Use the execute packet, get nothing output

Jul 25, 2015 at 9:49 AM
Hi all,

I use the execute packet (CNTK-20150415-Windows-64bit-MKL11.2-CUDA7.0+Argon.zip) directly, I test several config example using the cn.exe, but i can't get anything except for the xxx.log, I don't why?? who can help me.

Thanks.
Coordinator
Jul 25, 2015 at 5:40 PM
we don't ship data (we are not allowed to). you need to make sure you have data and change the config to point to your data. Another common error I observed is that some people forgot to specify configFile=
Jul 26, 2015 at 2:26 AM
yes, I had clone the source code, and test the Simple Demo. I use the data is the Simple folder. It contains SimpleDataTest.txt, SimpleDataTrain.txt, and SimpleMapping.txt. I copy the three data file and the Simple.config to the Execute Folder(contains cn.exe). I modify the Simple.config data part. Then I run the cn.exe configFile=Simple.config. it output nothing except for xxx.log file, and it run very quickly(just 5 seconds). I don't why?? I have debug whole day. Thanks. The modify config file is in the following list. Thanks..


RootDir=..
command=Simple_Demo:Simple_Demo_Output

deviceId=-1 for CPU, >=0 for GPU devices

DeviceNumber=-1
stderr=Demo

precision=float

modelPath=models/simple.dnn
deviceId=$DeviceNumber$

outputNodeNames=ScaledLogLikelihood
traceLevel=1

TRAINING CONFIG (Simple, Fixed LR)

Simple_Demo=[
action=train

# Notation xxx:yyy*n:zzz is equivalent to xxx,
#  then yyy repeated n times, then zzz
# example: 10:20*3:5 is equivalent to 10:20:20:20:5
SimpleNetworkBuilder=[
    # 2 input, 2 50-element hidden, 2 output
    layerSizes=2:50*2:2
    trainingCriterion=CrossEntropyWithSoftmax
    evalCriterion=ErrorPrediction
    layerTypes=Sigmoid
    initValueScale=1.0
    applyMeanVarNorm=true
    uniformInit=true
    needPrior=true
]

SGD=[
# epochSize=0 means epochSize is the size of 
    # the training set. Must be evenly divisible 
    # into number of data frames.
    epochSize=0 
    minibatchSize=25    
    learningRatesPerMB=0.5:0.2*20:0.1
    momentumPerMB=0.9
    dropoutRate=0.0
    maxEpochs=50
]

# Parameter values for the reader
reader=[
  # reader to use
  readerType=UCIFastReader
  file=SimpleDataTrain1.txt

  miniBatchMode=Partial
  randomize=Auto
  verbosity=1   

  features=[
      dim=2      # two-dimensional input data
      start=0    # Start with first element on line
  ]

  labels=[
    start=2      # Skip two elements
    dim=1        # One label dimension
    labelDim=2   # Two labels possible
    labelMappingFile=SimpleMapping1.txt
  ]
]
]

OUTPUT RESUTLS (Simple)

Simple_Demo_Output=[
action=write


# Parameter values for the reader
reader=[
  # reader to use
  readerType=UCIFastReader
  file=SimpleDataTest1.txt
  features=[
      dim=2
      start=0
  ]
  labels=[
  start=2
    dim=1
    labelDim=2
    labelMappingFile=SimpleMapping.txt
  ]
]
outputPath=output.txt    # Dump output as text
]
Coordinator
Jul 26, 2015 at 2:54 AM
all the output info has been saved in the log file. You just need to open that file to see how it goes.
Nov 26, 2015 at 1:21 PM
Edited Nov 27, 2015 at 12:39 AM
you find out what's wrong?

I got:
1 starting at epoch 0 counting lines to determine record count
2
3 234262 records found
4 starting epoch 0 at record count 0, and file position 0
5 already there from last epoch

but the output file get nothing...


It caused by wrong format test data...
Coordinator
Dec 9, 2015 at 10:35 PM
Hi flyaway1217,

there are now updated binaries as well as instructions to run the demos, e.g., http://cntk.codeplex.com/SourceControl/latest#Demos/Simple2d/README.md.

Would care having another try?

I'm not sure what went wrong last time...

Thanks, Mark