0.89143677 * 10240; Err = 0.28925781 * 10240; time = 0.0401s; samplesPerSecond = 255266.1
12/20/2016 15:28:44:  Epoch[19 of 25]-Minibatch[ 591- 600, 54.67%]: CE.SM = 0.89122925 * 10240; Err = 0.28867188 * 10240; time = 0.0403s; samplesPerSecond = 254359.4
12/20/2016 15:28:44:  Epoch[19 of 25]-Minibatch[ 601- 610, 55.58%]: CE.SM = 0.89847412 * 10240; Err = 0.29218750 * 10240; time = 0.0418s; samplesPerSecond = 244864.8
12/20/2016 15:28:44:  Epoch[19 of 25]-Minibatch[ 611- 620, 56.49%]: CE.SM = 0.88761597 * 10240; Err = 0.28789063 * 10240; time = 0.0421s; samplesPerSecond = 243409.6
12/20/2016 15:28:45:  Epoch[19 of 25]-Minibatch[ 621- 630, 57.40%]: CE.SM = 0.89013672 * 10240; Err = 0.28896484 * 10240; time = 0.0397s; samplesPerSecond = 258097.0
12/20/2016 15:28:45:  Epoch[19 of 25]-Minibatch[ 631- 640, 58.31%]: CE.SM = 0.86904297 * 10240; Err = 0.28076172 * 10240; time = 0.0401s; samplesPerSecond = 255234.3
12/20/2016 15:28:45:  Epoch[19 of 25]-Minibatch[ 641- 650, 59.23%]: CE.SM = 0.89935303 * 10240; Err = 0.29111328 * 10240; time = 0.0395s; samplesPerSecond = 259391.5
12/20/2016 15:28:45:  Epoch[19 of 25]-Minibatch[ 651- 660, 60.14%]: CE.SM = 0.88345337 * 10240; Err = 0.28525391 * 10240; time = 0.0400s; samplesPerSecond = 256070.4
12/20/2016 15:28:45:  Epoch[19 of 25]-Minibatch[ 661- 670, 61.05%]: CE.SM = 0.87100830 * 10240; Err = 0.28437500 * 10240; time = 0.0398s; samplesPerSecond = 257487.0
12/20/2016 15:28:45:  Epoch[19 of 25]-Minibatch[ 671- 680, 61.96%]: CE.SM = 0.89857178 * 10240; Err = 0.29394531 * 10240; time = 0.0405s; samplesPerSecond = 253001.9
12/20/2016 15:28:45:  Epoch[19 of 25]-Minibatch[ 681- 690, 62.87%]: CE.SM = 0.90263672 * 10240; Err = 0.29316406 * 10240; time = 0.0429s; samplesPerSecond = 238850.5
12/20/2016 15:28:45:  Epoch[19 of 25]-Minibatch[ 691- 700, 63.78%]: CE.SM = 0.89559937 * 10240; Err = 0.29960938 * 10240; time = 0.0440s; samplesPerSecond = 232864.9
12/20/2016 15:28:45:  Epoch[19 of 25]-Minibatch[ 701- 710, 64.69%]: CE.SM = 0.89122314 * 10240; Err = 0.28691406 * 10240; time = 0.0497s; samplesPerSecond = 205854.0
12/20/2016 15:28:45:  Epoch[19 of 25]-Minibatch[ 711- 720, 65.60%]: CE.SM = 0.88861694 * 10240; Err = 0.29560547 * 10240; time = 0.0402s; samplesPerSecond = 254840.5
12/20/2016 15:28:45:  Epoch[19 of 25]-Minibatch[ 721- 730, 66.51%]: CE.SM = 0.88264160 * 10240; Err = 0.28867188 * 10240; time = 0.0399s; samplesPerSecond = 256538.7
12/20/2016 15:28:45:  Epoch[19 of 25]-Minibatch[ 731- 740, 67.43%]: CE.SM = 0.88840942 * 10240; Err = 0.29248047 * 10240; time = 0.0439s; samplesPerSecond = 233310.5
12/20/2016 15:28:45:  Epoch[19 of 25]-Minibatch[ 741- 750, 68.34%]: CE.SM = 0.86893311 * 10240; Err = 0.28505859 * 10240; time = 0.0409s; samplesPerSecond = 250311.7
12/20/2016 15:28:45:  Epoch[19 of 25]-Minibatch[ 751- 760, 69.25%]: CE.SM = 0.86859741 * 10240; Err = 0.28974609 * 10240; time = 0.0399s; samplesPerSecond = 256718.8
12/20/2016 15:28:45:  Epoch[19 of 25]-Minibatch[ 761- 770, 70.16%]: CE.SM = 0.88505859 * 10240; Err = 0.28876953 * 10240; time = 0.0399s; samplesPerSecond = 256558.0
12/20/2016 15:28:45:  Epoch[19 of 25]-Minibatch[ 771- 780, 71.07%]: CE.SM = 0.90435181 * 10240; Err = 0.29472656 * 10240; time = 0.0403s; samplesPerSecond = 253829.8
12/20/2016 15:28:45:  Epoch[19 of 25]-Minibatch[ 781- 790, 71.98%]: CE.SM = 0.87858276 * 10240; Err = 0.29160156 * 10240; time = 0.0414s; samplesPerSecond = 247235.5
12/20/2016 15:28:45:  Epoch[19 of 25]-Minibatch[ 791- 800, 72.89%]: CE.SM = 0.90609741 * 10240; Err = 0.29550781 * 10240; time = 0.0410s; samplesPerSecond = 249987.8
12/20/2016 15:28:45:  Epoch[19 of 25]-Minibatch[ 801- 810, 73.80%]: CE.SM = 0.88635864 * 10240; Err = 0.28564453 * 10240; time = 0.0400s; samplesPerSecond = 255744.3
12/20/2016 15:28:45:  Epoch[19 of 25]-Minibatch[ 811- 820, 74.72%]: CE.SM = 0.86209717 * 10240; Err = 0.28535156 * 10240; time = 0.0404s; samplesPerSecond = 253641.1
12/20/2016 15:28:45:  Epoch[19 of 25]-Minibatch[ 821- 830, 75.63%]: CE.SM = 0.90347290 * 10240; Err = 0.29199219 * 10240; time = 0.0405s; samplesPerSecond = 252689.8
12/20/2016 15:28:45:  Epoch[19 of 25]-Minibatch[ 831- 840, 76.54%]: CE.SM = 0.89344482 * 10240; Err = 0.28876953 * 10240; time = 0.0409s; samplesPerSecond = 250073.3
12/20/2016 15:28:45:  Epoch[19 of 25]-Minibatch[ 841- 850, 77.45%]: CE.SM = 0.88660889 * 10240; Err = 0.28457031 * 10240; time = 0.0403s; samplesPerSecond = 254327.8
12/20/2016 15:28:45:  Epoch[19 of 25]-Minibatch[ 851- 860, 78.36%]: CE.SM = 0.88671265 * 10240; Err = 0.29023437 * 10240; time = 0.0399s; samplesPerSecond = 256532.3
12/20/2016 15:28:46:  Epoch[19 of 25]-Minibatch[ 861- 870, 79.27%]: CE.SM = 0.89844360 * 10240; Err = 0.29160156 * 10240; time = 0.0398s; samplesPerSecond = 257047.5
12/20/2016 15:28:46:  Epoch[19 of 25]-Minibatch[ 871- 880, 80.18%]: CE.SM = 0.88075562 * 10240; Err = 0.28710938 * 10240; time = 0.0397s; samplesPerSecond = 257694.3
12/20/2016 15:28:46:  Epoch[19 of 25]-Minibatch[ 881- 890, 81.09%]: CE.SM = 0.89108276 * 10240; Err = 0.29130859 * 10240; time = 0.0399s; samplesPerSecond = 256493.8
12/20/2016 15:28:46:  Epoch[19 of 25]-Minibatch[ 891- 900, 82.00%]: CE.SM = 0.87952881 * 10240; Err = 0.28652344 * 10240; time = 0.0401s; samplesPerSecond = 255540.0
12/20/2016 15:28:46:  Epoch[19 of 25]-Minibatch[ 901- 910, 82.92%]: CE.SM = 0.89105835 * 10240; Err = 0.28769531 * 10240; time = 0.0399s; samplesPerSecond = 256628.7
12/20/2016 15:28:46:  Epoch[19 of 25]-Minibatch[ 911- 920, 83.83%]: CE.SM = 0.89567871 * 10240; Err = 0.28828125 * 10240; time = 0.0398s; samplesPerSecond = 256989.4
12/20/2016 15:28:46:  Epoch[19 of 25]-Minibatch[ 921- 930, 84.74%]: CE.SM = 0.89037476 * 10240; Err = 0.29326172 * 10240; time = 0.0408s; samplesPerSecond = 251251.3
12/20/2016 15:28:46:  Epoch[19 of 25]-Minibatch[ 931- 940, 85.65%]: CE.SM = 0.88941040 * 10240; Err = 0.28925781 * 10240; time = 0.0401s; samplesPerSecond = 255151.6
12/20/2016 15:28:46:  Epoch[19 of 25]-Minibatch[ 941- 950, 86.56%]: CE.SM = 0.87369385 * 10240; Err = 0.28740234 * 10240; time = 0.0410s; samplesPerSecond = 249476.2
12/20/2016 15:28:46:  Epoch[19 of 25]-Minibatch[ 951- 960, 87.47%]: CE.SM = 0.87839355 * 10240; Err = 0.29345703 * 10240; time = 0.0397s; samplesPerSecond = 257993.0
12/20/2016 15:28:46:  Epoch[19 of 25]-Minibatch[ 961- 970, 88.38%]: CE.SM = 0.89838867 * 10240; Err = 0.29287109 * 10240; time = 0.0401s; samplesPerSecond = 255610.2
12/20/2016 15:28:46:  Epoch[19 of 25]-Minibatch[ 971- 980, 89.29%]: CE.SM = 0.88470459 * 10240; Err = 0.28320312 * 10240; time = 0.0397s; samplesPerSecond = 257791.7
12/20/2016 15:28:46:  Epoch[19 of 25]-Minibatch[ 981- 990, 90.21%]: CE.SM = 0.91701050 * 10240; Err = 0.30136719 * 10240; time = 0.0402s; samplesPerSecond = 255018.2
12/20/2016 15:28:46:  Epoch[19 of 25]-Minibatch[ 991-1000, 91.12%]: CE.SM = 0.89586792 * 10240; Err = 0.29140625 * 10240; time = 0.0397s; samplesPerSecond = 258084.0
12/20/2016 15:28:46:  Epoch[19 of 25]-Minibatch[1001-1010, 92.03%]: CE.SM = 0.90897827 * 10240; Err = 0.29580078 * 10240; time = 0.0400s; samplesPerSecond = 256032.0
12/20/2016 15:28:46:  Epoch[19 of 25]-Minibatch[1011-1020, 92.94%]: CE.SM = 0.88585205 * 10240; Err = 0.28964844 * 10240; time = 0.0394s; samplesPerSecond = 259918.3
12/20/2016 15:28:46:  Epoch[19 of 25]-Minibatch[1021-1030, 93.85%]: CE.SM = 0.90162964 * 10240; Err = 0.29853516 * 10240; time = 0.0396s; samplesPerSecond = 258814.6
12/20/2016 15:28:46:  Epoch[19 of 25]-Minibatch[1031-1040, 94.76%]: CE.SM = 0.88466797 * 10240; Err = 0.28857422 * 10240; time = 0.0399s; samplesPerSecond = 256635.2
12/20/2016 15:28:46:  Epoch[19 of 25]-Minibatch[1041-1050, 95.67%]: CE.SM = 0.90917358 * 10240; Err = 0.29970703 * 10240; time = 0.0401s; samplesPerSecond = 255158.0
12/20/2016 15:28:46:  Epoch[19 of 25]-Minibatch[1051-1060, 96.58%]: CE.SM = 0.89796143 * 10240; Err = 0.28603516 * 10240; time = 0.0397s; samplesPerSecond = 257908.5
12/20/2016 15:28:46:  Epoch[19 of 25]-Minibatch[1061-1070, 97.49%]: CE.SM = 0.90278931 * 10240; Err = 0.29658203 * 10240; time = 0.0393s; samplesPerSecond = 260255.2
12/20/2016 15:28:46:  Epoch[19 of 25]-Minibatch[1071-1080, 98.41%]: CE.SM = 0.91046753 * 10240; Err = 0.29179688 * 10240; time = 0.0393s; samplesPerSecond = 260420.6
12/20/2016 15:28:46:  Epoch[19 of 25]-Minibatch[1081-1090, 99.32%]: CE.SM = 0.87406616 * 10240; Err = 0.28574219 * 10240; time = 0.0396s; samplesPerSecond = 258318.4
12/20/2016 15:28:46: Finished Epoch[19 of 25]: [Training] CE.SM = 0.88792142 * 1124823; Err = 0.28962068 * 1124823; totalSamplesSeen = 21371637; learningRatePerSample = 7.8124998e-05; epochTime=4.64935s
12/20/2016 15:28:46: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn.19'

12/20/2016 15:28:47: Starting Epoch 20: learning rate per sample = 0.000078  effective momentum = 0.900000  momentum as time constant = 9719.0 samples

12/20/2016 15:28:47: Starting minibatch loop.
12/20/2016 15:28:47:  Epoch[20 of 25]-Minibatch[   1-  10, 0.91%]: CE.SM = 0.88768024 * 10240; Err = 0.28984375 * 10240; time = 0.0438s; samplesPerSecond = 233613.9
12/20/2016 15:28:47:  Epoch[20 of 25]-Minibatch[  11-  20, 1.82%]: CE.SM = 0.88095837 * 10240; Err = 0.28398438 * 10240; time = 0.0397s; samplesPerSecond = 257713.8
12/20/2016 15:28:47:  Epoch[20 of 25]-Minibatch[  21-  30, 2.73%]: CE.SM = 0.88968582 * 10240; Err = 0.28945312 * 10240; time = 0.0399s; samplesPerSecond = 256686.6
12/20/2016 15:28:47:  Epoch[20 of 25]-Minibatch[  31-  40, 3.64%]: CE.SM = 0.88780632 * 10240; Err = 0.28730469 * 10240; time = 0.0398s; samplesPerSecond = 257015.2
12/20/2016 15:28:47:  Epoch[20 of 25]-Minibatch[  41-  50, 4.56%]: CE.SM = 0.88793716 * 10240; Err = 0.28623047 * 10240; time = 0.0397s; samplesPerSecond = 257999.5
12/20/2016 15:28:47:  Epoch[20 of 25]-Minibatch[  51-  60, 5.47%]: CE.SM = 0.88001175 * 10240; Err = 0.28574219 * 10240; time = 0.0397s; samplesPerSecond = 258175.1
12/20/2016 15:28:47:  Epoch[20 of 25]-Minibatch[  61-  70, 6.38%]: CE.SM = 0.89198837 * 10240; Err = 0.28847656 * 10240; time = 0.0399s; samplesPerSecond = 256339.7
12/20/2016 15:28:47:  Epoch[20 of 25]-Minibatch[  71-  80, 7.29%]: CE.SM = 0.87966728 * 10240; Err = 0.28496094 * 10240; time = 0.0399s; samplesPerSecond = 256545.2
12/20/2016 15:28:47:  Epoch[20 of 25]-Minibatch[  81-  90, 8.20%]: CE.SM = 0.87180252 * 10240; Err = 0.28212891 * 10240; time = 0.0400s; samplesPerSecond = 255795.4
12/20/2016 15:28:47:  Epoch[20 of 25]-Minibatch[  91- 100, 9.11%]: CE.SM = 0.88184738 * 10240; Err = 0.29121094 * 10240; time = 0.0395s; samplesPerSecond = 258958.6
12/20/2016 15:28:47:  Epoch[20 of 25]-Minibatch[ 101- 110, 10.02%]: CE.SM = 0.89539413 * 10240; Err = 0.29130859 * 10240; time = 0.0400s; samplesPerSecond = 255916.8
12/20/2016 15:28:47:  Epoch[20 of 25]-Minibatch[ 111- 120, 10.93%]: CE.SM = 0.87660751 * 10240; Err = 0.28408203 * 10240; time = 0.0400s; samplesPerSecond = 256249.8
12/20/2016 15:28:47:  Epoch[20 of 25]-Minibatch[ 121- 130, 11.85%]: CE.SM = 0.89095459 * 10240; Err = 0.28593750 * 10240; time = 0.0402s; samplesPerSecond = 254732.7
12/20/2016 15:28:47:  Epoch[20 of 25]-Minibatch[ 131- 140, 12.76%]: CE.SM = 0.90195847 * 10240; Err = 0.29658203 * 10240; time = 0.0397s; samplesPerSecond = 258051.5
12/20/2016 15:28:47:  Epoch[20 of 25]-Minibatch[ 141- 150, 13.67%]: CE.SM = 0.86840210 * 10240; Err = 0.28671875 * 10240; time = 0.0402s; samplesPerSecond = 254732.7
12/20/2016 15:28:47:  Epoch[20 of 25]-Minibatch[ 151- 160, 14.58%]: CE.SM = 0.89550934 * 10240; Err = 0.29140625 * 10240; time = 0.0398s; samplesPerSecond = 257428.7
12/20/2016 15:28:47:  Epoch[20 of 25]-Minibatch[ 161- 170, 15.49%]: CE.SM = 0.89014740 * 10240; Err = 0.29111328 * 10240; time = 0.0400s; samplesPerSecond = 256064.0
12/20/2016 15:28:47:  Epoch[20 of 25]-Minibatch[ 171- 180, 16.40%]: CE.SM = 0.88664856 * 10240; Err = 0.28583984 * 10240; time = 0.0400s; samplesPerSecond = 255744.3
12/20/2016 15:28:47:  Epoch[20 of 25]-Minibatch[ 181- 190, 17.31%]: CE.SM = 0.88349915 * 10240; Err = 0.28691406 * 10240; time = 0.0400s; samplesPerSecond = 256211.4
12/20/2016 15:28:47:  Epoch[20 of 25]-Minibatch[ 191- 200, 18.22%]: CE.SM = 0.88416443 * 10240; Err = 0.28808594 * 10240; time = 0.0394s; samplesPerSecond = 260063.5
12/20/2016 15:28:48:  Epoch[20 of 25]-Minibatch[ 201- 210, 19.13%]: CE.SM = 0.87079773 * 10240; Err = 0.28535156 * 10240; time = 0.0396s; samplesPerSecond = 258755.7
12/20/2016 15:28:48:  Epoch[20 of 25]-Minibatch[ 211- 220, 20.05%]: CE.SM = 0.88174896 * 10240; Err = 0.28769531 * 10240; time = 0.0397s; samplesPerSecond = 257772.2
12/20/2016 15:28:48:  Epoch[20 of 25]-Minibatch[ 221- 230, 20.96%]: CE.SM = 0.89555969 * 10240; Err = 0.29101562 * 10240; time = 0.0400s; samplesPerSecond = 255744.3
12/20/2016 15:28:48:  Epoch[20 of 25]-Minibatch[ 231- 240, 21.87%]: CE.SM = 0.89794006 * 10240; Err = 0.29296875 * 10240; time = 0.0400s; samplesPerSecond = 256089.6
12/20/2016 15:28:48:  Epoch[20 of 25]-Minibatch[ 241- 250, 22.78%]: CE.SM = 0.89892883 * 10240; Err = 0.29423828 * 10240; time = 0.0396s; samplesPerSecond = 258664.2
12/20/2016 15:28:48:  Epoch[20 of 25]-Minibatch[ 251- 260, 23.69%]: CE.SM = 0.88218079 * 10240; Err = 0.29189453 * 10240; time = 0.0398s; samplesPerSecond = 257234.7
12/20/2016 15:28:48:  Epoch[20 of 25]-Minibatch[ 261- 270, 24.60%]: CE.SM = 0.89503174 * 10240; Err = 0.29287109 * 10240; time = 0.0395s; samplesPerSecond = 258939.0
12/20/2016 15:28:48:  Epoch[20 of 25]-Minibatch[ 271- 280, 25.51%]: CE.SM = 0.86927795 * 10240; Err = 0.28486328 * 10240; time = 0.0399s; samplesPerSecond = 256391.0
12/20/2016 15:28:48:  Epoch[20 of 25]-Minibatch[ 281- 290, 26.42%]: CE.SM = 0.90018158 * 10240; Err = 0.28828125 * 10240; time = 0.0400s; samplesPerSecond = 255686.8
12/20/2016 15:28:48:  Epoch[20 of 25]-Minibatch[ 291- 300, 27.33%]: CE.SM = 0.87864990 * 10240; Err = 0.29218750 * 10240; time = 0.0398s; samplesPerSecond = 257273.5
12/20/2016 15:28:48:  Epoch[20 of 25]-Minibatch[ 301- 310, 28.25%]: CE.SM = 0.90269775 * 10240; Err = 0.28828125 * 10240; time = 0.0399s; samplesPerSecond = 256352.5
12/20/2016 15:28:48:  Epoch[20 of 25]-Minibatch[ 311- 320, 29.16%]: CE.SM = 0.89104919 * 10240; Err = 0.29082031 * 10240; time = 0.0401s; samplesPerSecond = 255431.7
12/20/2016 15:28:48:  Epoch[20 of 25]-Minibatch[ 321- 330, 30.07%]: CE.SM = 0.89376526 * 10240; Err = 0.29218750 * 10240; time = 0.0398s; samplesPerSecond = 257577.7
12/20/2016 15:28:48:  Epoch[20 of 25]-Minibatch[ 331- 340, 30.98%]: CE.SM = 0.84693909 * 10240; Err = 0.27382812 * 10240; time = 0.0399s; samplesPerSecond = 256500.2
12/20/2016 15:28:48:  Epoch[20 of 25]-Minibatch[ 341- 350, 31.89%]: CE.SM = 0.90869751 * 10240; Err = 0.29541016 * 10240; time = 0.0401s; samplesPerSecond = 255253.4
12/20/2016 15:28:48:  Epoch[20 of 25]-Minibatch[ 351- 360, 32.80%]: CE.SM = 0.86951599 * 10240; Err = 0.28281250 * 10240; time = 0.0398s; samplesPerSecond = 257280.0
12/20/2016 15:28:48:  Epoch[20 of 25]-Minibatch[ 361- 370, 33.71%]: CE.SM = 0.89120178 * 10240; Err = 0.28789063 * 10240; time = 0.0400s; samplesPerSecond = 255718.7
12/20/2016 15:28:48:  Epoch[20 of 25]-Minibatch[ 371- 380, 34.62%]: CE.SM = 0.88424683 * 10240; Err = 0.28935547 * 10240; time = 0.0396s; samplesPerSecond = 258651.2
12/20/2016 15:28:48:  Epoch[20 of 25]-Minibatch[ 381- 390, 35.54%]: CE.SM = 0.87743530 * 10240; Err = 0.28837891 * 10240; time = 0.0401s; samplesPerSecond = 255342.5
12/20/2016 15:28:48:  Epoch[20 of 25]-Minibatch[ 391- 400, 36.45%]: CE.SM = 0.89331360 * 10240; Err = 0.29335937 * 10240; time = 0.0401s; samplesPerSecond = 255482.6
12/20/2016 15:28:48:  Epoch[20 of 25]-Minibatch[ 401- 410, 37.36%]: CE.SM = 0.89030762 * 10240; Err = 0.29257813 * 10240; time = 0.0399s; samplesPerSecond = 256506.6
12/20/2016 15:28:48:  Epoch[20 of 25]-Minibatch[ 411- 420, 38.27%]: CE.SM = 0.89311218 * 10240; Err = 0.29267578 * 10240; time = 0.0399s; samplesPerSecond = 256416.7
12/20/2016 15:28:48:  Epoch[20 of 25]-Minibatch[ 421- 430, 39.18%]: CE.SM = 0.86376953 * 10240; Err = 0.27958984 * 10240; time = 0.0399s; samplesPerSecond = 256808.9
12/20/2016 15:28:48:  Epoch[20 of 25]-Minibatch[ 431- 440, 40.09%]: CE.SM = 0.86089478 * 10240; Err = 0.28388672 * 10240; time = 0.0399s; samplesPerSecond = 256854.0
12/20/2016 15:28:48:  Epoch[20 of 25]-Minibatch[ 441- 450, 41.00%]: CE.SM = 0.89501343 * 10240; Err = 0.29218750 * 10240; time = 0.0399s; samplesPerSecond = 256796.1
12/20/2016 15:28:49:  Epoch[20 of 25]-Minibatch[ 451- 460, 41.91%]: CE.SM = 0.89939575 * 10240; Err = 0.29238281 * 10240; time = 0.0393s; samplesPerSecond = 260480.3
12/20/2016 15:28:49:  Epoch[20 of 25]-Minibatch[ 461- 470, 42.82%]: CE.SM = 0.88983459 * 10240; Err = 0.28798828 * 10240; time = 0.0394s; samplesPerSecond = 259707.3
12/20/2016 15:28:49:  Epoch[20 of 25]-Minibatch[ 471- 480, 43.74%]: CE.SM = 0.88051453 * 10240; Err = 0.28613281 * 10240; time = 0.0397s; samplesPerSecond = 257941.0
12/20/2016 15:28:49:  Epoch[20 of 25]-Minibatch[ 481- 490, 44.65%]: CE.SM = 0.89630432 * 10240; Err = 0.28808594 * 10240; time = 0.0394s; samplesPerSecond = 260215.5
12/20/2016 15:28:49:  Epoch[20 of 25]-Minibatch[ 491- 500, 45.56%]: CE.SM = 0.90645447 * 10240; Err = 0.29150391 * 10240; time = 0.0392s; samplesPerSecond = 261457.9
12/20/2016 15:28:49:  Epoch[20 of 25]-Minibatch[ 501- 510, 46.47%]: CE.SM = 0.88816528 * 10240; Err = 0.28300781 * 10240; time = 0.0396s; samplesPerSecond = 258847.3
12/20/2016 15:28:49:  Epoch[20 of 25]-Minibatch[ 511- 520, 47.38%]: CE.SM = 0.87147217 * 10240; Err = 0.29179688 * 10240; time = 0.0394s; samplesPerSecond = 259944.7
12/20/2016 15:28:49:  Epoch[20 of 25]-Minibatch[ 521- 530, 48.29%]: CE.SM = 0.90411377 * 10240; Err = 0.29687500 * 10240; time = 0.0400s; samplesPerSecond = 255686.8
12/20/2016 15:28:49:  Epoch[20 of 25]-Minibatch[ 531- 540, 49.20%]: CE.SM = 0.87564697 * 10240; Err = 0.28847656 * 10240; time = 0.0398s; samplesPerSecond = 257041.0
12/20/2016 15:28:49:  Epoch[20 of 25]-Minibatch[ 541- 550, 50.11%]: CE.SM = 0.88858337 * 10240; Err = 0.28847656 * 10240; time = 0.0397s; samplesPerSecond = 258110.0
12/20/2016 15:28:49:  Epoch[20 of 25]-Minibatch[ 551- 560, 51.03%]: CE.SM = 0.87669373 * 10240; Err = 0.28847656 * 10240; time = 0.0398s; samplesPerSecond = 257558.2
12/20/2016 15:28:49:  Epoch[20 of 25]-Minibatch[ 561- 570, 51.94%]: CE.SM = 0.88137817 * 10240; Err = 0.28740234 * 10240; time = 0.0400s; samplesPerSecond = 255948.8
12/20/2016 15:28:49:  Epoch[20 of 25]-Minibatch[ 571- 580, 52.85%]: CE.SM = 0.88656616 * 10240; Err = 0.29257813 * 10240; time = 0.0398s; samplesPerSecond = 257280.0
12/20/2016 15:28:49:  Epoch[20 of 25]-Minibatch[ 581- 590, 53.76%]: CE.SM = 0.87075806 * 10240; Err = 0.28798828 * 10240; time = 0.0400s; samplesPerSecond = 255961.6
12/20/2016 15:28:49:  Epoch[20 of 25]-Minibatch[ 591- 600, 54.67%]: CE.SM = 0.87952881 * 10240; Err = 0.29121094 * 10240; time = 0.0395s; samplesPerSecond = 259509.9
12/20/2016 15:28:49:  Epoch[20 of 25]-Minibatch[ 601- 610, 55.58%]: CE.SM = 0.87987061 * 10240; Err = 0.29511719 * 10240; time = 0.0396s; samplesPerSecond = 258259.8
12/20/2016 15:28:49:  Epoch[20 of 25]-Minibatch[ 611- 620, 56.49%]: CE.SM = 0.89714355 * 10240; Err = 0.29384766 * 10240; time = 0.0398s; samplesPerSecond = 257241.2
12/20/2016 15:28:49:  Epoch[20 of 25]-Minibatch[ 621- 630, 57.40%]: CE.SM = 0.87640381 * 10240; Err = 0.28925781 * 10240; time = 0.0404s; samplesPerSecond = 253408.9
12/20/2016 15:28:49:  Epoch[20 of 25]-Minibatch[ 631- 640, 58.31%]: CE.SM = 0.88957520 * 10240; Err = 0.29472656 * 10240; time = 0.0396s; samplesPerSecond = 258416.2
12/20/2016 15:28:49:  Epoch[20 of 25]-Minibatch[ 641- 650, 59.23%]: CE.SM = 0.88756104 * 10240; Err = 0.28720703 * 10240; time = 0.0395s; samplesPerSecond = 259096.2
12/20/2016 15:28:49:  Epoch[20 of 25]-Minibatch[ 651- 660, 60.14%]: CE.SM = 0.89672852 * 10240; Err = 0.28925781 * 10240; time = 0.0392s; samplesPerSecond = 261257.8
12/20/2016 15:28:49:  Epoch[20 of 25]-Minibatch[ 661- 670, 61.05%]: CE.SM = 0.88207397 * 10240; Err = 0.28037109 * 10240; time = 0.0390s; samplesPerSecond = 262799.9
12/20/2016 15:28:49:  Epoch[20 of 25]-Minibatch[ 671- 680, 61.96%]: CE.SM = 0.88452759 * 10240; Err = 0.28642578 * 10240; time = 0.0392s; samplesPerSecond = 261051.3
12/20/2016 15:28:49:  Epoch[20 of 25]-Minibatch[ 681- 690, 62.87%]: CE.SM = 0.88851929 * 10240; Err = 0.29335937 * 10240; time = 0.0391s; samplesPerSecond = 261591.5
12/20/2016 15:28:49:  Epoch[20 of 25]-Minibatch[ 691- 700, 63.78%]: CE.SM = 0.88453369 * 10240; Err = 0.28671875 * 10240; time = 0.0398s; samplesPerSecond = 257112.0
12/20/2016 15:28:50:  Epoch[20 of 25]-Minibatch[ 701- 710, 64.69%]: CE.SM = 0.88823242 * 10240; Err = 0.28886719 * 10240; time = 0.0399s; samplesPerSecond = 256879.8
12/20/2016 15:28:50:  Epoch[20 of 25]-Minibatch[ 711- 720, 65.60%]: CE.SM = 0.90439453 * 10240; Err = 0.29501953 * 10240; time = 0.0402s; samplesPerSecond = 254910.3
12/20/2016 15:28:50:  Epoch[20 of 25]-Minibatch[ 721- 730, 66.51%]: CE.SM = 0.88112793 * 10240; Err = 0.28886719 * 10240; time = 0.0402s; samplesPerSecond = 254631.4
12/20/2016 15:28:50:  Epoch[20 of 25]-Minibatch[ 731- 740, 67.43%]: CE.SM = 0.88806152 * 10240; Err = 0.28583984 * 10240; time = 0.0404s; samplesPerSecond = 253691.4
12/20/2016 15:28:50:  Epoch[20 of 25]-Minibatch[ 741- 750, 68.34%]: CE.SM = 0.89862671 * 10240; Err = 0.29414062 * 10240; time = 0.0396s; samplesPerSecond = 258795.0
12/20/2016 15:28:50:  Epoch[20 of 25]-Minibatch[ 751- 760, 69.25%]: CE.SM = 0.87491455 * 10240; Err = 0.29257813 * 10240; time = 0.0393s; samplesPerSecond = 260387.5
12/20/2016 15:28:50:  Epoch[20 of 25]-Minibatch[ 761- 770, 70.16%]: CE.SM = 0.88272095 * 10240; Err = 0.28271484 * 10240; time = 0.0391s; samplesPerSecond = 261885.9
12/20/2016 15:28:50:  Epoch[20 of 25]-Minibatch[ 771- 780, 71.07%]: CE.SM = 0.87257080 * 10240; Err = 0.28574219 * 10240; time = 0.0389s; samplesPerSecond = 262914.7
12/20/2016 15:28:50:  Epoch[20 of 25]-Minibatch[ 781- 790, 71.98%]: CE.SM = 0.88374023 * 10240; Err = 0.28505859 * 10240; time = 0.0392s; samplesPerSecond = 261311.1
12/20/2016 15:28:50:  Epoch[20 of 25]-Minibatch[ 791- 800, 72.89%]: CE.SM = 0.88535767 * 10240; Err = 0.28935547 * 10240; time = 0.0396s; samplesPerSecond = 258487.9
12/20/2016 15:28:50:  Epoch[20 of 25]-Minibatch[ 801- 810, 73.80%]: CE.SM = 0.87940674 * 10240; Err = 0.29013672 * 10240; time = 0.0395s; samplesPerSecond = 259437.5
12/20/2016 15:28:50:  Epoch[20 of 25]-Minibatch[ 811- 820, 74.72%]: CE.SM = 0.87711792 * 10240; Err = 0.29091797 * 10240; time = 0.0392s; samplesPerSecond = 261237.8
12/20/2016 15:28:50:  Epoch[20 of 25]-Minibatch[ 821- 830, 75.63%]: CE.SM = 0.87481079 * 10240; Err = 0.28544922 * 10240; time = 0.0392s; samplesPerSecond = 261271.1
12/20/2016 15:28:50:  Epoch[20 of 25]-Minibatch[ 831- 840, 76.54%]: CE.SM = 0.86143188 * 10240; Err = 0.28408203 * 10240; time = 0.0392s; samplesPerSecond = 261224.5
12/20/2016 15:28:50:  Epoch[20 of 25]-Minibatch[ 841- 850, 77.45%]: CE.SM = 0.88527832 * 10240; Err = 0.28662109 * 10240; time = 0.0393s; samplesPerSecond = 260752.2
12/20/2016 15:28:50:  Epoch[20 of 25]-Minibatch[ 851- 860, 78.36%]: CE.SM = 0.87584839 * 10240; Err = 0.28251953 * 10240; time = 0.0390s; samplesPerSecond = 262416.1
12/20/2016 15:28:50:  Epoch[20 of 25]-Minibatch[ 861- 870, 79.27%]: CE.SM = 0.89443359 * 10240; Err = 0.29042969 * 10240; time = 0.0391s; samplesPerSecond = 261845.7
12/20/2016 15:28:50:  Epoch[20 of 25]-Minibatch[ 871- 880, 80.18%]: CE.SM = 0.89382935 * 10240; Err = 0.29257813 * 10240; time = 0.0390s; samplesPerSecond = 262705.6
12/20/2016 15:28:50:  Epoch[20 of 25]-Minibatch[ 881- 890, 81.09%]: CE.SM = 0.88421631 * 10240; Err = 0.28867188 * 10240; time = 0.0392s; samplesPerSecond = 261051.3
12/20/2016 15:28:50:  Epoch[20 of 25]-Minibatch[ 891- 900, 82.00%]: CE.SM = 0.87937622 * 10240; Err = 0.28955078 * 10240; time = 0.0398s; samplesPerSecond = 257532.3
12/20/2016 15:28:50:  Epoch[20 of 25]-Minibatch[ 901- 910, 82.92%]: CE.SM = 0.89719238 * 10240; Err = 0.29003906 * 10240; time = 0.0393s; samplesPerSecond = 260626.1
12/20/2016 15:28:50:  Epoch[20 of 25]-Minibatch[ 911- 920, 83.83%]: CE.SM = 0.90004272 * 10240; Err = 0.28750000 * 10240; time = 0.0391s; samplesPerSecond = 262174.2
12/20/2016 15:28:50:  Epoch[20 of 25]-Minibatch[ 921- 930, 84.74%]: CE.SM = 0.87553101 * 10240; Err = 0.28281250 * 10240; time = 0.0392s; samplesPerSecond = 260898.4
12/20/2016 15:28:50:  Epoch[20 of 25]-Minibatch[ 931- 940, 85.65%]: CE.SM = 0.87621460 * 10240; Err = 0.28818359 * 10240; time = 0.0393s; samplesPerSecond = 260394.2
12/20/2016 15:28:50:  Epoch[20 of 25]-Minibatch[ 941- 950, 86.56%]: CE.SM = 0.89990234 * 10240; Err = 0.28984375 * 10240; time = 0.0393s; samplesPerSecond = 260261.8
12/20/2016 15:28:51:  Epoch[20 of 25]-Minibatch[ 951- 960, 87.47%]: CE.SM = 0.89854736 * 10240; Err = 0.29541016 * 10240; time = 0.0394s; samplesPerSecond = 259891.9
12/20/2016 15:28:51:  Epoch[20 of 25]-Minibatch[ 961- 970, 88.38%]: CE.SM = 0.87083740 * 10240; Err = 0.28330078 * 10240; time = 0.0395s; samplesPerSecond = 259457.3
12/20/2016 15:28:51:  Epoch[20 of 25]-Minibatch[ 971- 980, 89.29%]: CE.SM = 0.89030151 * 10240; Err = 0.29033203 * 10240; time = 0.0391s; samplesPerSecond = 261745.3
12/20/2016 15:28:51:  Epoch[20 of 25]-Minibatch[ 981- 990, 90.21%]: CE.SM = 0.90744019 * 10240; Err = 0.28837891 * 10240; time = 0.0393s; samplesPerSecond = 260685.8
12/20/2016 15:28:51:  Epoch[20 of 25]-Minibatch[ 991-1000, 91.12%]: CE.SM = 0.88539429 * 10240; Err = 0.28916016 * 10240; time = 0.0393s; samplesPerSecond = 260599.6
12/20/2016 15:28:51:  Epoch[20 of 25]-Minibatch[1001-1010, 92.03%]: CE.SM = 0.88861084 * 10240; Err = 0.29140625 * 10240; time = 0.0388s; samplesPerSecond = 263897.1
12/20/2016 15:28:51:  Epoch[20 of 25]-Minibatch[1011-1020, 92.94%]: CE.SM = 0.89020996 * 10240; Err = 0.28925781 * 10240; time = 0.0392s; samplesPerSecond = 261131.2
12/20/2016 15:28:51:  Epoch[20 of 25]-Minibatch[1021-1030, 93.85%]: CE.SM = 0.87376709 * 10240; Err = 0.28281250 * 10240; time = 0.0390s; samplesPerSecond = 262416.1
12/20/2016 15:28:51:  Epoch[20 of 25]-Minibatch[1031-1040, 94.76%]: CE.SM = 0.88791504 * 10240; Err = 0.28662109 * 10240; time = 0.0394s; samplesPerSecond = 259931.5
12/20/2016 15:28:51:  Epoch[20 of 25]-Minibatch[1041-1050, 95.67%]: CE.SM = 0.90556641 * 10240; Err = 0.29853516 * 10240; time = 0.0390s; samplesPerSecond = 262860.7
12/20/2016 15:28:51:  Epoch[20 of 25]-Minibatch[1051-1060, 96.58%]: CE.SM = 0.89429321 * 10240; Err = 0.29345703 * 10240; time = 0.0393s; samplesPerSecond = 260573.1
12/20/2016 15:28:51:  Epoch[20 of 25]-Minibatch[1061-1070, 97.49%]: CE.SM = 0.86805420 * 10240; Err = 0.28251953 * 10240; time = 0.0394s; samplesPerSecond = 260050.3
12/20/2016 15:28:51:  Epoch[20 of 25]-Minibatch[1071-1080, 98.41%]: CE.SM = 0.86126099 * 10240; Err = 0.28232422 * 10240; time = 0.0444s; samplesPerSecond = 230583.9
12/20/2016 15:28:51:  Epoch[20 of 25]-Minibatch[1081-1090, 99.32%]: CE.SM = 0.88621826 * 10240; Err = 0.29306641 * 10240; time = 0.0415s; samplesPerSecond = 246883.8
12/20/2016 15:28:51: Finished Epoch[20 of 25]: [Training] CE.SM = 0.88565240 * 1124823; Err = 0.28875476 * 1124823; totalSamplesSeen = 22496460; learningRatePerSample = 7.8124998e-05; epochTime=4.56113s
12/20/2016 15:28:51: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn.20'

12/20/2016 15:28:51: Starting Epoch 21: learning rate per sample = 0.000078  effective momentum = 0.900000  momentum as time constant = 9719.0 samples

12/20/2016 15:28:51: Starting minibatch loop.
12/20/2016 15:28:51:  Epoch[21 of 25]-Minibatch[   1-  10, 0.91%]: CE.SM = 0.89799147 * 10240; Err = 0.29326172 * 10240; time = 0.0451s; samplesPerSecond = 227257.6
12/20/2016 15:28:51:  Epoch[21 of 25]-Minibatch[  11-  20, 1.82%]: CE.SM = 0.90229063 * 10240; Err = 0.29238281 * 10240; time = 0.0405s; samplesPerSecond = 252633.7
12/20/2016 15:28:51:  Epoch[21 of 25]-Minibatch[  21-  30, 2.73%]: CE.SM = 0.89083729 * 10240; Err = 0.29404297 * 10240; time = 0.0412s; samplesPerSecond = 248338.7
12/20/2016 15:28:51:  Epoch[21 of 25]-Minibatch[  31-  40, 3.64%]: CE.SM = 0.88381653 * 10240; Err = 0.28779297 * 10240; time = 0.0404s; samplesPerSecond = 253339.9
12/20/2016 15:28:52:  Epoch[21 of 25]-Minibatch[  41-  50, 4.56%]: CE.SM = 0.88101616 * 10240; Err = 0.28935547 * 10240; time = 0.0398s; samplesPerSecond = 257053.9
12/20/2016 15:28:52:  Epoch[21 of 25]-Minibatch[  51-  60, 5.47%]: CE.SM = 0.90134697 * 10240; Err = 0.29521484 * 10240; time = 0.0402s; samplesPerSecond = 254929.3
12/20/2016 15:28:52:  Epoch[21 of 25]-Minibatch[  61-  70, 6.38%]: CE.SM = 0.87561264 * 10240; Err = 0.29169922 * 10240; time = 0.0404s; samplesPerSecond = 253628.6
12/20/2016 15:28:52:  Epoch[21 of 25]-Minibatch[  71-  80, 7.29%]: CE.SM = 0.88606644 * 10240; Err = 0.29101562 * 10240; time = 0.0401s; samplesPerSecond = 255183.4
12/20/2016 15:28:52:  Epoch[21 of 25]-Minibatch[  81-  90, 8.20%]: CE.SM = 0.89414368 * 10240; Err = 0.29638672 * 10240; time = 0.0403s; samplesPerSecond = 254006.1
12/20/2016 15:28:52:  Epoch[21 of 25]-Minibatch[  91- 100, 9.11%]: CE.SM = 0.85797729 * 10240; Err = 0.27812500 * 10240; time = 0.0400s; samplesPerSecond = 255993.6
12/20/2016 15:28:52:  Epoch[21 of 25]-Minibatch[ 101- 110, 10.02%]: CE.SM = 0.88020325 * 10240; Err = 0.28554687 * 10240; time = 0.0398s; samplesPerSecond = 257273.5
12/20/2016 15:28:52:  Epoch[21 of 25]-Minibatch[ 111- 120, 10.93%]: CE.SM = 0.86856842 * 10240; Err = 0.28154297 * 10240; time = 0.0402s; samplesPerSecond = 254511.1
12/20/2016 15:28:52:  Epoch[21 of 25]-Minibatch[ 121- 130, 11.85%]: CE.SM = 0.89083633 * 10240; Err = 0.28740234 * 10240; time = 0.0407s; samplesPerSecond = 251887.9
12/20/2016 15:28:52:  Epoch[21 of 25]-Minibatch[ 131- 140, 12.76%]: CE.SM = 0.88090973 * 10240; Err = 0.29140625 * 10240; time = 0.0404s; samplesPerSecond = 253584.6
12/20/2016 15:28:52:  Epoch[21 of 25]-Minibatch[ 141- 150, 13.67%]: CE.SM = 0.88349762 * 10240; Err = 0.28574219 * 10240; time = 0.0407s; samplesPerSecond = 251510.5
12/20/2016 15:28:52:  Epoch[21 of 25]-Minibatch[ 151- 160, 14.58%]: CE.SM = 0.89589691 * 10240; Err = 0.29121094 * 10240; time = 0.0401s; samplesPerSecond = 255119.8
12/20/2016 15:28:52:  Epoch[21 of 25]-Minibatch[ 161- 170, 15.49%]: CE.SM = 0.90134125 * 10240; Err = 0.29218750 * 10240; time = 0.0408s; samplesPerSecond = 250894.3
12/20/2016 15:28:52:  Epoch[21 of 25]-Minibatch[ 171- 180, 16.40%]: CE.SM = 0.91218872 * 10240; Err = 0.29853516 * 10240; time = 0.0410s; samplesPerSecond = 250000.0
12/20/2016 15:28:52:  Epoch[21 of 25]-Minibatch[ 181- 190, 17.31%]: CE.SM = 0.86560822 * 10240; Err = 0.28710938 * 10240; time = 0.0400s; samplesPerSecond = 255795.4
12/20/2016 15:28:52:  Epoch[21 of 25]-Minibatch[ 191- 200, 18.22%]: CE.SM = 0.90583191 * 10240; Err = 0.30058594 * 10240; time = 0.0404s; samplesPerSecond = 253321.1
12/20/2016 15:28:52:  Epoch[21 of 25]-Minibatch[ 201- 210, 19.13%]: CE.SM = 0.89282379 * 10240; Err = 0.28671875 * 10240; time = 0.0405s; samplesPerSecond = 252671.1
12/20/2016 15:28:52:  Epoch[21 of 25]-Minibatch[ 211- 220, 20.05%]: CE.SM = 0.89232330 * 10240; Err = 0.28701172 * 10240; time = 0.0405s; samplesPerSecond = 253139.5
12/20/2016 15:28:52:  Epoch[21 of 25]-Minibatch[ 221- 230, 20.96%]: CE.SM = 0.86123505 * 10240; Err = 0.27861328 * 10240; time = 0.0400s; samplesPerSecond = 256262.7
12/20/2016 15:28:52:  Epoch[21 of 25]-Minibatch[ 231- 240, 21.87%]: CE.SM = 0.88764801 * 10240; Err = 0.28857422 * 10240; time = 0.0399s; samplesPerSecond = 256667.3
12/20/2016 15:28:52:  Epoch[21 of 25]-Minibatch[ 241- 250, 22.78%]: CE.SM = 0.87450562 * 10240; Err = 0.28310547 * 10240; time = 0.0400s; samplesPerSecond = 256243.4
12/20/2016 15:28:52:  Epoch[21 of 25]-Minibatch[ 251- 260, 23.69%]: CE.SM = 0.86626587 * 10240; Err = 0.28212891 * 10240; time = 0.0399s; samplesPerSecond = 256680.2
12/20/2016 15:28:52:  Epoch[21 of 25]-Minibatch[ 261- 270, 24.60%]: CE.SM = 0.87709503 * 10240; Err = 0.28896484 * 10240; time = 0.0397s; samplesPerSecond = 258207.7
12/20/2016 15:28:52:  Epoch[21 of 25]-Minibatch[ 271- 280, 25.51%]: CE.SM = 0.89419250 * 10240; Err = 0.29160156 * 10240; time = 0.0403s; samplesPerSecond = 254207.8
12/20/2016 15:28:52:  Epoch[21 of 25]-Minibatch[ 281- 290, 26.42%]: CE.SM = 0.89423065 * 10240; Err = 0.29970703 * 10240; time = 0.0396s; samplesPerSecond = 258533.6
12/20/2016 15:28:53:  Epoch[21 of 25]-Minibatch[ 291- 300, 27.33%]: CE.SM = 0.88347778 * 10240; Err = 0.29521484 * 10240; time = 0.0400s; samplesPerSecond = 255744.3
12/20/2016 15:28:53:  Epoch[21 of 25]-Minibatch[ 301- 310, 28.25%]: CE.SM = 0.90356445 * 10240; Err = 0.29345703 * 10240; time = 0.0405s; samplesPerSecond = 253014.4
12/20/2016 15:28:53:  Epoch[21 of 25]-Minibatch[ 311- 320, 29.16%]: CE.SM = 0.87642822 * 10240; Err = 0.28359375 * 10240; time = 0.0402s; samplesPerSecond = 254859.5
12/20/2016 15:28:53:  Epoch[21 of 25]-Minibatch[ 321- 330, 30.07%]: CE.SM = 0.87874756 * 10240; Err = 0.29248047 * 10240; time = 0.0402s; samplesPerSecond = 254929.3
12/20/2016 15:28:53:  Epoch[21 of 25]-Minibatch[ 331- 340, 30.98%]: CE.SM = 0.89793701 * 10240; Err = 0.28876953 * 10240; time = 0.0408s; samplesPerSecond = 250992.7
12/20/2016 15:28:53:  Epoch[21 of 25]-Minibatch[ 341- 350, 31.89%]: CE.SM = 0.88714294 * 10240; Err = 0.28857422 * 10240; time = 0.0404s; samplesPerSecond = 253565.8
12/20/2016 15:28:53:  Epoch[21 of 25]-Minibatch[ 351- 360, 32.80%]: CE.SM = 0.88865662 * 10240; Err = 0.28837891 * 10240; time = 0.0406s; samplesPerSecond = 252316.2
12/20/2016 15:28:53:  Epoch[21 of 25]-Minibatch[ 361- 370, 33.71%]: CE.SM = 0.88202209 * 10240; Err = 0.29033203 * 10240; time = 0.0404s; samplesPerSecond = 253189.6
12/20/2016 15:28:53:  Epoch[21 of 25]-Minibatch[ 371- 380, 34.62%]: CE.SM = 0.89221497 * 10240; Err = 0.28769531 * 10240; time = 0.0407s; samplesPerSecond = 251405.6
12/20/2016 15:28:53:  Epoch[21 of 25]-Minibatch[ 381- 390, 35.54%]: CE.SM = 0.89779663 * 10240; Err = 0.29160156 * 10240; time = 0.0401s; samplesPerSecond = 255138.9
12/20/2016 15:28:53:  Epoch[21 of 25]-Minibatch[ 391- 400, 36.45%]: CE.SM = 0.88442993 * 10240; Err = 0.29082031 * 10240; time = 0.0410s; samplesPerSecond = 249853.6
12/20/2016 15:28:53:  Epoch[21 of 25]-Minibatch[ 401- 410, 37.36%]: CE.SM = 0.89200745 * 10240; Err = 0.29375000 * 10240; time = 0.0407s; samplesPerSecond = 251405.6
12/20/2016 15:28:53:  Epoch[21 of 25]-Minibatch[ 411- 420, 38.27%]: CE.SM = 0.88918457 * 10240; Err = 0.28261719 * 10240; time = 0.0402s; samplesPerSecond = 255030.9
12/20/2016 15:28:53:  Epoch[21 of 25]-Minibatch[ 421- 430, 39.18%]: CE.SM = 0.89075317 * 10240; Err = 0.28808594 * 10240; time = 0.0409s; samplesPerSecond = 250667.1
12/20/2016 15:28:53:  Epoch[21 of 25]-Minibatch[ 431- 440, 40.09%]: CE.SM = 0.88851318 * 10240; Err = 0.28359375 * 10240; time = 0.0402s; samplesPerSecond = 254447.9
12/20/2016 15:28:53:  Epoch[21 of 25]-Minibatch[ 441- 450, 41.00%]: CE.SM = 0.89571533 * 10240; Err = 0.28818359 * 10240; time = 0.0405s; samplesPerSecond = 252902.0
12/20/2016 15:28:53:  Epoch[21 of 25]-Minibatch[ 451- 460, 41.91%]: CE.SM = 0.88857422 * 10240; Err = 0.28847656 * 10240; time = 0.0399s; samplesPerSecond = 256545.2
12/20/2016 15:28:53:  Epoch[21 of 25]-Minibatch[ 461- 470, 42.82%]: CE.SM = 0.88558655 * 10240; Err = 0.28535156 * 10240; time = 0.0405s; samplesPerSecond = 253039.4
12/20/2016 15:28:53:  Epoch[21 of 25]-Minibatch[ 471- 480, 43.74%]: CE.SM = 0.86889038 * 10240; Err = 0.28642578 * 10240; time = 0.0412s; samplesPerSecond = 248260.5
12/20/2016 15:28:53:  Epoch[21 of 25]-Minibatch[ 481- 490, 44.65%]: CE.SM = 0.87821655 * 10240; Err = 0.28330078 * 10240; time = 0.0408s; samplesPerSecond = 250814.4
12/20/2016 15:28:53:  Epoch[21 of 25]-Minibatch[ 491- 500, 45.56%]: CE.SM = 0.88918457 * 10240; Err = 0.29472656 * 10240; time = 0.0403s; samplesPerSecond = 254119.5
12/20/2016 15:28:53:  Epoch[21 of 25]-Minibatch[ 501- 510, 46.47%]: CE.SM = 0.90068359 * 10240; Err = 0.29550781 * 10240; time = 0.0409s; samplesPerSecond = 250605.7
12/20/2016 15:28:53:  Epoch[21 of 25]-Minibatch[ 511- 520, 47.38%]: CE.SM = 0.86630859 * 10240; Err = 0.28486328 * 10240; time = 0.0404s; samplesPerSecond = 253660.0
12/20/2016 15:28:53:  Epoch[21 of 25]-Minibatch[ 521- 530, 48.29%]: CE.SM = 0.87844849 * 10240; Err = 0.28828125 * 10240; time = 0.0410s; samplesPerSecond = 250018.3
12/20/2016 15:28:54:  Epoch[21 of 25]-Minibatch[ 531- 540, 49.20%]: CE.SM = 0.88147278 * 10240; Err = 0.28652344 * 10240; time = 0.0402s; samplesPerSecond = 254517.4
12/20/2016 15:28:54:  Epoch[21 of 25]-Minibatch[ 541- 550, 50.11%]: CE.SM = 0.86651001 * 10240; Err = 0.28505859 * 10240; time = 0.0403s; samplesPerSecond = 254056.5
12/20/2016 15:28:54:  Epoch[21 of 25]-Minibatch[ 551- 560, 51.03%]: CE.SM = 0.88726807 * 10240; Err = 0.29121094 * 10240; time = 0.0408s; samplesPerSecond = 250992.7
12/20/2016 15:28:54:  Epoch[21 of 25]-Minibatch[ 561- 570, 51.94%]: CE.SM = 0.86210937 * 10240; Err = 0.28408203 * 10240; time = 0.0433s; samplesPerSecond = 236555.2
12/20/2016 15:28:54:  Epoch[21 of 25]-Minibatch[ 571- 580, 52.85%]: CE.SM = 0.88630676 * 10240; Err = 0.29423828 * 10240; time = 0.0408s; samplesPerSecond = 251091.2
12/20/2016 15:28:54:  Epoch[21 of 25]-Minibatch[ 581- 590, 53.76%]: CE.SM = 0.88088379 * 10240; Err = 0.28789063 * 10240; time = 0.0403s; samplesPerSecond = 253817.2
12/20/2016 15:28:54:  Epoch[21 of 25]-Minibatch[ 591- 600, 54.67%]: CE.SM = 0.87077637 * 10240; Err = 0.28779297 * 10240; time = 0.0400s; samplesPerSecond = 255731.5
12/20/2016 15:28:54:  Epoch[21 of 25]-Minibatch[ 601- 610, 55.58%]: CE.SM = 0.88183594 * 10240; Err = 0.28574219 * 10240; time = 0.0402s; samplesPerSecond = 254637.7
12/20/2016 15:28:54:  Epoch[21 of 25]-Minibatch[ 611- 620, 56.49%]: CE.SM = 0.87261963 * 10240; Err = 0.28466797 * 10240; time = 0.0409s; samplesPerSecond = 250287.2
12/20/2016 15:28:54:  Epoch[21 of 25]-Minibatch[ 621- 630, 57.40%]: CE.SM = 0.88535767 * 10240; Err = 0.29023437 * 10240; time = 0.0406s; samplesPerSecond = 251925.1
12/20/2016 15:28:54:  Epoch[21 of 25]-Minibatch[ 631- 640, 58.31%]: CE.SM = 0.90818481 * 10240; Err = 0.29316406 * 10240; time = 0.0402s; samplesPerSecond = 254492.1
12/20/2016 15:28:54:  Epoch[21 of 25]-Minibatch[ 641- 650, 59.23%]: CE.SM = 0.86754150 * 10240; Err = 0.28183594 * 10240; time = 0.0407s; samplesPerSecond = 251615.6
12/20/2016 15:28:54:  Epoch[21 of 25]-Minibatch[ 651- 660, 60.14%]: CE.SM = 0.87567749 * 10240; Err = 0.28115234 * 10240; time = 0.0409s; samplesPerSecond = 250366.7
12/20/2016 15:28:54:  Epoch[21 of 25]-Minibatch[ 661- 670, 61.05%]: CE.SM = 0.88480835 * 10240; Err = 0.29189453 * 10240; time = 0.0404s; samplesPerSecond = 253710.3
12/20/2016 15:28:54:  Epoch[21 of 25]-Minibatch[ 671- 680, 61.96%]: CE.SM = 0.89306641 * 10240; Err = 0.29052734 * 10240; time = 0.0408s; samplesPerSecond = 250710.0
12/20/2016 15:28:54:  Epoch[21 of 25]-Minibatch[ 681- 690, 62.87%]: CE.SM = 0.89364014 * 10240; Err = 0.29296875 * 10240; time = 0.0404s; samplesPerSecond = 253509.3
12/20/2016 15:28:54:  Epoch[21 of 25]-Minibatch[ 691- 700, 63.78%]: CE.SM = 0.89699097 * 10240; Err = 0.29228516 * 10240; time = 0.0401s; samplesPerSecond = 255183.4
12/20/2016 15:28:54:  Epoch[21 of 25]-Minibatch[ 701- 710, 64.69%]: CE.SM = 0.88986206 * 10240; Err = 0.28701172 * 10240; time = 0.0409s; samplesPerSecond = 250330.0
12/20/2016 15:28:54:  Epoch[21 of 25]-Minibatch[ 711- 720, 65.60%]: CE.SM = 0.89188843 * 10240; Err = 0.28710938 * 10240; time = 0.0400s; samplesPerSecond = 255725.1
12/20/2016 15:28:54:  Epoch[21 of 25]-Minibatch[ 721- 730, 66.51%]: CE.SM = 0.88198242 * 10240; Err = 0.28173828 * 10240; time = 0.0398s; samplesPerSecond = 257467.6
12/20/2016 15:28:54:  Epoch[21 of 25]-Minibatch[ 731- 740, 67.43%]: CE.SM = 0.89028320 * 10240; Err = 0.29462891 * 10240; time = 0.0394s; samplesPerSecond = 259944.7
12/20/2016 15:28:54:  Epoch[21 of 25]-Minibatch[ 741- 750, 68.34%]: CE.SM = 0.87610474 * 10240; Err = 0.28378906 * 10240; time = 0.0390s; samplesPerSecond = 262274.9
12/20/2016 15:28:54:  Epoch[21 of 25]-Minibatch[ 751- 760, 69.25%]: CE.SM = 0.86769409 * 10240; Err = 0.28037109 * 10240; time = 0.0393s; samplesPerSecond = 260520.0
12/20/2016 15:28:54:  Epoch[21 of 25]-Minibatch[ 761- 770, 70.16%]: CE.SM = 0.87484741 * 10240; Err = 0.28496094 * 10240; time = 0.0395s; samplesPerSecond = 259089.6
12/20/2016 15:28:54:  Epoch[21 of 25]-Minibatch[ 771- 780, 71.07%]: CE.SM = 0.87394409 * 10240; Err = 0.28408203 * 10240; time = 0.0397s; samplesPerSecond = 258214.2
12/20/2016 15:28:55:  Epoch[21 of 25]-Minibatch[ 781- 790, 71.98%]: CE.SM = 0.88486328 * 10240; Err = 0.29453125 * 10240; time = 0.0398s; samplesPerSecond = 257034.6
12/20/2016 15:28:55:  Epoch[21 of 25]-Minibatch[ 791- 800, 72.89%]: CE.SM = 0.87249146 * 10240; Err = 0.28496094 * 10240; time = 0.0408s; samplesPerSecond = 250863.6
12/20/2016 15:28:55:  Epoch[21 of 25]-Minibatch[ 801- 810, 73.80%]: CE.SM = 0.88563843 * 10240; Err = 0.28935547 * 10240; time = 0.0403s; samplesPerSecond = 254025.0
12/20/2016 15:28:55:  Epoch[21 of 25]-Minibatch[ 811- 820, 74.72%]: CE.SM = 0.86105347 * 10240; Err = 0.28486328 * 10240; time = 0.0397s; samplesPerSecond = 258123.1
12/20/2016 15:28:55:  Epoch[21 of 25]-Minibatch[ 821- 830, 75.63%]: CE.SM = 0.86783447 * 10240; Err = 0.28261719 * 10240; time = 0.0393s; samplesPerSecond = 260460.4
12/20/2016 15:28:55:  Epoch[21 of 25]-Minibatch[ 831- 840, 76.54%]: CE.SM = 0.87417603 * 10240; Err = 0.28593750 * 10240; time = 0.0398s; samplesPerSecond = 257558.2
12/20/2016 15:28:55:  Epoch[21 of 25]-Minibatch[ 841- 850, 77.45%]: CE.SM = 0.87366943 * 10240; Err = 0.28105469 * 10240; time = 0.0393s; samplesPerSecond = 260885.1
12/20/2016 15:28:55:  Epoch[21 of 25]-Minibatch[ 851- 860, 78.36%]: CE.SM = 0.89128418 * 10240; Err = 0.28417969 * 10240; time = 0.0395s; samplesPerSecond = 258945.5
12/20/2016 15:28:55:  Epoch[21 of 25]-Minibatch[ 861- 870, 79.27%]: CE.SM = 0.89707642 * 10240; Err = 0.28925781 * 10240; time = 0.0391s; samplesPerSecond = 261731.9
12/20/2016 15:28:55:  Epoch[21 of 25]-Minibatch[ 871- 880, 80.18%]: CE.SM = 0.88634644 * 10240; Err = 0.28945312 * 10240; time = 0.0396s; samplesPerSecond = 258696.9
12/20/2016 15:28:55:  Epoch[21 of 25]-Minibatch[ 881- 890, 81.09%]: CE.SM = 0.88139648 * 10240; Err = 0.28212891 * 10240; time = 0.0389s; samplesPerSecond = 263401.6
12/20/2016 15:28:55:  Epoch[21 of 25]-Minibatch[ 891- 900, 82.00%]: CE.SM = 0.87524414 * 10240; Err = 0.28544922 * 10240; time = 0.0391s; samplesPerSecond = 261785.5
12/20/2016 15:28:55:  Epoch[21 of 25]-Minibatch[ 901- 910, 82.92%]: CE.SM = 0.90508423 * 10240; Err = 0.29443359 * 10240; time = 0.0392s; samplesPerSecond = 261411.2
12/20/2016 15:28:55:  Epoch[21 of 25]-Minibatch[ 911- 920, 83.83%]: CE.SM = 0.88687134 * 10240; Err = 0.28847656 * 10240; time = 0.0391s; samplesPerSecond = 261792.2
12/20/2016 15:28:55:  Epoch[21 of 25]-Minibatch[ 921- 930, 84.74%]: CE.SM = 0.89563599 * 10240; Err = 0.28701172 * 10240; time = 0.0392s; samplesPerSecond = 261484.6
12/20/2016 15:28:55:  Epoch[21 of 25]-Minibatch[ 931- 940, 85.65%]: CE.SM = 0.88105469 * 10240; Err = 0.28779297 * 10240; time = 0.0394s; samplesPerSecond = 259694.1
12/20/2016 15:28:55:  Epoch[21 of 25]-Minibatch[ 941- 950, 86.56%]: CE.SM = 0.87406006 * 10240; Err = 0.28701172 * 10240; time = 0.0401s; samplesPerSecond = 255463.5
12/20/2016 15:28:55:  Epoch[21 of 25]-Minibatch[ 951- 960, 87.47%]: CE.SM = 0.86514282 * 10240; Err = 0.28330078 * 10240; time = 0.0408s; samplesPerSecond = 251226.7
12/20/2016 15:28:55:  Epoch[21 of 25]-Minibatch[ 961- 970, 88.38%]: CE.SM = 0.88926392 * 10240; Err = 0.29013672 * 10240; time = 0.0403s; samplesPerSecond = 254132.1
12/20/2016 15:28:55:  Epoch[21 of 25]-Minibatch[ 971- 980, 89.29%]: CE.SM = 0.88462524 * 10240; Err = 0.29101562 * 10240; time = 0.0400s; samplesPerSecond = 255942.4
12/20/2016 15:28:55:  Epoch[21 of 25]-Minibatch[ 981- 990, 90.21%]: CE.SM = 0.87756348 * 10240; Err = 0.29072266 * 10240; time = 0.0399s; samplesPerSecond = 256770.3
12/20/2016 15:28:55:  Epoch[21 of 25]-Minibatch[ 991-1000, 91.12%]: CE.SM = 0.88082886 * 10240; Err = 0.28564453 * 10240; time = 0.0400s; samplesPerSecond = 255776.2
12/20/2016 15:28:55:  Epoch[21 of 25]-Minibatch[1001-1010, 92.03%]: CE.SM = 0.87758789 * 10240; Err = 0.28847656 * 10240; time = 0.0399s; samplesPerSecond = 256802.5
12/20/2016 15:28:55:  Epoch[21 of 25]-Minibatch[1011-1020, 92.94%]: CE.SM = 0.88175659 * 10240; Err = 0.29121094 * 10240; time = 0.0400s; samplesPerSecond = 255852.9
12/20/2016 15:28:55:  Epoch[21 of 25]-Minibatch[1021-1030, 93.85%]: CE.SM = 0.86475830 * 10240; Err = 0.27666016 * 10240; time = 0.0402s; samplesPerSecond = 254840.5
12/20/2016 15:28:56:  Epoch[21 of 25]-Minibatch[1031-1040, 94.76%]: CE.SM = 0.88187256 * 10240; Err = 0.28261719 * 10240; time = 0.0395s; samplesPerSecond = 258958.6
12/20/2016 15:28:56:  Epoch[21 of 25]-Minibatch[1041-1050, 95.67%]: CE.SM = 0.88224487 * 10240; Err = 0.29121094 * 10240; time = 0.0404s; samplesPerSecond = 253628.6
12/20/2016 15:28:56:  Epoch[21 of 25]-Minibatch[1051-1060, 96.58%]: CE.SM = 0.87214355 * 10240; Err = 0.28701172 * 10240; time = 0.0401s; samplesPerSecond = 255444.4
12/20/2016 15:28:56:  Epoch[21 of 25]-Minibatch[1061-1070, 97.49%]: CE.SM = 0.88802490 * 10240; Err = 0.29091797 * 10240; time = 0.0442s; samplesPerSecond = 231878.8
12/20/2016 15:28:56:  Epoch[21 of 25]-Minibatch[1071-1080, 98.41%]: CE.SM = 0.88450317 * 10240; Err = 0.29257813 * 10240; time = 0.0491s; samplesPerSecond = 208498.8
12/20/2016 15:28:56:  Epoch[21 of 25]-Minibatch[1081-1090, 99.32%]: CE.SM = 0.89425659 * 10240; Err = 0.29511719 * 10240; time = 0.0400s; samplesPerSecond = 256217.8
12/20/2016 15:28:56: Finished Epoch[21 of 25]: [Training] CE.SM = 0.88385328 * 1124823; Err = 0.28835737 * 1124823; totalSamplesSeen = 23621283; learningRatePerSample = 7.8124998e-05; epochTime=4.62536s
12/20/2016 15:28:56: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn.21'

12/20/2016 15:28:56: Starting Epoch 22: learning rate per sample = 0.000078  effective momentum = 0.900000  momentum as time constant = 9719.0 samples

12/20/2016 15:28:56: Starting minibatch loop.
12/20/2016 15:28:56:  Epoch[22 of 25]-Minibatch[   1-  10, 0.91%]: CE.SM = 0.88326931 * 10240; Err = 0.28906250 * 10240; time = 0.0445s; samplesPerSecond = 229936.7
12/20/2016 15:28:56:  Epoch[22 of 25]-Minibatch[  11-  20, 1.82%]: CE.SM = 0.88533039 * 10240; Err = 0.29130859 * 10240; time = 0.0412s; samplesPerSecond = 248483.4
12/20/2016 15:28:56:  Epoch[22 of 25]-Minibatch[  21-  30, 2.73%]: CE.SM = 0.86798325 * 10240; Err = 0.28544922 * 10240; time = 0.0452s; samplesPerSecond = 226709.2
12/20/2016 15:28:56:  Epoch[22 of 25]-Minibatch[  31-  40, 3.64%]: CE.SM = 0.88669281 * 10240; Err = 0.28535156 * 10240; time = 0.0403s; samplesPerSecond = 253779.4
12/20/2016 15:28:56:  Epoch[22 of 25]-Minibatch[  41-  50, 4.56%]: CE.SM = 0.89012794 * 10240; Err = 0.28876953 * 10240; time = 0.0400s; samplesPerSecond = 256083.2
12/20/2016 15:28:56:  Epoch[22 of 25]-Minibatch[  51-  60, 5.47%]: CE.SM = 0.90409966 * 10240; Err = 0.29707031 * 10240; time = 0.0403s; samplesPerSecond = 253867.5
12/20/2016 15:28:56:  Epoch[22 of 25]-Minibatch[  61-  70, 6.38%]: CE.SM = 0.87449036 * 10240; Err = 0.28818359 * 10240; time = 0.0406s; samplesPerSecond = 252322.4
12/20/2016 15:28:56:  Epoch[22 of 25]-Minibatch[  71-  80, 7.29%]: CE.SM = 0.86446991 * 10240; Err = 0.28505859 * 10240; time = 0.0407s; samplesPerSecond = 251294.5
12/20/2016 15:28:56:  Epoch[22 of 25]-Minibatch[  81-  90, 8.20%]: CE.SM = 0.88067169 * 10240; Err = 0.28554687 * 10240; time = 0.0454s; samplesPerSecond = 225565.6
12/20/2016 15:28:56:  Epoch[22 of 25]-Minibatch[  91- 100, 9.11%]: CE.SM = 0.87791443 * 10240; Err = 0.28857422 * 10240; time = 0.0411s; samplesPerSecond = 249130.2
12/20/2016 15:28:56:  Epoch[22 of 25]-Minibatch[ 101- 110, 10.02%]: CE.SM = 0.87764893 * 10240; Err = 0.28466797 * 10240; time = 0.0410s; samplesPerSecond = 249676.9
12/20/2016 15:28:57:  Epoch[22 of 25]-Minibatch[ 111- 120, 10.93%]: CE.SM = 0.90743790 * 10240; Err = 0.29560547 * 10240; time = 0.0398s; samplesPerSecond = 257060.4
12/20/2016 15:28:57:  Epoch[22 of 25]-Minibatch[ 121- 130, 11.85%]: CE.SM = 0.85701447 * 10240; Err = 0.27890625 * 10240; time = 0.0401s; samplesPerSecond = 255069.0
12/20/2016 15:28:57:  Epoch[22 of 25]-Minibatch[ 131- 140, 12.76%]: CE.SM = 0.89113007 * 10240; Err = 0.29013672 * 10240; time = 0.0403s; samplesPerSecond = 253968.3
12/20/2016 15:28:57:  Epoch[22 of 25]-Minibatch[ 141- 150, 13.67%]: CE.SM = 0.88912888 * 10240; Err = 0.29003906 * 10240; time = 0.0401s; samplesPerSecond = 255266.1
12/20/2016 15:28:57:  Epoch[22 of 25]-Minibatch[ 151- 160, 14.58%]: CE.SM = 0.87105713 * 10240; Err = 0.28808594 * 10240; time = 0.0400s; samplesPerSecond = 255846.5
12/20/2016 15:28:57:  Epoch[22 of 25]-Minibatch[ 161- 170, 15.49%]: CE.SM = 0.88159790 * 10240; Err = 0.29101562 * 10240; time = 0.0396s; samplesPerSecond = 258585.9
12/20/2016 15:28:57:  Epoch[22 of 25]-Minibatch[ 171- 180, 16.40%]: CE.SM = 0.88873291 * 10240; Err = 0.28798828 * 10240; time = 0.0402s; samplesPerSecond = 254707.4
12/20/2016 15:28:57:  Epoch[22 of 25]-Minibatch[ 181- 190, 17.31%]: CE.SM = 0.88604126 * 10240; Err = 0.28681641 * 10240; time = 0.0433s; samplesPerSecond = 236653.6
12/20/2016 15:28:57:  Epoch[22 of 25]-Minibatch[ 191- 200, 18.22%]: CE.SM = 0.87510376 * 10240; Err = 0.28837891 * 10240; time = 0.0442s; samplesPerSecond = 231569.4
12/20/2016 15:28:57:  Epoch[22 of 25]-Minibatch[ 201- 210, 19.13%]: CE.SM = 0.90517731 * 10240; Err = 0.29658203 * 10240; time = 0.0422s; samplesPerSecond = 242867.0
12/20/2016 15:28:57:  Epoch[22 of 25]-Minibatch[ 211- 220, 20.05%]: CE.SM = 0.88616943 * 10240; Err = 0.29296875 * 10240; time = 0.0441s; samplesPerSecond = 232131.1
12/20/2016 15:28:57:  Epoch[22 of 25]-Minibatch[ 221- 230, 20.96%]: CE.SM = 0.87748260 * 10240; Err = 0.28554687 * 10240; time = 0.0406s; samplesPerSecond = 252490.4
12/20/2016 15:28:57:  Epoch[22 of 25]-Minibatch[ 231- 240, 21.87%]: CE.SM = 0.89553375 * 10240; Err = 0.28964844 * 10240; time = 0.0402s; samplesPerSecond = 254631.4
12/20/2016 15:28:57:  Epoch[22 of 25]-Minibatch[ 241- 250, 22.78%]: CE.SM = 0.86972809 * 10240; Err = 0.28359375 * 10240; time = 0.0407s; samplesPerSecond = 251541.4
12/20/2016 15:28:57:  Epoch[22 of 25]-Minibatch[ 251- 260, 23.69%]: CE.SM = 0.87246246 * 10240; Err = 0.28164062 * 10240; time = 0.0417s; samplesPerSecond = 245657.8
12/20/2016 15:28:57:  Epoch[22 of 25]-Minibatch[ 261- 270, 24.60%]: CE.SM = 0.89271393 * 10240; Err = 0.28554687 * 10240; time = 0.0413s; samplesPerSecond = 247726.0
12/20/2016 15:28:57:  Epoch[22 of 25]-Minibatch[ 271- 280, 25.51%]: CE.SM = 0.87055054 * 10240; Err = 0.28886719 * 10240; time = 0.0407s; samplesPerSecond = 251461.1
12/20/2016 15:28:57:  Epoch[22 of 25]-Minibatch[ 281- 290, 26.42%]: CE.SM = 0.89481354 * 10240; Err = 0.28925781 * 10240; time = 0.0407s; samplesPerSecond = 251584.7
12/20/2016 15:28:57:  Epoch[22 of 25]-Minibatch[ 291- 300, 27.33%]: CE.SM = 0.86885986 * 10240; Err = 0.28457031 * 10240; time = 0.0409s; samplesPerSecond = 250648.7
12/20/2016 15:28:57:  Epoch[22 of 25]-Minibatch[ 301- 310, 28.25%]: CE.SM = 0.90033264 * 10240; Err = 0.29335937 * 10240; time = 0.0404s; samplesPerSecond = 253158.3
12/20/2016 15:28:57:  Epoch[22 of 25]-Minibatch[ 311- 320, 29.16%]: CE.SM = 0.87864990 * 10240; Err = 0.29228516 * 10240; time = 0.0405s; samplesPerSecond = 252671.1
12/20/2016 15:28:57:  Epoch[22 of 25]-Minibatch[ 321- 330, 30.07%]: CE.SM = 0.87808838 * 10240; Err = 0.28476563 * 10240; time = 0.0413s; samplesPerSecond = 247833.9
12/20/2016 15:28:57:  Epoch[22 of 25]-Minibatch[ 331- 340, 30.98%]: CE.SM = 0.87539673 * 10240; Err = 0.28818359 * 10240; time = 0.0403s; samplesPerSecond = 254359.4
12/20/2016 15:28:57:  Epoch[22 of 25]-Minibatch[ 341- 350, 31.89%]: CE.SM = 0.87786865 * 10240; Err = 0.28750000 * 10240; time = 0.0409s; samplesPerSecond = 250268.8
12/20/2016 15:28:58:  Epoch[22 of 25]-Minibatch[ 351- 360, 32.80%]: CE.SM = 0.89135132 * 10240; Err = 0.28759766 * 10240; time = 0.0403s; samplesPerSecond = 254340.4
12/20/2016 15:28:58:  Epoch[22 of 25]-Minibatch[ 361- 370, 33.71%]: CE.SM = 0.88810120 * 10240; Err = 0.28955078 * 10240; time = 0.0409s; samplesPerSecond = 250581.2
12/20/2016 15:28:58:  Epoch[22 of 25]-Minibatch[ 371- 380, 34.62%]: CE.SM = 0.87589111 * 10240; Err = 0.28486328 * 10240; time = 0.0401s; samplesPerSecond = 255450.8
12/20/2016 15:28:58:  Epoch[22 of 25]-Minibatch[ 381- 390, 35.54%]: CE.SM = 0.88117371 * 10240; Err = 0.28750000 * 10240; time = 0.0396s; samplesPerSecond = 258338.0
12/20/2016 15:28:58:  Epoch[22 of 25]-Minibatch[ 391- 400, 36.45%]: CE.SM = 0.89205017 * 10240; Err = 0.28935547 * 10240; time = 0.0404s; samplesPerSecond = 253490.4
12/20/2016 15:28:58:  Epoch[22 of 25]-Minibatch[ 401- 410, 37.36%]: CE.SM = 0.88506775 * 10240; Err = 0.29082031 * 10240; time = 0.0402s; samplesPerSecond = 254663.0
12/20/2016 15:28:58:  Epoch[22 of 25]-Minibatch[ 411- 420, 38.27%]: CE.SM = 0.89814453 * 10240; Err = 0.29384766 * 10240; time = 0.0403s; samplesPerSecond = 253798.3
12/20/2016 15:28:58:  Epoch[22 of 25]-Minibatch[ 421- 430, 39.18%]: CE.SM = 0.88552246 * 10240; Err = 0.28925781 * 10240; time = 0.0398s; samplesPerSecond = 257383.4
12/20/2016 15:28:58:  Epoch[22 of 25]-Minibatch[ 431- 440, 40.09%]: CE.SM = 0.89413147 * 10240; Err = 0.29140625 * 10240; time = 0.0403s; samplesPerSecond = 254220.5
12/20/2016 15:28:58:  Epoch[22 of 25]-Minibatch[ 441- 450, 41.00%]: CE.SM = 0.87808838 * 10240; Err = 0.28603516 * 10240; time = 0.0410s; samplesPerSecond = 249798.7
12/20/2016 15:28:58:  Epoch[22 of 25]-Minibatch[ 451- 460, 41.91%]: CE.SM = 0.88217163 * 10240; Err = 0.29091797 * 10240; time = 0.0403s; samplesPerSecond = 254144.7
12/20/2016 15:28:58:  Epoch[22 of 25]-Minibatch[ 461- 470, 42.82%]: CE.SM = 0.90577393 * 10240; Err = 0.29384766 * 10240; time = 0.0404s; samplesPerSecond = 253220.9
12/20/2016 15:28:58:  Epoch[22 of 25]-Minibatch[ 471- 480, 43.74%]: CE.SM = 0.88435059 * 10240; Err = 0.28623047 * 10240; time = 0.0411s; samplesPerSecond = 249051.5
12/20/2016 15:28:58:  Epoch[22 of 25]-Minibatch[ 481- 490, 44.65%]: CE.SM = 0.86630554 * 10240; Err = 0.28867188 * 10240; time = 0.0407s; samplesPerSecond = 251689.8
12/20/2016 15:28:58:  Epoch[22 of 25]-Minibatch[ 491- 500, 45.56%]: CE.SM = 0.89235229 * 10240; Err = 0.29101562 * 10240; time = 0.0403s; samplesPerSecond = 254006.1
12/20/2016 15:28:58:  Epoch[22 of 25]-Minibatch[ 501- 510, 46.47%]: CE.SM = 0.88765564 * 10240; Err = 0.28662109 * 10240; time = 0.0403s; samplesPerSecond = 254043.9
12/20/2016 15:28:58:  Epoch[22 of 25]-Minibatch[ 511- 520, 47.38%]: CE.SM = 0.88115540 * 10240; Err = 0.28681641 * 10240; time = 0.0400s; samplesPerSecond = 256025.6
12/20/2016 15:28:58:  Epoch[22 of 25]-Minibatch[ 521- 530, 48.29%]: CE.SM = 0.87304993 * 10240; Err = 0.28359375 * 10240; time = 0.0404s; samplesPerSecond = 253735.4
12/20/2016 15:28:58:  Epoch[22 of 25]-Minibatch[ 531- 540, 49.20%]: CE.SM = 0.87439575 * 10240; Err = 0.28691406 * 10240; time = 0.0398s; samplesPerSecond = 257228.3
12/20/2016 15:28:58:  Epoch[22 of 25]-Minibatch[ 541- 550, 50.11%]: CE.SM = 0.87151184 * 10240; Err = 0.28164062 * 10240; time = 0.0404s; samplesPerSecond = 253521.8
12/20/2016 15:28:58:  Epoch[22 of 25]-Minibatch[ 551- 560, 51.03%]: CE.SM = 0.88522644 * 10240; Err = 0.28994141 * 10240; time = 0.0405s; samplesPerSecond = 252720.9
12/20/2016 15:28:58:  Epoch[22 of 25]-Minibatch[ 561- 570, 51.94%]: CE.SM = 0.89984436 * 10240; Err = 0.28896484 * 10240; time = 0.0399s; samplesPerSecond = 256957.2
12/20/2016 15:28:58:  Epoch[22 of 25]-Minibatch[ 571- 580, 52.85%]: CE.SM = 0.89393921 * 10240; Err = 0.28925781 * 10240; time = 0.0403s; samplesPerSecond = 253848.6
12/20/2016 15:28:58:  Epoch[22 of 25]-Minibatch[ 581- 590, 53.76%]: CE.SM = 0.91099243 * 10240; Err = 0.29833984 * 10240; time = 0.0405s; samplesPerSecond = 253051.9
12/20/2016 15:28:58:  Epoch[22 of 25]-Minibatch[ 591- 600, 54.67%]: CE.SM = 0.86447144 * 10240; Err = 0.28466797 * 10240; time = 0.0401s; samplesPerSecond = 255623.0
12/20/2016 15:28:59:  Epoch[22 of 25]-Minibatch[ 601- 610, 55.58%]: CE.SM = 0.90496216 * 10240; Err = 0.29160156 * 10240; time = 0.0397s; samplesPerSecond = 257895.5
12/20/2016 15:28:59:  Epoch[22 of 25]-Minibatch[ 611- 620, 56.49%]: CE.SM = 0.87996216 * 10240; Err = 0.28398438 * 10240; time = 0.0402s; samplesPerSecond = 254574.4
12/20/2016 15:28:59:  Epoch[22 of 25]-Minibatch[ 621- 630, 57.40%]: CE.SM = 0.87733765 * 10240; Err = 0.28115234 * 10240; time = 0.0405s; samplesPerSecond = 252702.2
12/20/2016 15:28:59:  Epoch[22 of 25]-Minibatch[ 631- 640, 58.31%]: CE.SM = 0.85890503 * 10240; Err = 0.28212891 * 10240; time = 0.0404s; samplesPerSecond = 253528.1
12/20/2016 15:28:59:  Epoch[22 of 25]-Minibatch[ 641- 650, 59.23%]: CE.SM = 0.88942871 * 10240; Err = 0.28828125 * 10240; time = 0.0399s; samplesPerSecond = 256847.6
12/20/2016 15:28:59:  Epoch[22 of 25]-Minibatch[ 651- 660, 60.14%]: CE.SM = 0.84776611 * 10240; Err = 0.27568359 * 10240; time = 0.0400s; samplesPerSecond = 256147.3
12/20/2016 15:28:59:  Epoch[22 of 25]-Minibatch[ 661- 670, 61.05%]: CE.SM = 0.89132690 * 10240; Err = 0.29169922 * 10240; time = 0.0474s; samplesPerSecond = 215906.2
12/20/2016 15:28:59:  Epoch[22 of 25]-Minibatch[ 671- 680, 61.96%]: CE.SM = 0.88533936 * 10240; Err = 0.28925781 * 10240; time = 0.0489s; samplesPerSecond = 209471.2
12/20/2016 15:28:59:  Epoch[22 of 25]-Minibatch[ 681- 690, 62.87%]: CE.SM = 0.89024658 * 10240; Err = 0.28876953 * 10240; time = 0.0465s; samplesPerSecond = 220082.5
12/20/2016 15:28:59:  Epoch[22 of 25]-Minibatch[ 691- 700, 63.78%]: CE.SM = 0.86568604 * 10240; Err = 0.28369141 * 10240; time = 0.0435s; samplesPerSecond = 235402.3
12/20/2016 15:28:59:  Epoch[22 of 25]-Minibatch[ 701- 710, 64.69%]: CE.SM = 0.86996460 * 10240; Err = 0.28359375 * 10240; time = 0.0444s; samplesPerSecond = 230729.4
12/20/2016 15:28:59:  Epoch[22 of 25]-Minibatch[ 711- 720, 65.60%]: CE.SM = 0.86973877 * 10240; Err = 0.28437500 * 10240; time = 0.0406s; samplesPerSecond = 252384.6
12/20/2016 15:28:59:  Epoch[22 of 25]-Minibatch[ 721- 730, 66.51%]: CE.SM = 0.88712158 * 10240; Err = 0.29140625 * 10240; time = 0.0399s; samplesPerSecond = 256448.8
12/20/2016 15:28:59:  Epoch[22 of 25]-Minibatch[ 731- 740, 67.43%]: CE.SM = 0.88612671 * 10240; Err = 0.29072266 * 10240; time = 0.0403s; samplesPerSecond = 254056.5
12/20/2016 15:28:59:  Epoch[22 of 25]-Minibatch[ 741- 750, 68.34%]: CE.SM = 0.86585083 * 10240; Err = 0.28642578 * 10240; time = 0.0400s; samplesPerSecond = 256064.0
12/20/2016 15:28:59:  Epoch[22 of 25]-Minibatch[ 751- 760, 69.25%]: CE.SM = 0.87653198 * 10240; Err = 0.29023437 * 10240; time = 0.0404s; samplesPerSecond = 253415.2
12/20/2016 15:28:59:  Epoch[22 of 25]-Minibatch[ 761- 770, 70.16%]: CE.SM = 0.88202515 * 10240; Err = 0.28544922 * 10240; time = 0.0406s; samplesPerSecond = 252403.3
12/20/2016 15:28:59:  Epoch[22 of 25]-Minibatch[ 771- 780, 71.07%]: CE.SM = 0.85845337 * 10240; Err = 0.28056641 * 10240; time = 0.0396s; samplesPerSecond = 258507.5
12/20/2016 15:28:59:  Epoch[22 of 25]-Minibatch[ 781- 790, 71.98%]: CE.SM = 0.85651855 * 10240; Err = 0.28076172 * 10240; time = 0.0397s; samplesPerSecond = 257791.7
12/20/2016 15:28:59:  Epoch[22 of 25]-Minibatch[ 791- 800, 72.89%]: CE.SM = 0.87976074 * 10240; Err = 0.29218750 * 10240; time = 0.0395s; samplesPerSecond = 259286.5
12/20/2016 15:28:59:  Epoch[22 of 25]-Minibatch[ 801- 810, 73.80%]: CE.SM = 0.88628540 * 10240; Err = 0.29013672 * 10240; time = 0.0398s; samplesPerSecond = 257422.3
12/20/2016 15:28:59:  Epoch[22 of 25]-Minibatch[ 811- 820, 74.72%]: CE.SM = 0.87674561 * 10240; Err = 0.28447266 * 10240; time = 0.0397s; samplesPerSecond = 257856.6
12/20/2016 15:28:59:  Epoch[22 of 25]-Minibatch[ 821- 830, 75.63%]: CE.SM = 0.88734131 * 10240; Err = 0.28876953 * 10240; time = 0.0401s; samplesPerSecond = 255049.9
12/20/2016 15:28:59:  Epoch[22 of 25]-Minibatch[ 831- 840, 76.54%]: CE.SM = 0.88290405 * 10240; Err = 0.28847656 * 10240; time = 0.0416s; samplesPerSecond = 246106.5
12/20/2016 15:29:00:  Epoch[22 of 25]-Minibatch[ 841- 850, 77.45%]: CE.SM = 0.90643921 * 10240; Err = 0.29501953 * 10240; time = 0.0400s; samplesPerSecond = 256032.0
12/20/2016 15:29:00:  Epoch[22 of 25]-Minibatch[ 851- 860, 78.36%]: CE.SM = 0.87778320 * 10240; Err = 0.28818359 * 10240; time = 0.0402s; samplesPerSecond = 254846.8
12/20/2016 15:29:00:  Epoch[22 of 25]-Minibatch[ 861- 870, 79.27%]: CE.SM = 0.87542725 * 10240; Err = 0.28408203 * 10240; time = 0.0398s; samplesPerSecond = 257331.7
12/20/2016 15:29:00:  Epoch[22 of 25]-Minibatch[ 871- 880, 80.18%]: CE.SM = 0.88619995 * 10240; Err = 0.28623047 * 10240; time = 0.0397s; samplesPerSecond = 257798.1
12/20/2016 15:29:00:  Epoch[22 of 25]-Minibatch[ 881- 890, 81.09%]: CE.SM = 0.88354492 * 10240; Err = 0.28779297 * 10240; time = 0.0399s; samplesPerSecond = 256712.4
12/20/2016 15:29:00:  Epoch[22 of 25]-Minibatch[ 891- 900, 82.00%]: CE.SM = 0.87291260 * 10240; Err = 0.28339844 * 10240; time = 0.0401s; samplesPerSecond = 255412.6
12/20/2016 15:29:00:  Epoch[22 of 25]-Minibatch[ 901- 910, 82.92%]: CE.SM = 0.87844849 * 10240; Err = 0.29072266 * 10240; time = 0.0400s; samplesPerSecond = 256025.6
12/20/2016 15:29:00:  Epoch[22 of 25]-Minibatch[ 911- 920, 83.83%]: CE.SM = 0.91080933 * 10240; Err = 0.29638672 * 10240; time = 0.0407s; samplesPerSecond = 251757.9
12/20/2016 15:29:00:  Epoch[22 of 25]-Minibatch[ 921- 930, 84.74%]: CE.SM = 0.87034912 * 10240; Err = 0.28457031 * 10240; time = 0.0423s; samplesPerSecond = 242263.7
12/20/2016 15:29:00:  Epoch[22 of 25]-Minibatch[ 931- 940, 85.65%]: CE.SM = 0.86542358 * 10240; Err = 0.28300781 * 10240; time = 0.0404s; samplesPerSecond = 253264.7
12/20/2016 15:29:00:  Epoch[22 of 25]-Minibatch[ 941- 950, 86.56%]: CE.SM = 0.87987061 * 10240; Err = 0.29052734 * 10240; time = 0.0401s; samplesPerSecond = 255635.7
12/20/2016 15:29:00:  Epoch[22 of 25]-Minibatch[ 951- 960, 87.47%]: CE.SM = 0.88959351 * 10240; Err = 0.29179688 * 10240; time = 0.0405s; samplesPerSecond = 252908.2
12/20/2016 15:29:00:  Epoch[22 of 25]-Minibatch[ 961- 970, 88.38%]: CE.SM = 0.89266968 * 10240; Err = 0.29042969 * 10240; time = 0.0406s; samplesPerSecond = 252024.3
12/20/2016 15:29:00:  Epoch[22 of 25]-Minibatch[ 971- 980, 89.29%]: CE.SM = 0.86838379 * 10240; Err = 0.27929688 * 10240; time = 0.0405s; samplesPerSecond = 252970.7
12/20/2016 15:29:00:  Epoch[22 of 25]-Minibatch[ 981- 990, 90.21%]: CE.SM = 0.89120483 * 10240; Err = 0.28710938 * 10240; time = 0.0407s; samplesPerSecond = 251560.0
12/20/2016 15:29:00:  Epoch[22 of 25]-Minibatch[ 991-1000, 91.12%]: CE.SM = 0.86557007 * 10240; Err = 0.28447266 * 10240; time = 0.0405s; samplesPerSecond = 252845.7
12/20/2016 15:29:00:  Epoch[22 of 25]-Minibatch[1001-1010, 92.03%]: CE.SM = 0.87616577 * 10240; Err = 0.28720703 * 10240; time = 0.0397s; samplesPerSecond = 257967.0
12/20/2016 15:29:00:  Epoch[22 of 25]-Minibatch[1011-1020, 92.94%]: CE.SM = 0.86434326 * 10240; Err = 0.28007813 * 10240; time = 0.0429s; samplesPerSecond = 238928.6
12/20/2016 15:29:00:  Epoch[22 of 25]-Minibatch[1021-1030, 93.85%]: CE.SM = 0.87040405 * 10240; Err = 0.28017578 * 10240; time = 0.0411s; samplesPerSecond = 248918.3
12/20/2016 15:29:00:  Epoch[22 of 25]-Minibatch[1031-1040, 94.76%]: CE.SM = 0.89610596 * 10240; Err = 0.28603516 * 10240; time = 0.0409s; samplesPerSecond = 250611.8
12/20/2016 15:29:00:  Epoch[22 of 25]-Minibatch[1041-1050, 95.67%]: CE.SM = 0.90671997 * 10240; Err = 0.29414062 * 10240; time = 0.0401s; samplesPerSecond = 255571.9
12/20/2016 15:29:00:  Epoch[22 of 25]-Minibatch[1051-1060, 96.58%]: CE.SM = 0.88673096 * 10240; Err = 0.28613281 * 10240; time = 0.0399s; samplesPerSecond = 256326.8
12/20/2016 15:29:00:  Epoch[22 of 25]-Minibatch[1061-1070, 97.49%]: CE.SM = 0.89990845 * 10240; Err = 0.29658203 * 10240; time = 0.0406s; samplesPerSecond = 252372.2
12/20/2016 15:29:00:  Epoch[22 of 25]-Minibatch[1071-1080, 98.41%]: CE.SM = 0.87387695 * 10240; Err = 0.28398438 * 10240; time = 0.0401s; samplesPerSecond = 255368.0
12/20/2016 15:29:01:  Epoch[22 of 25]-Minibatch[1081-1090, 99.32%]: CE.SM = 0.89566650 * 10240; Err = 0.29052734 * 10240; time = 0.0548s; samplesPerSecond = 186967.1
12/20/2016 15:29:01: Finished Epoch[22 of 25]: [Training] CE.SM = 0.88229865 * 1124823; Err = 0.28779995 * 1124823; totalSamplesSeen = 24746106; learningRatePerSample = 7.8124998e-05; epochTime=4.69942s
12/20/2016 15:29:01: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn.22'

12/20/2016 15:29:01: Starting Epoch 23: learning rate per sample = 0.000078  effective momentum = 0.900000  momentum as time constant = 9719.0 samples

12/20/2016 15:29:01: Starting minibatch loop.
12/20/2016 15:29:01:  Epoch[23 of 25]-Minibatch[   1-  10, 0.91%]: CE.SM = 0.88651304 * 10240; Err = 0.28466797 * 10240; time = 0.0455s; samplesPerSecond = 225124.2
12/20/2016 15:29:01:  Epoch[23 of 25]-Minibatch[  11-  20, 1.82%]: CE.SM = 0.86095362 * 10240; Err = 0.27978516 * 10240; time = 0.0417s; samplesPerSecond = 245381.1
12/20/2016 15:29:01:  Epoch[23 of 25]-Minibatch[  21-  30, 2.73%]: CE.SM = 0.89792061 * 10240; Err = 0.29775391 * 10240; time = 0.0403s; samplesPerSecond = 254188.9
12/20/2016 15:29:01:  Epoch[23 of 25]-Minibatch[  31-  40, 3.64%]: CE.SM = 0.88487358 * 10240; Err = 0.28847656 * 10240; time = 0.0403s; samplesPerSecond = 254151.1
12/20/2016 15:29:01:  Epoch[23 of 25]-Minibatch[  41-  50, 4.56%]: CE.SM = 0.89632874 * 10240; Err = 0.28886719 * 10240; time = 0.0405s; samplesPerSecond = 252858.2
12/20/2016 15:29:01:  Epoch[23 of 25]-Minibatch[  51-  60, 5.47%]: CE.SM = 0.88440361 * 10240; Err = 0.28974609 * 10240; time = 0.0402s; samplesPerSecond = 255018.2
12/20/2016 15:29:01:  Epoch[23 of 25]-Minibatch[  61-  70, 6.38%]: CE.SM = 0.86850281 * 10240; Err = 0.28281250 * 10240; time = 0.0406s; samplesPerSecond = 252465.5
12/20/2016 15:29:01:  Epoch[23 of 25]-Minibatch[  71-  80, 7.29%]: CE.SM = 0.87453842 * 10240; Err = 0.28496094 * 10240; time = 0.0399s; samplesPerSecond = 256686.6
12/20/2016 15:29:01:  Epoch[23 of 25]-Minibatch[  81-  90, 8.20%]: CE.SM = 0.87718124 * 10240; Err = 0.28769531 * 10240; time = 0.0404s; samplesPerSecond = 253471.6
12/20/2016 15:29:01:  Epoch[23 of 25]-Minibatch[  91- 100, 9.11%]: CE.SM = 0.89955978 * 10240; Err = 0.29521484 * 10240; time = 0.0405s; samplesPerSecond = 252534.0
12/20/2016 15:29:01:  Epoch[23 of 25]-Minibatch[ 101- 110, 10.02%]: CE.SM = 0.87087250 * 10240; Err = 0.28349609 * 10240; time = 0.0409s; samplesPerSecond = 250593.4
12/20/2016 15:29:01:  Epoch[23 of 25]-Minibatch[ 111- 120, 10.93%]: CE.SM = 0.88819962 * 10240; Err = 0.28828125 * 10240; time = 0.0406s; samplesPerSecond = 251943.7
12/20/2016 15:29:01:  Epoch[23 of 25]-Minibatch[ 121- 130, 11.85%]: CE.SM = 0.87607117 * 10240; Err = 0.28330078 * 10240; time = 0.0405s; samplesPerSecond = 253008.2
12/20/2016 15:29:01:  Epoch[23 of 25]-Minibatch[ 131- 140, 12.76%]: CE.SM = 0.89461746 * 10240; Err = 0.28984375 * 10240; time = 0.0404s; samplesPerSecond = 253358.7
12/20/2016 15:29:01:  Epoch[23 of 25]-Minibatch[ 141- 150, 13.67%]: CE.SM = 0.89164734 * 10240; Err = 0.29179688 * 10240; time = 0.0403s; samplesPerSecond = 254132.1
12/20/2016 15:29:01:  Epoch[23 of 25]-Minibatch[ 151- 160, 14.58%]: CE.SM = 0.87912750 * 10240; Err = 0.28466797 * 10240; time = 0.0406s; samplesPerSecond = 252272.7
12/20/2016 15:29:02:  Epoch[23 of 25]-Minibatch[ 161- 170, 15.49%]: CE.SM = 0.87551117 * 10240; Err = 0.27753906 * 10240; time = 0.0401s; samplesPerSecond = 255336.1
12/20/2016 15:29:02:  Epoch[23 of 25]-Minibatch[ 171- 180, 16.40%]: CE.SM = 0.89368744 * 10240; Err = 0.29365234 * 10240; time = 0.0406s; samplesPerSecond = 252204.3
12/20/2016 15:29:02:  Epoch[23 of 25]-Minibatch[ 181- 190, 17.31%]: CE.SM = 0.88565369 * 10240; Err = 0.29140625 * 10240; time = 0.0401s; samplesPerSecond = 255177.1
12/20/2016 15:29:02:  Epoch[23 of 25]-Minibatch[ 191- 200, 18.22%]: CE.SM = 0.87652435 * 10240; Err = 0.28349609 * 10240; time = 0.0406s; samplesPerSecond = 252378.4
12/20/2016 15:29:02:  Epoch[23 of 25]-Minibatch[ 201- 210, 19.13%]: CE.SM = 0.88084412 * 10240; Err = 0.28642578 * 10240; time = 0.0408s; samplesPerSecond = 251140.4
12/20/2016 15:29:02:  Epoch[23 of 25]-Minibatch[ 211- 220, 20.05%]: CE.SM = 0.86899567 * 10240; Err = 0.28437500 * 10240; time = 0.0406s; samplesPerSecond = 252142.2
12/20/2016 15:29:02:  Epoch[23 of 25]-Minibatch[ 221- 230, 20.96%]: CE.SM = 0.91273041 * 10240; Err = 0.29570313 * 10240; time = 0.0402s; samplesPerSecond = 254669.4
12/20/2016 15:29:02:  Epoch[23 of 25]-Minibatch[ 231- 240, 21.87%]: CE.SM = 0.87550659 * 10240; Err = 0.28554687 * 10240; time = 0.0402s; samplesPerSecond = 254720.0
12/20/2016 15:29:02:  Epoch[23 of 25]-Minibatch[ 241- 250, 22.78%]: CE.SM = 0.87972717 * 10240; Err = 0.28378906 * 10240; time = 0.0403s; samplesPerSecond = 254144.7
12/20/2016 15:29:02:  Epoch[23 of 25]-Minibatch[ 251- 260, 23.69%]: CE.SM = 0.86157227 * 10240; Err = 0.28652344 * 10240; time = 0.0406s; samplesPerSecond = 252477.9
12/20/2016 15:29:02:  Epoch[23 of 25]-Minibatch[ 261- 270, 24.60%]: CE.SM = 0.88950958 * 10240; Err = 0.28427734 * 10240; time = 0.0403s; samplesPerSecond = 254277.3
12/20/2016 15:29:02:  Epoch[23 of 25]-Minibatch[ 271- 280, 25.51%]: CE.SM = 0.88445282 * 10240; Err = 0.28769531 * 10240; time = 0.0405s; samplesPerSecond = 252639.9
12/20/2016 15:29:02:  Epoch[23 of 25]-Minibatch[ 281- 290, 26.42%]: CE.SM = 0.86918945 * 10240; Err = 0.28261719 * 10240; time = 0.0399s; samplesPerSecond = 256615.9
12/20/2016 15:29:02:  Epoch[23 of 25]-Minibatch[ 291- 300, 27.33%]: CE.SM = 0.86468048 * 10240; Err = 0.28535156 * 10240; time = 0.0405s; samplesPerSecond = 252571.3
12/20/2016 15:29:02:  Epoch[23 of 25]-Minibatch[ 301- 310, 28.25%]: CE.SM = 0.86579285 * 10240; Err = 0.28222656 * 10240; time = 0.0404s; samplesPerSecond = 253616.0
12/20/2016 15:29:02:  Epoch[23 of 25]-Minibatch[ 311- 320, 29.16%]: CE.SM = 0.87547302 * 10240; Err = 0.28378906 * 10240; time = 0.0403s; samplesPerSecond = 253949.4
12/20/2016 15:29:02:  Epoch[23 of 25]-Minibatch[ 321- 330, 30.07%]: CE.SM = 0.87590637 * 10240; Err = 0.28496094 * 10240; time = 0.0412s; samplesPerSecond = 248779.2
12/20/2016 15:29:02:  Epoch[23 of 25]-Minibatch[ 331- 340, 30.98%]: CE.SM = 0.86558228 * 10240; Err = 0.28154297 * 10240; time = 0.0401s; samplesPerSecond = 255476.3
12/20/2016 15:29:02:  Epoch[23 of 25]-Minibatch[ 341- 350, 31.89%]: CE.SM = 0.87654114 * 10240; Err = 0.28535156 * 10240; time = 0.0402s; samplesPerSecond = 254663.0
12/20/2016 15:29:02:  Epoch[23 of 25]-Minibatch[ 351- 360, 32.80%]: CE.SM = 0.88417664 * 10240; Err = 0.29121094 * 10240; time = 0.0402s; samplesPerSecond = 254466.8
12/20/2016 15:29:02:  Epoch[23 of 25]-Minibatch[ 361- 370, 33.71%]: CE.SM = 0.89947205 * 10240; Err = 0.29404297 * 10240; time = 0.0407s; samplesPerSecond = 251887.9
12/20/2016 15:29:02:  Epoch[23 of 25]-Minibatch[ 371- 380, 34.62%]: CE.SM = 0.89145203 * 10240; Err = 0.29335937 * 10240; time = 0.0404s; samplesPerSecond = 253584.6
12/20/2016 15:29:02:  Epoch[23 of 25]-Minibatch[ 381- 390, 35.54%]: CE.SM = 0.86369934 * 10240; Err = 0.28583984 * 10240; time = 0.0405s; samplesPerSecond = 252677.3
12/20/2016 15:29:02:  Epoch[23 of 25]-Minibatch[ 391- 400, 36.45%]: CE.SM = 0.89335632 * 10240; Err = 0.29492188 * 10240; time = 0.0406s; samplesPerSecond = 252297.5
12/20/2016 15:29:02:  Epoch[23 of 25]-Minibatch[ 401- 410, 37.36%]: CE.SM = 0.87629089 * 10240; Err = 0.28984375 * 10240; time = 0.0406s; samplesPerSecond = 252036.7
12/20/2016 15:29:03:  Epoch[23 of 25]-Minibatch[ 411- 420, 38.27%]: CE.SM = 0.88121643 * 10240; Err = 0.28730469 * 10240; time = 0.0401s; samplesPerSecond = 255559.2
12/20/2016 15:29:03:  Epoch[23 of 25]-Minibatch[ 421- 430, 39.18%]: CE.SM = 0.86733704 * 10240; Err = 0.28447266 * 10240; time = 0.0404s; samplesPerSecond = 253553.2
12/20/2016 15:29:03:  Epoch[23 of 25]-Minibatch[ 431- 440, 40.09%]: CE.SM = 0.88125916 * 10240; Err = 0.28935547 * 10240; time = 0.0419s; samplesPerSecond = 244578.2
12/20/2016 15:29:03:  Epoch[23 of 25]-Minibatch[ 441- 450, 41.00%]: CE.SM = 0.87631836 * 10240; Err = 0.29121094 * 10240; time = 0.0403s; samplesPerSecond = 254372.0
12/20/2016 15:29:03:  Epoch[23 of 25]-Minibatch[ 451- 460, 41.91%]: CE.SM = 0.87915955 * 10240; Err = 0.28593750 * 10240; time = 0.0405s; samplesPerSecond = 252783.3
12/20/2016 15:29:03:  Epoch[23 of 25]-Minibatch[ 461- 470, 42.82%]: CE.SM = 0.89221191 * 10240; Err = 0.28935547 * 10240; time = 0.0403s; samplesPerSecond = 254252.0
12/20/2016 15:29:03:  Epoch[23 of 25]-Minibatch[ 471- 480, 43.74%]: CE.SM = 0.85569458 * 10240; Err = 0.28476563 * 10240; time = 0.0405s; samplesPerSecond = 252789.6
12/20/2016 15:29:03:  Epoch[23 of 25]-Minibatch[ 481- 490, 44.65%]: CE.SM = 0.86928101 * 10240; Err = 0.28398438 * 10240; time = 0.0410s; samplesPerSecond = 249823.1
12/20/2016 15:29:03:  Epoch[23 of 25]-Minibatch[ 491- 500, 45.56%]: CE.SM = 0.90631409 * 10240; Err = 0.29765625 * 10240; time = 0.0401s; samplesPerSecond = 255247.0
12/20/2016 15:29:03:  Epoch[23 of 25]-Minibatch[ 501- 510, 46.47%]: CE.SM = 0.89281311 * 10240; Err = 0.29384766 * 10240; time = 0.0410s; samplesPerSecond = 249713.5
12/20/2016 15:29:03:  Epoch[23 of 25]-Minibatch[ 511- 520, 47.38%]: CE.SM = 0.88060913 * 10240; Err = 0.29013672 * 10240; time = 0.0401s; samplesPerSecond = 255259.7
12/20/2016 15:29:03:  Epoch[23 of 25]-Minibatch[ 521- 530, 48.29%]: CE.SM = 0.89348145 * 10240; Err = 0.29296875 * 10240; time = 0.0410s; samplesPerSecond = 249512.7
12/20/2016 15:29:03:  Epoch[23 of 25]-Minibatch[ 531- 540, 49.20%]: CE.SM = 0.88818359 * 10240; Err = 0.29160156 * 10240; time = 0.0416s; samplesPerSecond = 246408.5
12/20/2016 15:29:03:  Epoch[23 of 25]-Minibatch[ 541- 550, 50.11%]: CE.SM = 0.88176880 * 10240; Err = 0.28212891 * 10240; time = 0.0400s; samplesPerSecond = 255974.4
12/20/2016 15:29:03:  Epoch[23 of 25]-Minibatch[ 551- 560, 51.03%]: CE.SM = 0.85799255 * 10240; Err = 0.28046875 * 10240; time = 0.0400s; samplesPerSecond = 256217.8
12/20/2016 15:29:03:  Epoch[23 of 25]-Minibatch[ 561- 570, 51.94%]: CE.SM = 0.88323059 * 10240; Err = 0.28554687 * 10240; time = 0.0400s; samplesPerSecond = 256051.2
12/20/2016 15:29:03:  Epoch[23 of 25]-Minibatch[ 571- 580, 52.85%]: CE.SM = 0.88445129 * 10240; Err = 0.28720703 * 10240; time = 0.0403s; samplesPerSecond = 253987.2
12/20/2016 15:29:03:  Epoch[23 of 25]-Minibatch[ 581- 590, 53.76%]: CE.SM = 0.88396301 * 10240; Err = 0.28837891 * 10240; time = 0.0401s; samplesPerSecond = 255043.6
12/20/2016 15:29:03:  Epoch[23 of 25]-Minibatch[ 591- 600, 54.67%]: CE.SM = 0.87291870 * 10240; Err = 0.28837891 * 10240; time = 0.0402s; samplesPerSecond = 254663.0
12/20/2016 15:29:03:  Epoch[23 of 25]-Minibatch[ 601- 610, 55.58%]: CE.SM = 0.89089966 * 10240; Err = 0.29091797 * 10240; time = 0.0400s; samplesPerSecond = 256051.2
12/20/2016 15:29:03:  Epoch[23 of 25]-Minibatch[ 611- 620, 56.49%]: CE.SM = 0.88615112 * 10240; Err = 0.28183594 * 10240; time = 0.0399s; samplesPerSecond = 256944.3
12/20/2016 15:29:03:  Epoch[23 of 25]-Minibatch[ 621- 630, 57.40%]: CE.SM = 0.86563721 * 10240; Err = 0.27822266 * 10240; time = 0.0413s; samplesPerSecond = 247690.0
12/20/2016 15:29:03:  Epoch[23 of 25]-Minibatch[ 631- 640, 58.31%]: CE.SM = 0.89259033 * 10240; Err = 0.29003906 * 10240; time = 0.0426s; samplesPerSecond = 240341.7
12/20/2016 15:29:03:  Epoch[23 of 25]-Minibatch[ 641- 650, 59.23%]: CE.SM = 0.87581787 * 10240; Err = 0.28300781 * 10240; time = 0.0405s; samplesPerSecond = 253001.9
12/20/2016 15:29:03:  Epoch[23 of 25]-Minibatch[ 651- 660, 60.14%]: CE.SM = 0.89867554 * 10240; Err = 0.28974609 * 10240; time = 0.0411s; samplesPerSecond = 249203.0
12/20/2016 15:29:04:  Epoch[23 of 25]-Minibatch[ 661- 670, 61.05%]: CE.SM = 0.88766479 * 10240; Err = 0.28779297 * 10240; time = 0.0403s; samplesPerSecond = 254391.0
12/20/2016 15:29:04:  Epoch[23 of 25]-Minibatch[ 671- 680, 61.96%]: CE.SM = 0.88594360 * 10240; Err = 0.29160156 * 10240; time = 0.0400s; samplesPerSecond = 255808.1
12/20/2016 15:29:04:  Epoch[23 of 25]-Minibatch[ 681- 690, 62.87%]: CE.SM = 0.86351929 * 10240; Err = 0.28662109 * 10240; time = 0.0398s; samplesPerSecond = 257234.7
12/20/2016 15:29:04:  Epoch[23 of 25]-Minibatch[ 691- 700, 63.78%]: CE.SM = 0.88809204 * 10240; Err = 0.28710938 * 10240; time = 0.0399s; samplesPerSecond = 256879.8
12/20/2016 15:29:04:  Epoch[23 of 25]-Minibatch[ 701- 710, 64.69%]: CE.SM = 0.87951050 * 10240; Err = 0.28486328 * 10240; time = 0.0402s; samplesPerSecond = 254517.4
12/20/2016 15:29:04:  Epoch[23 of 25]-Minibatch[ 711- 720, 65.60%]: CE.SM = 0.87427368 * 10240; Err = 0.28388672 * 10240; time = 0.0402s; samplesPerSecond = 254587.0
12/20/2016 15:29:04:  Epoch[23 of 25]-Minibatch[ 721- 730, 66.51%]: CE.SM = 0.87376709 * 10240; Err = 0.28242187 * 10240; time = 0.0404s; samplesPerSecond = 253503.0
12/20/2016 15:29:04:  Epoch[23 of 25]-Minibatch[ 731- 740, 67.43%]: CE.SM = 0.87812500 * 10240; Err = 0.28027344 * 10240; time = 0.0395s; samplesPerSecond = 258939.0
12/20/2016 15:29:04:  Epoch[23 of 25]-Minibatch[ 741- 750, 68.34%]: CE.SM = 0.87862549 * 10240; Err = 0.29218750 * 10240; time = 0.0406s; samplesPerSecond = 252496.6
12/20/2016 15:29:04:  Epoch[23 of 25]-Minibatch[ 751- 760, 69.25%]: CE.SM = 0.87747803 * 10240; Err = 0.28759766 * 10240; time = 0.0391s; samplesPerSecond = 261631.6
12/20/2016 15:29:04:  Epoch[23 of 25]-Minibatch[ 761- 770, 70.16%]: CE.SM = 0.86751709 * 10240; Err = 0.29033203 * 10240; time = 0.0399s; samplesPerSecond = 256519.5
12/20/2016 15:29:04:  Epoch[23 of 25]-Minibatch[ 771- 780, 71.07%]: CE.SM = 0.89135132 * 10240; Err = 0.29316406 * 10240; time = 0.0397s; samplesPerSecond = 257960.5
12/20/2016 15:29:04:  Epoch[23 of 25]-Minibatch[ 781- 790, 71.98%]: CE.SM = 0.90701904 * 10240; Err = 0.30048828 * 10240; time = 0.0398s; samplesPerSecond = 257183.0
12/20/2016 15:29:04:  Epoch[23 of 25]-Minibatch[ 791- 800, 72.89%]: CE.SM = 0.87629395 * 10240; Err = 0.28183594 * 10240; time = 0.0406s; samplesPerSecond = 252334.8
12/20/2016 15:29:04:  Epoch[23 of 25]-Minibatch[ 801- 810, 73.80%]: CE.SM = 0.88806152 * 10240; Err = 0.29394531 * 10240; time = 0.0399s; samplesPerSecond = 256834.7
12/20/2016 15:29:04:  Epoch[23 of 25]-Minibatch[ 811- 820, 74.72%]: CE.SM = 0.88731079 * 10240; Err = 0.29101562 * 10240; time = 0.0400s; samplesPerSecond = 255904.0
12/20/2016 15:29:04:  Epoch[23 of 25]-Minibatch[ 821- 830, 75.63%]: CE.SM = 0.90524902 * 10240; Err = 0.29443359 * 10240; time = 0.0433s; samplesPerSecond = 236391.3
12/20/2016 15:29:04:  Epoch[23 of 25]-Minibatch[ 831- 840, 76.54%]: CE.SM = 0.87328491 * 10240; Err = 0.28662109 * 10240; time = 0.0409s; samplesPerSecond = 250323.9
12/20/2016 15:29:04:  Epoch[23 of 25]-Minibatch[ 841- 850, 77.45%]: CE.SM = 0.87066650 * 10240; Err = 0.28476563 * 10240; time = 0.0393s; samplesPerSecond = 260526.7
12/20/2016 15:29:04:  Epoch[23 of 25]-Minibatch[ 851- 860, 78.36%]: CE.SM = 0.86732788 * 10240; Err = 0.28476563 * 10240; time = 0.0398s; samplesPerSecond = 257499.9
12/20/2016 15:29:04:  Epoch[23 of 25]-Minibatch[ 861- 870, 79.27%]: CE.SM = 0.89156494 * 10240; Err = 0.28759766 * 10240; time = 0.0407s; samplesPerSecond = 251424.1
12/20/2016 15:29:04:  Epoch[23 of 25]-Minibatch[ 871- 880, 80.18%]: CE.SM = 0.87503662 * 10240; Err = 0.28125000 * 10240; time = 0.0397s; samplesPerSecond = 257739.7
12/20/2016 15:29:04:  Epoch[23 of 25]-Minibatch[ 881- 890, 81.09%]: CE.SM = 0.86363525 * 10240; Err = 0.28281250 * 10240; time = 0.0401s; samplesPerSecond = 255565.5
12/20/2016 15:29:04:  Epoch[23 of 25]-Minibatch[ 891- 900, 82.00%]: CE.SM = 0.86847534 * 10240; Err = 0.28437500 * 10240; time = 0.0400s; samplesPerSecond = 255776.2
12/20/2016 15:29:05:  Epoch[23 of 25]-Minibatch[ 901- 910, 82.92%]: CE.SM = 0.86210937 * 10240; Err = 0.28496094 * 10240; time = 0.0400s; samplesPerSecond = 255980.8
12/20/2016 15:29:05:  Epoch[23 of 25]-Minibatch[ 911- 920, 83.83%]: CE.SM = 0.89123535 * 10240; Err = 0.29267578 * 10240; time = 0.0395s; samplesPerSecond = 259129.0
12/20/2016 15:29:05:  Epoch[23 of 25]-Minibatch[ 921- 930, 84.74%]: CE.SM = 0.88458252 * 10240; Err = 0.29267578 * 10240; time = 0.0412s; samplesPerSecond = 248393.0
12/20/2016 15:29:05:  Epoch[23 of 25]-Minibatch[ 931- 940, 85.65%]: CE.SM = 0.89018555 * 10240; Err = 0.29169922 * 10240; time = 0.0396s; samplesPerSecond = 258514.0
12/20/2016 15:29:05:  Epoch[23 of 25]-Minibatch[ 941- 950, 86.56%]: CE.SM = 0.86936035 * 10240; Err = 0.28251953 * 10240; time = 0.0404s; samplesPerSecond = 253572.0
12/20/2016 15:29:05:  Epoch[23 of 25]-Minibatch[ 951- 960, 87.47%]: CE.SM = 0.88899536 * 10240; Err = 0.29121094 * 10240; time = 0.0405s; samplesPerSecond = 253051.9
12/20/2016 15:29:05:  Epoch[23 of 25]-Minibatch[ 961- 970, 88.38%]: CE.SM = 0.88395996 * 10240; Err = 0.28779297 * 10240; time = 0.0401s; samplesPerSecond = 255674.0
12/20/2016 15:29:05:  Epoch[23 of 25]-Minibatch[ 971- 980, 89.29%]: CE.SM = 0.88723755 * 10240; Err = 0.28984375 * 10240; time = 0.0400s; samplesPerSecond = 255827.3
12/20/2016 15:29:05:  Epoch[23 of 25]-Minibatch[ 981- 990, 90.21%]: CE.SM = 0.87328491 * 10240; Err = 0.28476563 * 10240; time = 0.0409s; samplesPerSecond = 250452.5
12/20/2016 15:29:05:  Epoch[23 of 25]-Minibatch[ 991-1000, 91.12%]: CE.SM = 0.87924194 * 10240; Err = 0.28447266 * 10240; time = 0.0406s; samplesPerSecond = 252191.9
12/20/2016 15:29:05:  Epoch[23 of 25]-Minibatch[1001-1010, 92.03%]: CE.SM = 0.88939819 * 10240; Err = 0.29082031 * 10240; time = 0.0401s; samplesPerSecond = 255278.8
12/20/2016 15:29:05:  Epoch[23 of 25]-Minibatch[1011-1020, 92.94%]: CE.SM = 0.87916260 * 10240; Err = 0.29033203 * 10240; time = 0.0403s; samplesPerSecond = 254308.8
12/20/2016 15:29:05:  Epoch[23 of 25]-Minibatch[1021-1030, 93.85%]: CE.SM = 0.87203369 * 10240; Err = 0.28173828 * 10240; time = 0.0400s; samplesPerSecond = 255859.3
12/20/2016 15:29:05:  Epoch[23 of 25]-Minibatch[1031-1040, 94.76%]: CE.SM = 0.88570557 * 10240; Err = 0.28535156 * 10240; time = 0.0403s; samplesPerSecond = 253823.5
12/20/2016 15:29:05:  Epoch[23 of 25]-Minibatch[1041-1050, 95.67%]: CE.SM = 0.88486328 * 10240; Err = 0.29101562 * 10240; time = 0.0411s; samplesPerSecond = 249391.1
12/20/2016 15:29:05:  Epoch[23 of 25]-Minibatch[1051-1060, 96.58%]: CE.SM = 0.89329834 * 10240; Err = 0.29335937 * 10240; time = 0.0403s; samplesPerSecond = 254050.2
12/20/2016 15:29:05:  Epoch[23 of 25]-Minibatch[1061-1070, 97.49%]: CE.SM = 0.87462158 * 10240; Err = 0.28388672 * 10240; time = 0.0400s; samplesPerSecond = 255929.6
12/20/2016 15:29:05:  Epoch[23 of 25]-Minibatch[1071-1080, 98.41%]: CE.SM = 0.87763672 * 10240; Err = 0.28193359 * 10240; time = 0.0401s; samplesPerSecond = 255412.6
12/20/2016 15:29:05:  Epoch[23 of 25]-Minibatch[1081-1090, 99.32%]: CE.SM = 0.86640625 * 10240; Err = 0.28056641 * 10240; time = 0.0400s; samplesPerSecond = 256153.7
12/20/2016 15:29:05: Finished Epoch[23 of 25]: [Training] CE.SM = 0.88100756 * 1124823; Err = 0.28743989 * 1124823; totalSamplesSeen = 25870929; learningRatePerSample = 7.8124998e-05; epochTime=4.6354s
12/20/2016 15:29:05: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn.23'

12/20/2016 15:29:05: Starting Epoch 24: learning rate per sample = 0.000078  effective momentum = 0.900000  momentum as time constant = 9719.0 samples

12/20/2016 15:29:06: Starting minibatch loop.
12/20/2016 15:29:06:  Epoch[24 of 25]-Minibatch[   1-  10, 0.91%]: CE.SM = 0.87919397 * 10240; Err = 0.28710938 * 10240; time = 0.0445s; samplesPerSecond = 229859.3
12/20/2016 15:29:06:  Epoch[24 of 25]-Minibatch[  11-  20, 1.82%]: CE.SM = 0.87253027 * 10240; Err = 0.28369141 * 10240; time = 0.0403s; samplesPerSecond = 253854.9
12/20/2016 15:29:06:  Epoch[24 of 25]-Minibatch[  21-  30, 2.73%]: CE.SM = 0.85156651 * 10240; Err = 0.28037109 * 10240; time = 0.0398s; samplesPerSecond = 257079.7
12/20/2016 15:29:06:  Epoch[24 of 25]-Minibatch[  31-  40, 3.64%]: CE.SM = 0.86256351 * 10240; Err = 0.28378906 * 10240; time = 0.0405s; samplesPerSecond = 252964.4
12/20/2016 15:29:06:  Epoch[24 of 25]-Minibatch[  41-  50, 4.56%]: CE.SM = 0.86636963 * 10240; Err = 0.28027344 * 10240; time = 0.0411s; samplesPerSecond = 249221.2
12/20/2016 15:29:06:  Epoch[24 of 25]-Minibatch[  51-  60, 5.47%]: CE.SM = 0.91239967 * 10240; Err = 0.29335937 * 10240; time = 0.0407s; samplesPerSecond = 251788.8
12/20/2016 15:29:06:  Epoch[24 of 25]-Minibatch[  61-  70, 6.38%]: CE.SM = 0.88147392 * 10240; Err = 0.28779297 * 10240; time = 0.0401s; samplesPerSecond = 255374.3
12/20/2016 15:29:06:  Epoch[24 of 25]-Minibatch[  71-  80, 7.29%]: CE.SM = 0.87586403 * 10240; Err = 0.28906250 * 10240; time = 0.0403s; samplesPerSecond = 253943.1
12/20/2016 15:29:06:  Epoch[24 of 25]-Minibatch[  81-  90, 8.20%]: CE.SM = 0.86768188 * 10240; Err = 0.28417969 * 10240; time = 0.0407s; samplesPerSecond = 251467.3
12/20/2016 15:29:06:  Epoch[24 of 25]-Minibatch[  91- 100, 9.11%]: CE.SM = 0.85453262 * 10240; Err = 0.28066406 * 10240; time = 0.0403s; samplesPerSecond = 254264.6
12/20/2016 15:29:06:  Epoch[24 of 25]-Minibatch[ 101- 110, 10.02%]: CE.SM = 0.88602066 * 10240; Err = 0.29121094 * 10240; time = 0.0413s; samplesPerSecond = 247941.9
12/20/2016 15:29:06:  Epoch[24 of 25]-Minibatch[ 111- 120, 10.93%]: CE.SM = 0.88729401 * 10240; Err = 0.29228516 * 10240; time = 0.0406s; samplesPerSecond = 251949.9
12/20/2016 15:29:06:  Epoch[24 of 25]-Minibatch[ 121- 130, 11.85%]: CE.SM = 0.88421555 * 10240; Err = 0.28681641 * 10240; time = 0.0408s; samplesPerSecond = 250722.3
12/20/2016 15:29:06:  Epoch[24 of 25]-Minibatch[ 131- 140, 12.76%]: CE.SM = 0.88191605 * 10240; Err = 0.28867188 * 10240; time = 0.0409s; samplesPerSecond = 250177.1
12/20/2016 15:29:06:  Epoch[24 of 25]-Minibatch[ 141- 150, 13.67%]: CE.SM = 0.87585983 * 10240; Err = 0.28330078 * 10240; time = 0.0402s; samplesPerSecond = 254815.1
12/20/2016 15:29:06:  Epoch[24 of 25]-Minibatch[ 151- 160, 14.58%]: CE.SM = 0.87839661 * 10240; Err = 0.28466797 * 10240; time = 0.0399s; samplesPerSecond = 256821.8
12/20/2016 15:29:06:  Epoch[24 of 25]-Minibatch[ 161- 170, 15.49%]: CE.SM = 0.86909943 * 10240; Err = 0.28574219 * 10240; time = 0.0397s; samplesPerSecond = 258136.1
12/20/2016 15:29:06:  Epoch[24 of 25]-Minibatch[ 171- 180, 16.40%]: CE.SM = 0.87240753 * 10240; Err = 0.28632812 * 10240; time = 0.0397s; samplesPerSecond = 258214.2
12/20/2016 15:29:06:  Epoch[24 of 25]-Minibatch[ 181- 190, 17.31%]: CE.SM = 0.87971344 * 10240; Err = 0.29072266 * 10240; time = 0.0397s; samplesPerSecond = 257980.0
12/20/2016 15:29:06:  Epoch[24 of 25]-Minibatch[ 191- 200, 18.22%]: CE.SM = 0.89180756 * 10240; Err = 0.28701172 * 10240; time = 0.0396s; samplesPerSecond = 258853.9
12/20/2016 15:29:06:  Epoch[24 of 25]-Minibatch[ 201- 210, 19.13%]: CE.SM = 0.88923340 * 10240; Err = 0.28974609 * 10240; time = 0.0396s; samplesPerSecond = 258474.9
12/20/2016 15:29:06:  Epoch[24 of 25]-Minibatch[ 211- 220, 20.05%]: CE.SM = 0.89096832 * 10240; Err = 0.29082031 * 10240; time = 0.0399s; samplesPerSecond = 256738.1
12/20/2016 15:29:06:  Epoch[24 of 25]-Minibatch[ 221- 230, 20.96%]: CE.SM = 0.88581848 * 10240; Err = 0.28925781 * 10240; time = 0.0401s; samplesPerSecond = 255527.3
12/20/2016 15:29:06:  Epoch[24 of 25]-Minibatch[ 231- 240, 21.87%]: CE.SM = 0.87536774 * 10240; Err = 0.28642578 * 10240; time = 0.0401s; samplesPerSecond = 255584.7
12/20/2016 15:29:07:  Epoch[24 of 25]-Minibatch[ 241- 250, 22.78%]: CE.SM = 0.87635956 * 10240; Err = 0.28593750 * 10240; time = 0.0398s; samplesPerSecond = 257512.9
12/20/2016 15:29:07:  Epoch[24 of 25]-Minibatch[ 251- 260, 23.69%]: CE.SM = 0.87325745 * 10240; Err = 0.28593750 * 10240; time = 0.0399s; samplesPerSecond = 256615.9
12/20/2016 15:29:07:  Epoch[24 of 25]-Minibatch[ 261- 270, 24.60%]: CE.SM = 0.87829132 * 10240; Err = 0.28916016 * 10240; time = 0.0397s; samplesPerSecond = 258220.7
12/20/2016 15:29:07:  Epoch[24 of 25]-Minibatch[ 271- 280, 25.51%]: CE.SM = 0.87548065 * 10240; Err = 0.28496094 * 10240; time = 0.0397s; samplesPerSecond = 257902.0
12/20/2016 15:29:07:  Epoch[24 of 25]-Minibatch[ 281- 290, 26.42%]: CE.SM = 0.88986816 * 10240; Err = 0.28613281 * 10240; time = 0.0395s; samplesPerSecond = 259083.1
12/20/2016 15:29:07:  Epoch[24 of 25]-Minibatch[ 291- 300, 27.33%]: CE.SM = 0.88257141 * 10240; Err = 0.29169922 * 10240; time = 0.0399s; samplesPerSecond = 256551.6
12/20/2016 15:29:07:  Epoch[24 of 25]-Minibatch[ 301- 310, 28.25%]: CE.SM = 0.88671875 * 10240; Err = 0.28798828 * 10240; time = 0.0397s; samplesPerSecond = 258149.1
12/20/2016 15:29:07:  Epoch[24 of 25]-Minibatch[ 311- 320, 29.16%]: CE.SM = 0.86683044 * 10240; Err = 0.28398438 * 10240; time = 0.0402s; samplesPerSecond = 254637.7
12/20/2016 15:29:07:  Epoch[24 of 25]-Minibatch[ 321- 330, 30.07%]: CE.SM = 0.89520874 * 10240; Err = 0.29208984 * 10240; time = 0.0393s; samplesPerSecond = 260573.1
12/20/2016 15:29:07:  Epoch[24 of 25]-Minibatch[ 331- 340, 30.98%]: CE.SM = 0.88999329 * 10240; Err = 0.29082031 * 10240; time = 0.0395s; samplesPerSecond = 259194.6
12/20/2016 15:29:07:  Epoch[24 of 25]-Minibatch[ 341- 350, 31.89%]: CE.SM = 0.88322449 * 10240; Err = 0.28876953 * 10240; time = 0.0396s; samplesPerSecond = 258507.5
12/20/2016 15:29:07:  Epoch[24 of 25]-Minibatch[ 351- 360, 32.80%]: CE.SM = 0.86781006 * 10240; Err = 0.28144531 * 10240; time = 0.0395s; samplesPerSecond = 258997.9
12/20/2016 15:29:07:  Epoch[24 of 25]-Minibatch[ 361- 370, 33.71%]: CE.SM = 0.86748047 * 10240; Err = 0.28457031 * 10240; time = 0.0400s; samplesPerSecond = 255789.0
12/20/2016 15:29:07:  Epoch[24 of 25]-Minibatch[ 371- 380, 34.62%]: CE.SM = 0.87294006 * 10240; Err = 0.28486328 * 10240; time = 0.0397s; samplesPerSecond = 257649.0
12/20/2016 15:29:07:  Epoch[24 of 25]-Minibatch[ 381- 390, 35.54%]: CE.SM = 0.88226624 * 10240; Err = 0.28437500 * 10240; time = 0.0395s; samplesPerSecond = 259266.8
12/20/2016 15:29:07:  Epoch[24 of 25]-Minibatch[ 391- 400, 36.45%]: CE.SM = 0.86514587 * 10240; Err = 0.28886719 * 10240; time = 0.0393s; samplesPerSecond = 260447.1
12/20/2016 15:29:07:  Epoch[24 of 25]-Minibatch[ 401- 410, 37.36%]: CE.SM = 0.86791687 * 10240; Err = 0.28876953 * 10240; time = 0.0397s; samplesPerSecond = 258038.5
12/20/2016 15:29:07:  Epoch[24 of 25]-Minibatch[ 411- 420, 38.27%]: CE.SM = 0.87650452 * 10240; Err = 0.28496094 * 10240; time = 0.0400s; samplesPerSecond = 255737.9
12/20/2016 15:29:07:  Epoch[24 of 25]-Minibatch[ 421- 430, 39.18%]: CE.SM = 0.88462830 * 10240; Err = 0.28818359 * 10240; time = 0.0394s; samplesPerSecond = 260116.3
12/20/2016 15:29:07:  Epoch[24 of 25]-Minibatch[ 431- 440, 40.09%]: CE.SM = 0.87790527 * 10240; Err = 0.28681641 * 10240; time = 0.0396s; samplesPerSecond = 258266.3
12/20/2016 15:29:07:  Epoch[24 of 25]-Minibatch[ 441- 450, 41.00%]: CE.SM = 0.87976074 * 10240; Err = 0.28583984 * 10240; time = 0.0396s; samplesPerSecond = 258527.1
12/20/2016 15:29:07:  Epoch[24 of 25]-Minibatch[ 451- 460, 41.91%]: CE.SM = 0.87291565 * 10240; Err = 0.28369141 * 10240; time = 0.0395s; samplesPerSecond = 259194.6
12/20/2016 15:29:07:  Epoch[24 of 25]-Minibatch[ 461- 470, 42.82%]: CE.SM = 0.87991333 * 10240; Err = 0.28427734 * 10240; time = 0.0394s; samplesPerSecond = 259687.6
12/20/2016 15:29:07:  Epoch[24 of 25]-Minibatch[ 471- 480, 43.74%]: CE.SM = 0.88432617 * 10240; Err = 0.28828125 * 10240; time = 0.0395s; samplesPerSecond = 259096.2
12/20/2016 15:29:07:  Epoch[24 of 25]-Minibatch[ 481- 490, 44.65%]: CE.SM = 0.86301575 * 10240; Err = 0.28242187 * 10240; time = 0.0396s; samplesPerSecond = 258749.2
12/20/2016 15:29:08:  Epoch[24 of 25]-Minibatch[ 491- 500, 45.56%]: CE.SM = 0.87623291 * 10240; Err = 0.28232422 * 10240; time = 0.0395s; samplesPerSecond = 259024.1
12/20/2016 15:29:08:  Epoch[24 of 25]-Minibatch[ 501- 510, 46.47%]: CE.SM = 0.85718079 * 10240; Err = 0.27939453 * 10240; time = 0.0396s; samplesPerSecond = 258331.4
12/20/2016 15:29:08:  Epoch[24 of 25]-Minibatch[ 511- 520, 47.38%]: CE.SM = 0.86884766 * 10240; Err = 0.28154297 * 10240; time = 0.0401s; samplesPerSecond = 255661.2
12/20/2016 15:29:08:  Epoch[24 of 25]-Minibatch[ 521- 530, 48.29%]: CE.SM = 0.86895447 * 10240; Err = 0.28095703 * 10240; time = 0.0397s; samplesPerSecond = 258227.2
12/20/2016 15:29:08:  Epoch[24 of 25]-Minibatch[ 531- 540, 49.20%]: CE.SM = 0.86946716 * 10240; Err = 0.28242187 * 10240; time = 0.0396s; samplesPerSecond = 258566.3
12/20/2016 15:29:08:  Epoch[24 of 25]-Minibatch[ 541- 550, 50.11%]: CE.SM = 0.87769165 * 10240; Err = 0.28701172 * 10240; time = 0.0394s; samplesPerSecond = 259951.3
12/20/2016 15:29:08:  Epoch[24 of 25]-Minibatch[ 551- 560, 51.03%]: CE.SM = 0.88883057 * 10240; Err = 0.29345703 * 10240; time = 0.0398s; samplesPerSecond = 257519.4
12/20/2016 15:29:08:  Epoch[24 of 25]-Minibatch[ 561- 570, 51.94%]: CE.SM = 0.89376526 * 10240; Err = 0.29150391 * 10240; time = 0.0403s; samplesPerSecond = 254233.1
12/20/2016 15:29:08:  Epoch[24 of 25]-Minibatch[ 571- 580, 52.85%]: CE.SM = 0.86233521 * 10240; Err = 0.28242187 * 10240; time = 0.0394s; samplesPerSecond = 260017.3
12/20/2016 15:29:08:  Epoch[24 of 25]-Minibatch[ 581- 590, 53.76%]: CE.SM = 0.89376221 * 10240; Err = 0.29472656 * 10240; time = 0.0410s; samplesPerSecond = 249689.1
12/20/2016 15:29:08:  Epoch[24 of 25]-Minibatch[ 591- 600, 54.67%]: CE.SM = 0.87392578 * 10240; Err = 0.28427734 * 10240; time = 0.0406s; samplesPerSecond = 252117.4
12/20/2016 15:29:08:  Epoch[24 of 25]-Minibatch[ 601- 610, 55.58%]: CE.SM = 0.88525391 * 10240; Err = 0.28808594 * 10240; time = 0.0410s; samplesPerSecond = 249762.2
12/20/2016 15:29:08:  Epoch[24 of 25]-Minibatch[ 611- 620, 56.49%]: CE.SM = 0.89202271 * 10240; Err = 0.28984375 * 10240; time = 0.0414s; samplesPerSecond = 247462.5
12/20/2016 15:29:08:  Epoch[24 of 25]-Minibatch[ 621- 630, 57.40%]: CE.SM = 0.89587402 * 10240; Err = 0.29189453 * 10240; time = 0.0429s; samplesPerSecond = 238906.3
12/20/2016 15:29:08:  Epoch[24 of 25]-Minibatch[ 631- 640, 58.31%]: CE.SM = 0.89346313 * 10240; Err = 0.29384766 * 10240; time = 0.0419s; samplesPerSecond = 244292.3
12/20/2016 15:29:08:  Epoch[24 of 25]-Minibatch[ 641- 650, 59.23%]: CE.SM = 0.88958130 * 10240; Err = 0.28837891 * 10240; time = 0.0435s; samplesPerSecond = 235380.7
12/20/2016 15:29:08:  Epoch[24 of 25]-Minibatch[ 651- 660, 60.14%]: CE.SM = 0.86804810 * 10240; Err = 0.28193359 * 10240; time = 0.0406s; samplesPerSecond = 252210.5
12/20/2016 15:29:08:  Epoch[24 of 25]-Minibatch[ 661- 670, 61.05%]: CE.SM = 0.87697754 * 10240; Err = 0.28232422 * 10240; time = 0.0410s; samplesPerSecond = 249920.7
12/20/2016 15:29:08:  Epoch[24 of 25]-Minibatch[ 671- 680, 61.96%]: CE.SM = 0.89741821 * 10240; Err = 0.29443359 * 10240; time = 0.0407s; samplesPerSecond = 251696.0
12/20/2016 15:29:08:  Epoch[24 of 25]-Minibatch[ 681- 690, 62.87%]: CE.SM = 0.87470093 * 10240; Err = 0.29326172 * 10240; time = 0.0435s; samplesPerSecond = 235391.5
12/20/2016 15:29:08:  Epoch[24 of 25]-Minibatch[ 691- 700, 63.78%]: CE.SM = 0.88763428 * 10240; Err = 0.29052734 * 10240; time = 0.0431s; samplesPerSecond = 237785.6
12/20/2016 15:29:08:  Epoch[24 of 25]-Minibatch[ 701- 710, 64.69%]: CE.SM = 0.87606812 * 10240; Err = 0.28574219 * 10240; time = 0.0404s; samplesPerSecond = 253314.9
12/20/2016 15:29:08:  Epoch[24 of 25]-Minibatch[ 711- 720, 65.60%]: CE.SM = 0.87131958 * 10240; Err = 0.28798828 * 10240; time = 0.0409s; samplesPerSecond = 250342.3
12/20/2016 15:29:08:  Epoch[24 of 25]-Minibatch[ 721- 730, 66.51%]: CE.SM = 0.88819580 * 10240; Err = 0.28876953 * 10240; time = 0.0412s; samplesPerSecond = 248284.6
12/20/2016 15:29:09:  Epoch[24 of 25]-Minibatch[ 731- 740, 67.43%]: CE.SM = 0.87957764 * 10240; Err = 0.27949219 * 10240; time = 0.0436s; samplesPerSecond = 234733.2
12/20/2016 15:29:09:  Epoch[24 of 25]-Minibatch[ 741- 750, 68.34%]: CE.SM = 0.86903076 * 10240; Err = 0.28359375 * 10240; time = 0.0408s; samplesPerSecond = 250980.4
12/20/2016 15:29:09:  Epoch[24 of 25]-Minibatch[ 751- 760, 69.25%]: CE.SM = 0.87500000 * 10240; Err = 0.28437500 * 10240; time = 0.0405s; samplesPerSecond = 252939.4
12/20/2016 15:29:09:  Epoch[24 of 25]-Minibatch[ 761- 770, 70.16%]: CE.SM = 0.87287598 * 10240; Err = 0.28681641 * 10240; time = 0.0413s; samplesPerSecond = 247809.9
12/20/2016 15:29:09:  Epoch[24 of 25]-Minibatch[ 771- 780, 71.07%]: CE.SM = 0.87509766 * 10240; Err = 0.28603516 * 10240; time = 0.0403s; samplesPerSecond = 254043.9
12/20/2016 15:29:09:  Epoch[24 of 25]-Minibatch[ 781- 790, 71.98%]: CE.SM = 0.87227783 * 10240; Err = 0.28574219 * 10240; time = 0.0407s; samplesPerSecond = 251485.8
12/20/2016 15:29:09:  Epoch[24 of 25]-Minibatch[ 791- 800, 72.89%]: CE.SM = 0.89391479 * 10240; Err = 0.29130859 * 10240; time = 0.0409s; samplesPerSecond = 250624.1
12/20/2016 15:29:09:  Epoch[24 of 25]-Minibatch[ 801- 810, 73.80%]: CE.SM = 0.87792358 * 10240; Err = 0.28437500 * 10240; time = 0.0410s; samplesPerSecond = 249981.7
12/20/2016 15:29:09:  Epoch[24 of 25]-Minibatch[ 811- 820, 74.72%]: CE.SM = 0.89371338 * 10240; Err = 0.28955078 * 10240; time = 0.0404s; samplesPerSecond = 253333.7
12/20/2016 15:29:09:  Epoch[24 of 25]-Minibatch[ 821- 830, 75.63%]: CE.SM = 0.86367798 * 10240; Err = 0.27968750 * 10240; time = 0.0407s; samplesPerSecond = 251683.6
12/20/2016 15:29:09:  Epoch[24 of 25]-Minibatch[ 831- 840, 76.54%]: CE.SM = 0.89010010 * 10240; Err = 0.28964844 * 10240; time = 0.0407s; samplesPerSecond = 251479.7
12/20/2016 15:29:09:  Epoch[24 of 25]-Minibatch[ 841- 850, 77.45%]: CE.SM = 0.88697510 * 10240; Err = 0.28828125 * 10240; time = 0.0403s; samplesPerSecond = 253810.9
12/20/2016 15:29:09:  Epoch[24 of 25]-Minibatch[ 851- 860, 78.36%]: CE.SM = 0.89348755 * 10240; Err = 0.29521484 * 10240; time = 0.0412s; samplesPerSecond = 248296.6
12/20/2016 15:29:09:  Epoch[24 of 25]-Minibatch[ 861- 870, 79.27%]: CE.SM = 0.89772949 * 10240; Err = 0.29082031 * 10240; time = 0.0404s; samplesPerSecond = 253616.0
12/20/2016 15:29:09:  Epoch[24 of 25]-Minibatch[ 871- 880, 80.18%]: CE.SM = 0.87285767 * 10240; Err = 0.28437500 * 10240; time = 0.0411s; samplesPerSecond = 249263.6
12/20/2016 15:29:09:  Epoch[24 of 25]-Minibatch[ 881- 890, 81.09%]: CE.SM = 0.86743774 * 10240; Err = 0.28349609 * 10240; time = 0.0409s; samplesPerSecond = 250342.3
12/20/2016 15:29:09:  Epoch[24 of 25]-Minibatch[ 891- 900, 82.00%]: CE.SM = 0.88441772 * 10240; Err = 0.29121094 * 10240; time = 0.0407s; samplesPerSecond = 251726.9
12/20/2016 15:29:09:  Epoch[24 of 25]-Minibatch[ 901- 910, 82.92%]: CE.SM = 0.89645996 * 10240; Err = 0.28896484 * 10240; time = 0.0407s; samplesPerSecond = 251566.1
12/20/2016 15:29:09:  Epoch[24 of 25]-Minibatch[ 911- 920, 83.83%]: CE.SM = 0.87686157 * 10240; Err = 0.28496094 * 10240; time = 0.0396s; samplesPerSecond = 258899.7
12/20/2016 15:29:09:  Epoch[24 of 25]-Minibatch[ 921- 930, 84.74%]: CE.SM = 0.88500977 * 10240; Err = 0.29042969 * 10240; time = 0.0401s; samplesPerSecond = 255342.5
12/20/2016 15:29:09:  Epoch[24 of 25]-Minibatch[ 931- 940, 85.65%]: CE.SM = 0.85800781 * 10240; Err = 0.28447266 * 10240; time = 0.0399s; samplesPerSecond = 256551.6
12/20/2016 15:29:09:  Epoch[24 of 25]-Minibatch[ 941- 950, 86.56%]: CE.SM = 0.90014648 * 10240; Err = 0.29345703 * 10240; time = 0.0397s; samplesPerSecond = 257882.5
12/20/2016 15:29:09:  Epoch[24 of 25]-Minibatch[ 951- 960, 87.47%]: CE.SM = 0.88172607 * 10240; Err = 0.28662109 * 10240; time = 0.0402s; samplesPerSecond = 254770.7
12/20/2016 15:29:09:  Epoch[24 of 25]-Minibatch[ 961- 970, 88.38%]: CE.SM = 0.88383789 * 10240; Err = 0.28691406 * 10240; time = 0.0415s; samplesPerSecond = 247032.7
12/20/2016 15:29:09:  Epoch[24 of 25]-Minibatch[ 971- 980, 89.29%]: CE.SM = 0.88377075 * 10240; Err = 0.28925781 * 10240; time = 0.0412s; samplesPerSecond = 248712.7
12/20/2016 15:29:10:  Epoch[24 of 25]-Minibatch[ 981- 990, 90.21%]: CE.SM = 0.90206299 * 10240; Err = 0.28984375 * 10240; time = 0.0407s; samplesPerSecond = 251621.8
12/20/2016 15:29:10:  Epoch[24 of 25]-Minibatch[ 991-1000, 91.12%]: CE.SM = 0.88867188 * 10240; Err = 0.29130859 * 10240; time = 0.0404s; samplesPerSecond = 253170.8
12/20/2016 15:29:10:  Epoch[24 of 25]-Minibatch[1001-1010, 92.03%]: CE.SM = 0.89605103 * 10240; Err = 0.29355469 * 10240; time = 0.0406s; samplesPerSecond = 252216.7
12/20/2016 15:29:10:  Epoch[24 of 25]-Minibatch[1011-1020, 92.94%]: CE.SM = 0.87774658 * 10240; Err = 0.29062500 * 10240; time = 0.0403s; samplesPerSecond = 253880.1
12/20/2016 15:29:10:  Epoch[24 of 25]-Minibatch[1021-1030, 93.85%]: CE.SM = 0.90101929 * 10240; Err = 0.28642578 * 10240; time = 0.0407s; samplesPerSecond = 251584.7
12/20/2016 15:29:10:  Epoch[24 of 25]-Minibatch[1031-1040, 94.76%]: CE.SM = 0.87816772 * 10240; Err = 0.28437500 * 10240; time = 0.0402s; samplesPerSecond = 254422.6
12/20/2016 15:29:10:  Epoch[24 of 25]-Minibatch[1041-1050, 95.67%]: CE.SM = 0.87081909 * 10240; Err = 0.28242187 * 10240; time = 0.0403s; samplesPerSecond = 253873.8
12/20/2016 15:29:10:  Epoch[24 of 25]-Minibatch[1051-1060, 96.58%]: CE.SM = 0.87282104 * 10240; Err = 0.27939453 * 10240; time = 0.0404s; samplesPerSecond = 253597.2
12/20/2016 15:29:10:  Epoch[24 of 25]-Minibatch[1061-1070, 97.49%]: CE.SM = 0.87446899 * 10240; Err = 0.28232422 * 10240; time = 0.0401s; samplesPerSecond = 255329.8
12/20/2016 15:29:10:  Epoch[24 of 25]-Minibatch[1071-1080, 98.41%]: CE.SM = 0.89628906 * 10240; Err = 0.28496094 * 10240; time = 0.0408s; samplesPerSecond = 250851.3
12/20/2016 15:29:10:  Epoch[24 of 25]-Minibatch[1081-1090, 99.32%]: CE.SM = 0.87015991 * 10240; Err = 0.28457031 * 10240; time = 0.0399s; samplesPerSecond = 256371.7
12/20/2016 15:29:10: Finished Epoch[24 of 25]: [Training] CE.SM = 0.87973847 * 1124823; Err = 0.28688158 * 1124823; totalSamplesSeen = 26995752; learningRatePerSample = 7.8124998e-05; epochTime=4.63386s
12/20/2016 15:29:10: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn.24'

12/20/2016 15:29:10: Starting Epoch 25: learning rate per sample = 0.000078  effective momentum = 0.900000  momentum as time constant = 9719.0 samples

12/20/2016 15:29:10: Starting minibatch loop.
12/20/2016 15:29:10:  Epoch[25 of 25]-Minibatch[   1-  10, 0.91%]: CE.SM = 0.87378635 * 10240; Err = 0.28925781 * 10240; time = 0.0449s; samplesPerSecond = 227894.9
12/20/2016 15:29:10:  Epoch[25 of 25]-Minibatch[  11-  20, 1.82%]: CE.SM = 0.88379650 * 10240; Err = 0.28515625 * 10240; time = 0.0405s; samplesPerSecond = 252852.0
12/20/2016 15:29:10:  Epoch[25 of 25]-Minibatch[  21-  30, 2.73%]: CE.SM = 0.89495049 * 10240; Err = 0.29736328 * 10240; time = 0.0398s; samplesPerSecond = 257389.9
12/20/2016 15:29:10:  Epoch[25 of 25]-Minibatch[  31-  40, 3.64%]: CE.SM = 0.87824211 * 10240; Err = 0.28369141 * 10240; time = 0.0400s; samplesPerSecond = 255878.5
12/20/2016 15:29:10:  Epoch[25 of 25]-Minibatch[  41-  50, 4.56%]: CE.SM = 0.86171608 * 10240; Err = 0.28398438 * 10240; time = 0.0405s; samplesPerSecond = 252833.3
12/20/2016 15:29:10:  Epoch[25 of 25]-Minibatch[  51-  60, 5.47%]: CE.SM = 0.89038200 * 10240; Err = 0.28847656 * 10240; time = 0.0402s; samplesPerSecond = 255024.5
12/20/2016 15:29:11:  Epoch[25 of 25]-Minibatch[  61-  70, 6.38%]: CE.SM = 0.86953850 * 10240; Err = 0.27919922 * 10240; time = 0.0401s; samplesPerSecond = 255119.8
12/20/2016 15:29:11:  Epoch[25 of 25]-Minibatch[  71-  80, 7.29%]: CE.SM = 0.86587029 * 10240; Err = 0.28212891 * 10240; time = 0.0405s; samplesPerSecond = 252590.0
12/20/2016 15:29:11:  Epoch[25 of 25]-Minibatch[  81-  90, 8.20%]: CE.SM = 0.86918030 * 10240; Err = 0.28085938 * 10240; time = 0.0412s; samplesPerSecond = 248797.3
12/20/2016 15:29:11:  Epoch[25 of 25]-Minibatch[  91- 100, 9.11%]: CE.SM = 0.87550659 * 10240; Err = 0.28701172 * 10240; time = 0.0408s; samplesPerSecond = 250937.3
12/20/2016 15:29:11:  Epoch[25 of 25]-Minibatch[ 101- 110, 10.02%]: CE.SM = 0.88347549 * 10240; Err = 0.28349609 * 10240; time = 0.0404s; samplesPerSecond = 253365.0
12/20/2016 15:29:11:  Epoch[25 of 25]-Minibatch[ 111- 120, 10.93%]: CE.SM = 0.88424911 * 10240; Err = 0.28632812 * 10240; time = 0.0406s; samplesPerSecond = 251943.7
12/20/2016 15:29:11:  Epoch[25 of 25]-Minibatch[ 121- 130, 11.85%]: CE.SM = 0.86623077 * 10240; Err = 0.28183594 * 10240; time = 0.0403s; samplesPerSecond = 254289.9
12/20/2016 15:29:11:  Epoch[25 of 25]-Minibatch[ 131- 140, 12.76%]: CE.SM = 0.88372345 * 10240; Err = 0.28466797 * 10240; time = 0.0403s; samplesPerSecond = 254195.2
12/20/2016 15:29:11:  Epoch[25 of 25]-Minibatch[ 141- 150, 13.67%]: CE.SM = 0.88094559 * 10240; Err = 0.28251953 * 10240; time = 0.0410s; samplesPerSecond = 249932.9
12/20/2016 15:29:11:  Epoch[25 of 25]-Minibatch[ 151- 160, 14.58%]: CE.SM = 0.85607605 * 10240; Err = 0.27880859 * 10240; time = 0.0400s; samplesPerSecond = 256320.4
12/20/2016 15:29:11:  Epoch[25 of 25]-Minibatch[ 161- 170, 15.49%]: CE.SM = 0.87388763 * 10240; Err = 0.28710938 * 10240; time = 0.0407s; samplesPerSecond = 251720.7
12/20/2016 15:29:11:  Epoch[25 of 25]-Minibatch[ 171- 180, 16.40%]: CE.SM = 0.87303925 * 10240; Err = 0.28613281 * 10240; time = 0.0412s; samplesPerSecond = 248314.7
12/20/2016 15:29:11:  Epoch[25 of 25]-Minibatch[ 181- 190, 17.31%]: CE.SM = 0.86883087 * 10240; Err = 0.28125000 * 10240; time = 0.0409s; samplesPerSecond = 250532.1
12/20/2016 15:29:11:  Epoch[25 of 25]-Minibatch[ 191- 200, 18.22%]: CE.SM = 0.89238739 * 10240; Err = 0.28847656 * 10240; time = 0.0407s; samplesPerSecond = 251621.8
12/20/2016 15:29:11:  Epoch[25 of 25]-Minibatch[ 201- 210, 19.13%]: CE.SM = 0.87737427 * 10240; Err = 0.28466797 * 10240; time = 0.0403s; samplesPerSecond = 254302.5
12/20/2016 15:29:11:  Epoch[25 of 25]-Minibatch[ 211- 220, 20.05%]: CE.SM = 0.87682800 * 10240; Err = 0.28837891 * 10240; time = 0.0409s; samplesPerSecond = 250654.8
12/20/2016 15:29:11:  Epoch[25 of 25]-Minibatch[ 221- 230, 20.96%]: CE.SM = 0.87815399 * 10240; Err = 0.28417969 * 10240; time = 0.0409s; samplesPerSecond = 250152.7
12/20/2016 15:29:11:  Epoch[25 of 25]-Minibatch[ 231- 240, 21.87%]: CE.SM = 0.86195068 * 10240; Err = 0.28037109 * 10240; time = 0.0410s; samplesPerSecond = 249890.2
12/20/2016 15:29:11:  Epoch[25 of 25]-Minibatch[ 241- 250, 22.78%]: CE.SM = 0.86418915 * 10240; Err = 0.28310547 * 10240; time = 0.0413s; samplesPerSecond = 247821.9
12/20/2016 15:29:11:  Epoch[25 of 25]-Minibatch[ 251- 260, 23.69%]: CE.SM = 0.87015076 * 10240; Err = 0.28378906 * 10240; time = 0.0411s; samplesPerSecond = 249203.0
12/20/2016 15:29:11:  Epoch[25 of 25]-Minibatch[ 261- 270, 24.60%]: CE.SM = 0.88921814 * 10240; Err = 0.29062500 * 10240; time = 0.0404s; samplesPerSecond = 253565.8
12/20/2016 15:29:11:  Epoch[25 of 25]-Minibatch[ 271- 280, 25.51%]: CE.SM = 0.89143372 * 10240; Err = 0.29277344 * 10240; time = 0.0409s; samplesPerSecond = 250073.3
12/20/2016 15:29:11:  Epoch[25 of 25]-Minibatch[ 281- 290, 26.42%]: CE.SM = 0.87731628 * 10240; Err = 0.27958984 * 10240; time = 0.0404s; samplesPerSecond = 253578.3
12/20/2016 15:29:11:  Epoch[25 of 25]-Minibatch[ 291- 300, 27.33%]: CE.SM = 0.87510071 * 10240; Err = 0.28750000 * 10240; time = 0.0400s; samplesPerSecond = 256147.3
12/20/2016 15:29:12:  Epoch[25 of 25]-Minibatch[ 301- 310, 28.25%]: CE.SM = 0.86876526 * 10240; Err = 0.28642578 * 10240; time = 0.0404s; samplesPerSecond = 253653.7
12/20/2016 15:29:12:  Epoch[25 of 25]-Minibatch[ 311- 320, 29.16%]: CE.SM = 0.88739929 * 10240; Err = 0.28593750 * 10240; time = 0.0406s; samplesPerSecond = 252191.9
12/20/2016 15:29:12:  Epoch[25 of 25]-Minibatch[ 321- 330, 30.07%]: CE.SM = 0.88685303 * 10240; Err = 0.29619141 * 10240; time = 0.0404s; samplesPerSecond = 253214.6
12/20/2016 15:29:12:  Epoch[25 of 25]-Minibatch[ 331- 340, 30.98%]: CE.SM = 0.87343750 * 10240; Err = 0.28349609 * 10240; time = 0.0409s; samplesPerSecond = 250067.2
12/20/2016 15:29:12:  Epoch[25 of 25]-Minibatch[ 341- 350, 31.89%]: CE.SM = 0.86928711 * 10240; Err = 0.28476563 * 10240; time = 0.0407s; samplesPerSecond = 251813.6
12/20/2016 15:29:12:  Epoch[25 of 25]-Minibatch[ 351- 360, 32.80%]: CE.SM = 0.87275391 * 10240; Err = 0.27880859 * 10240; time = 0.0413s; samplesPerSecond = 247989.9
12/20/2016 15:29:12:  Epoch[25 of 25]-Minibatch[ 361- 370, 33.71%]: CE.SM = 0.86646423 * 10240; Err = 0.28027344 * 10240; time = 0.0410s; samplesPerSecond = 249537.0
12/20/2016 15:29:12:  Epoch[25 of 25]-Minibatch[ 371- 380, 34.62%]: CE.SM = 0.88729858 * 10240; Err = 0.29345703 * 10240; time = 0.0403s; samplesPerSecond = 254289.9
12/20/2016 15:29:12:  Epoch[25 of 25]-Minibatch[ 381- 390, 35.54%]: CE.SM = 0.87984619 * 10240; Err = 0.28134766 * 10240; time = 0.0406s; samplesPerSecond = 252434.4
12/20/2016 15:29:12:  Epoch[25 of 25]-Minibatch[ 391- 400, 36.45%]: CE.SM = 0.87705383 * 10240; Err = 0.28427734 * 10240; time = 0.0406s; samplesPerSecond = 252179.5
12/20/2016 15:29:12:  Epoch[25 of 25]-Minibatch[ 401- 410, 37.36%]: CE.SM = 0.88798828 * 10240; Err = 0.29189453 * 10240; time = 0.0406s; samplesPerSecond = 252073.9
12/20/2016 15:29:12:  Epoch[25 of 25]-Minibatch[ 411- 420, 38.27%]: CE.SM = 0.86606750 * 10240; Err = 0.28183594 * 10240; time = 0.0403s; samplesPerSecond = 254031.3
12/20/2016 15:29:12:  Epoch[25 of 25]-Minibatch[ 421- 430, 39.18%]: CE.SM = 0.87674255 * 10240; Err = 0.28691406 * 10240; time = 0.0400s; samplesPerSecond = 256032.0
12/20/2016 15:29:12:  Epoch[25 of 25]-Minibatch[ 431- 440, 40.09%]: CE.SM = 0.87055359 * 10240; Err = 0.28564453 * 10240; time = 0.0416s; samplesPerSecond = 246390.8
12/20/2016 15:29:12:  Epoch[25 of 25]-Minibatch[ 441- 450, 41.00%]: CE.SM = 0.86878357 * 10240; Err = 0.28808594 * 10240; time = 0.0459s; samplesPerSecond = 223084.0
12/20/2016 15:29:12:  Epoch[25 of 25]-Minibatch[ 451- 460, 41.91%]: CE.SM = 0.88420105 * 10240; Err = 0.28515625 * 10240; time = 0.0453s; samplesPerSecond = 225928.9
12/20/2016 15:29:12:  Epoch[25 of 25]-Minibatch[ 461- 470, 42.82%]: CE.SM = 0.85320740 * 10240; Err = 0.27744141 * 10240; time = 0.0425s; samplesPerSecond = 241088.7
12/20/2016 15:29:12:  Epoch[25 of 25]-Minibatch[ 471- 480, 43.74%]: CE.SM = 0.88195496 * 10240; Err = 0.29072266 * 10240; time = 0.0466s; samplesPerSecond = 219832.1
12/20/2016 15:29:12:  Epoch[25 of 25]-Minibatch[ 481- 490, 44.65%]: CE.SM = 0.87927856 * 10240; Err = 0.28945312 * 10240; time = 0.0395s; samplesPerSecond = 259352.1
12/20/2016 15:29:12:  Epoch[25 of 25]-Minibatch[ 491- 500, 45.56%]: CE.SM = 0.87138062 * 10240; Err = 0.28476563 * 10240; time = 0.0408s; samplesPerSecond = 250869.7
12/20/2016 15:29:12:  Epoch[25 of 25]-Minibatch[ 501- 510, 46.47%]: CE.SM = 0.87334900 * 10240; Err = 0.28740234 * 10240; time = 0.0415s; samplesPerSecond = 246883.8
12/20/2016 15:29:12:  Epoch[25 of 25]-Minibatch[ 511- 520, 47.38%]: CE.SM = 0.90605469 * 10240; Err = 0.29902344 * 10240; time = 0.0411s; samplesPerSecond = 248972.7
12/20/2016 15:29:12:  Epoch[25 of 25]-Minibatch[ 521- 530, 48.29%]: CE.SM = 0.89900513 * 10240; Err = 0.29013672 * 10240; time = 0.0404s; samplesPerSecond = 253183.3
12/20/2016 15:29:12:  Epoch[25 of 25]-Minibatch[ 531- 540, 49.20%]: CE.SM = 0.88554687 * 10240; Err = 0.28710938 * 10240; time = 0.0402s; samplesPerSecond = 254859.5
12/20/2016 15:29:12:  Epoch[25 of 25]-Minibatch[ 541- 550, 50.11%]: CE.SM = 0.88118591 * 10240; Err = 0.28730469 * 10240; time = 0.0399s; samplesPerSecond = 256480.9
12/20/2016 15:29:13:  Epoch[25 of 25]-Minibatch[ 551- 560, 51.03%]: CE.SM = 0.87124023 * 10240; Err = 0.28789063 * 10240; time = 0.0431s; samplesPerSecond = 237647.7
12/20/2016 15:29:13:  Epoch[25 of 25]-Minibatch[ 561- 570, 51.94%]: CE.SM = 0.89294434 * 10240; Err = 0.29257813 * 10240; time = 0.0403s; samplesPerSecond = 253974.6
12/20/2016 15:29:13:  Epoch[25 of 25]-Minibatch[ 571- 580, 52.85%]: CE.SM = 0.85587158 * 10240; Err = 0.28046875 * 10240; time = 0.0409s; samplesPerSecond = 250391.2
12/20/2016 15:29:13:  Epoch[25 of 25]-Minibatch[ 581- 590, 53.76%]: CE.SM = 0.89339905 * 10240; Err = 0.29003906 * 10240; time = 0.0402s; samplesPerSecond = 254580.7
12/20/2016 15:29:13:  Epoch[25 of 25]-Minibatch[ 591- 600, 54.67%]: CE.SM = 0.88289795 * 10240; Err = 0.28457031 * 10240; time = 0.0400s; samplesPerSecond = 255872.1
12/20/2016 15:29:13:  Epoch[25 of 25]-Minibatch[ 601- 610, 55.58%]: CE.SM = 0.87572021 * 10240; Err = 0.28447266 * 10240; time = 0.0407s; samplesPerSecond = 251374.7
12/20/2016 15:29:13:  Epoch[25 of 25]-Minibatch[ 611- 620, 56.49%]: CE.SM = 0.89519043 * 10240; Err = 0.29023437 * 10240; time = 0.0410s; samplesPerSecond = 250018.3
12/20/2016 15:29:13:  Epoch[25 of 25]-Minibatch[ 621- 630, 57.40%]: CE.SM = 0.88944092 * 10240; Err = 0.29550781 * 10240; time = 0.0409s; samplesPerSecond = 250128.2
12/20/2016 15:29:13:  Epoch[25 of 25]-Minibatch[ 631- 640, 58.31%]: CE.SM = 0.88719482 * 10240; Err = 0.29218750 * 10240; time = 0.0411s; samplesPerSecond = 249348.6
12/20/2016 15:29:13:  Epoch[25 of 25]-Minibatch[ 641- 650, 59.23%]: CE.SM = 0.84609985 * 10240; Err = 0.27343750 * 10240; time = 0.0407s; samplesPerSecond = 251547.6
12/20/2016 15:29:13:  Epoch[25 of 25]-Minibatch[ 651- 660, 60.14%]: CE.SM = 0.89089355 * 10240; Err = 0.29218750 * 10240; time = 0.0403s; samplesPerSecond = 254315.2
12/20/2016 15:29:13:  Epoch[25 of 25]-Minibatch[ 661- 670, 61.05%]: CE.SM = 0.88078003 * 10240; Err = 0.29150391 * 10240; time = 0.0406s; samplesPerSecond = 251949.9
12/20/2016 15:29:13:  Epoch[25 of 25]-Minibatch[ 671- 680, 61.96%]: CE.SM = 0.89481812 * 10240; Err = 0.29267578 * 10240; time = 0.0405s; samplesPerSecond = 253014.4
12/20/2016 15:29:13:  Epoch[25 of 25]-Minibatch[ 681- 690, 62.87%]: CE.SM = 0.87452393 * 10240; Err = 0.28779297 * 10240; time = 0.0399s; samplesPerSecond = 256789.6
12/20/2016 15:29:13:  Epoch[25 of 25]-Minibatch[ 691- 700, 63.78%]: CE.SM = 0.90339355 * 10240; Err = 0.29521484 * 10240; time = 0.0408s; samplesPerSecond = 250882.0
12/20/2016 15:29:13:  Epoch[25 of 25]-Minibatch[ 701- 710, 64.69%]: CE.SM = 0.89127197 * 10240; Err = 0.28896484 * 10240; time = 0.0401s; samplesPerSecond = 255272.5
12/20/2016 15:29:13:  Epoch[25 of 25]-Minibatch[ 711- 720, 65.60%]: CE.SM = 0.87361450 * 10240; Err = 0.28310547 * 10240; time = 0.0404s; samplesPerSecond = 253616.0
12/20/2016 15:29:13:  Epoch[25 of 25]-Minibatch[ 721- 730, 66.51%]: CE.SM = 0.87463989 * 10240; Err = 0.28925781 * 10240; time = 0.0414s; samplesPerSecond = 247372.9
12/20/2016 15:29:13:  Epoch[25 of 25]-Minibatch[ 731- 740, 67.43%]: CE.SM = 0.88566895 * 10240; Err = 0.28759766 * 10240; time = 0.0414s; samplesPerSecond = 247343.0
12/20/2016 15:29:13:  Epoch[25 of 25]-Minibatch[ 741- 750, 68.34%]: CE.SM = 0.88154297 * 10240; Err = 0.28720703 * 10240; time = 0.0404s; samplesPerSecond = 253616.0
12/20/2016 15:29:13:  Epoch[25 of 25]-Minibatch[ 751- 760, 69.25%]: CE.SM = 0.86257324 * 10240; Err = 0.27900391 * 10240; time = 0.0403s; samplesPerSecond = 253974.6
12/20/2016 15:29:13:  Epoch[25 of 25]-Minibatch[ 761- 770, 70.16%]: CE.SM = 0.86390991 * 10240; Err = 0.27792969 * 10240; time = 0.0397s; samplesPerSecond = 257889.0
12/20/2016 15:29:13:  Epoch[25 of 25]-Minibatch[ 771- 780, 71.07%]: CE.SM = 0.87026367 * 10240; Err = 0.29111328 * 10240; time = 0.0396s; samplesPerSecond = 258618.5
12/20/2016 15:29:13:  Epoch[25 of 25]-Minibatch[ 781- 790, 71.98%]: CE.SM = 0.88863525 * 10240; Err = 0.28691406 * 10240; time = 0.0396s; samplesPerSecond = 258840.8
12/20/2016 15:29:14:  Epoch[25 of 25]-Minibatch[ 791- 800, 72.89%]: CE.SM = 0.88026123 * 10240; Err = 0.28554687 * 10240; time = 0.0395s; samplesPerSecond = 258991.4
12/20/2016 15:29:14:  Epoch[25 of 25]-Minibatch[ 801- 810, 73.80%]: CE.SM = 0.90115967 * 10240; Err = 0.29101562 * 10240; time = 0.0398s; samplesPerSecond = 257318.8
12/20/2016 15:29:14:  Epoch[25 of 25]-Minibatch[ 811- 820, 74.72%]: CE.SM = 0.86712036 * 10240; Err = 0.28164062 * 10240; time = 0.0404s; samplesPerSecond = 253170.8
12/20/2016 15:29:14:  Epoch[25 of 25]-Minibatch[ 821- 830, 75.63%]: CE.SM = 0.87908936 * 10240; Err = 0.28271484 * 10240; time = 0.0400s; samplesPerSecond = 255948.8
12/20/2016 15:29:14:  Epoch[25 of 25]-Minibatch[ 831- 840, 76.54%]: CE.SM = 0.86826782 * 10240; Err = 0.28349609 * 10240; time = 0.0398s; samplesPerSecond = 257034.6
12/20/2016 15:29:14:  Epoch[25 of 25]-Minibatch[ 841- 850, 77.45%]: CE.SM = 0.87559814 * 10240; Err = 0.28828125 * 10240; time = 0.0401s; samplesPerSecond = 255285.2
12/20/2016 15:29:14:  Epoch[25 of 25]-Minibatch[ 851- 860, 78.36%]: CE.SM = 0.88126221 * 10240; Err = 0.28662109 * 10240; time = 0.0399s; samplesPerSecond = 256615.9
12/20/2016 15:29:14:  Epoch[25 of 25]-Minibatch[ 861- 870, 79.27%]: CE.SM = 0.88843994 * 10240; Err = 0.29521484 * 10240; time = 0.0399s; samplesPerSecond = 256500.2
12/20/2016 15:29:14:  Epoch[25 of 25]-Minibatch[ 871- 880, 80.18%]: CE.SM = 0.89079590 * 10240; Err = 0.28955078 * 10240; time = 0.0400s; samplesPerSecond = 256076.8
12/20/2016 15:29:14:  Epoch[25 of 25]-Minibatch[ 881- 890, 81.09%]: CE.SM = 0.86031494 * 10240; Err = 0.28632812 * 10240; time = 0.0397s; samplesPerSecond = 257713.8
12/20/2016 15:29:14:  Epoch[25 of 25]-Minibatch[ 891- 900, 82.00%]: CE.SM = 0.89233398 * 10240; Err = 0.28955078 * 10240; time = 0.0402s; samplesPerSecond = 254511.1
12/20/2016 15:29:14:  Epoch[25 of 25]-Minibatch[ 901- 910, 82.92%]: CE.SM = 0.88482666 * 10240; Err = 0.28642578 * 10240; time = 0.0399s; samplesPerSecond = 256905.6
12/20/2016 15:29:14:  Epoch[25 of 25]-Minibatch[ 911- 920, 83.83%]: CE.SM = 0.89938354 * 10240; Err = 0.29658203 * 10240; time = 0.0400s; samplesPerSecond = 255712.3
12/20/2016 15:29:14:  Epoch[25 of 25]-Minibatch[ 921- 930, 84.74%]: CE.SM = 0.87849121 * 10240; Err = 0.28769531 * 10240; time = 0.0399s; samplesPerSecond = 256519.5
12/20/2016 15:29:14:  Epoch[25 of 25]-Minibatch[ 931- 940, 85.65%]: CE.SM = 0.86999512 * 10240; Err = 0.28193359 * 10240; time = 0.0394s; samplesPerSecond = 259694.1
12/20/2016 15:29:14:  Epoch[25 of 25]-Minibatch[ 941- 950, 86.56%]: CE.SM = 0.86710205 * 10240; Err = 0.27929688 * 10240; time = 0.0399s; samplesPerSecond = 256583.7
12/20/2016 15:29:14:  Epoch[25 of 25]-Minibatch[ 951- 960, 87.47%]: CE.SM = 0.88892212 * 10240; Err = 0.28974609 * 10240; time = 0.0401s; samplesPerSecond = 255495.4
12/20/2016 15:29:14:  Epoch[25 of 25]-Minibatch[ 961- 970, 88.38%]: CE.SM = 0.85966797 * 10240; Err = 0.27832031 * 10240; time = 0.0397s; samplesPerSecond = 257908.5
12/20/2016 15:29:14:  Epoch[25 of 25]-Minibatch[ 971- 980, 89.29%]: CE.SM = 0.88410645 * 10240; Err = 0.29238281 * 10240; time = 0.0392s; samplesPerSecond = 260971.5
12/20/2016 15:29:14:  Epoch[25 of 25]-Minibatch[ 981- 990, 90.21%]: CE.SM = 0.88132324 * 10240; Err = 0.28496094 * 10240; time = 0.0407s; samplesPerSecond = 251313.0
12/20/2016 15:29:14:  Epoch[25 of 25]-Minibatch[ 991-1000, 91.12%]: CE.SM = 0.87382202 * 10240; Err = 0.28671875 * 10240; time = 0.0397s; samplesPerSecond = 258246.7
12/20/2016 15:29:14:  Epoch[25 of 25]-Minibatch[1001-1010, 92.03%]: CE.SM = 0.87658691 * 10240; Err = 0.28798828 * 10240; time = 0.0402s; samplesPerSecond = 254910.3
12/20/2016 15:29:14:  Epoch[25 of 25]-Minibatch[1011-1020, 92.94%]: CE.SM = 0.87371216 * 10240; Err = 0.28515625 * 10240; time = 0.0402s; samplesPerSecond = 254669.4
12/20/2016 15:29:14:  Epoch[25 of 25]-Minibatch[1021-1030, 93.85%]: CE.SM = 0.88092651 * 10240; Err = 0.28574219 * 10240; time = 0.0407s; samplesPerSecond = 251813.6
12/20/2016 15:29:14:  Epoch[25 of 25]-Minibatch[1031-1040, 94.76%]: CE.SM = 0.89332886 * 10240; Err = 0.28515625 * 10240; time = 0.0410s; samplesPerSecond = 250036.6
12/20/2016 15:29:15:  Epoch[25 of 25]-Minibatch[1041-1050, 95.67%]: CE.SM = 0.86897583 * 10240; Err = 0.28769531 * 10240; time = 0.0400s; samplesPerSecond = 255910.4
12/20/2016 15:29:15:  Epoch[25 of 25]-Minibatch[1051-1060, 96.58%]: CE.SM = 0.89067383 * 10240; Err = 0.28994141 * 10240; time = 0.0398s; samplesPerSecond = 257228.3
12/20/2016 15:29:15:  Epoch[25 of 25]-Minibatch[1061-1070, 97.49%]: CE.SM = 0.87401733 * 10240; Err = 0.28623047 * 10240; time = 0.0406s; samplesPerSecond = 252310.0
12/20/2016 15:29:15:  Epoch[25 of 25]-Minibatch[1071-1080, 98.41%]: CE.SM = 0.87929077 * 10240; Err = 0.28750000 * 10240; time = 0.0399s; samplesPerSecond = 256641.6
12/20/2016 15:29:15:  Epoch[25 of 25]-Minibatch[1081-1090, 99.32%]: CE.SM = 0.89161987 * 10240; Err = 0.29287109 * 10240; time = 0.0401s; samplesPerSecond = 255412.6
12/20/2016 15:29:15: Finished Epoch[25 of 25]: [Training] CE.SM = 0.87857001 * 1124823; Err = 0.28655442 * 1124823; totalSamplesSeen = 28120575; learningRatePerSample = 7.8124998e-05; epochTime=4.65598s
12/20/2016 15:29:15: SGD: Saving checkpoint model '/tmp/cntk-test-20161220143826.605487/Speech/HTKDeserializers/TIMIT_TrainWithPreTrain@release_gpu/exp/TrainWithPreTrain/model/cntkSpeech.dnn'

12/20/2016 15:29:15: Action "train" complete.

12/20/2016 15:29:15: __COMPLETED__
