textsampler.lua:38: bad argument #3

[ThorJonsson]

Hi! I'm getting the following error running the recurrentlanguagemodel code.

`FileLogger: log will be written to /home/user/save/comp:1458431781:1/log    
==> epoch # 1 for optimizer :   
 [======================================== 4000/4000 ==================================>]  Tot: 12s48ms | Step: 4ms     
==> example speed = 216.87828023852 examples/s  
 [======================================== 24000/24000 ================================>]  Tot: 2m49s | Step: 7ms       
/home/user/torch/install/bin/luajit: ...2/torch/install/share/lua/5.1/dp/sampler/textsampler.lua:38: bad argument #3 to 'narrow' (out of range at /home/user/torch/pkg/torch/lib/TH/generic/THTensor.c:351)
stack traceback:
        [C]: in function 'narrow'
        ...2/torch/install/share/lua/5.1/dp/sampler/textsampler.lua:38: in function 'sampler'
        ...torch/install/share/lua/5.1/dp/propagator/propagator.lua:117: in function 'propagateEpoch'
        ...torch/install/share/lua/5.1/dp/propagator/experiment.lua:116: in function 'run'
        lstm_lm_example.lua:360: in main chunk
        [C]: in function 'dofile'
        ...hj92/torch/install/lib/luarocks/rocks/trepl/scm-1/bin/th:145: in main chunk
        [C]: at 0x00405d70

I can get around this error by setting

trainOnly = true`

This is the model:

{
   accUpdate : false
   batchSize : 64
   bidirectional : true
   cuda : false
   cutoffNorm : -1
   dataPath : "/home/user/data"
   dataset : "PennTreeBank"
   decayFactor : 0.001
   dropout : true
   dropoutProb : 0.5
   evalSize : 100
   forestGaterSize : "{}"
   hiddenSize : {200,200}
   learningRate : 0.1
   lrDecay : "linear"
   lstm : true
   maxEpoch : 400
   maxOutNorm : 2
   maxTries : 30
   maxWait : 4
   minLR : 1e-05
   momentum : 0
   progress : true
   rho : 20
   saturateEpoch : 300
   schedule : {}
   silent : false
   small : false
   softmaxforest : false
   softmaxtree : false
   testFile : "test.txt"
   tiny : false
   trainEpochSize : 4000
   trainFile : "train.txt"
   trainOnly : false
   uniform : -1
   useDevice : 1
   validEpochSize : 24000
   validFile : "valid.txt"
   xpPath : ""
   zeroFirst : false
}       
Warning : the Perplexity of a bidirectional RNN/LSTM isn't necessarily mathematically valid as it uses P(x_t|x_{/neq t}) instead of P(x_t|x_{<t}), which is used for unidirectional RNN/LSTMs. You can however still use predictions to measure pseudo-likelihood.      
Language Model :        
nn.Sequential {
  [input -> (1) -> (2) -> (3) -> (4) -> (5) -> output]
  (1): nn.LookupTable
  (2): nn.Dropout(0.5, busy)
  (3): nn.SplitTable
  (4): nn.BiSequencerLM {
    (  fwd  ): nn.Sequential {
      |      [input -> (1) -> (2) -> (3) -> (4) -> output]
      |      (1): nn.Sequencer @ nn.FastLSTM(200 -> 200)
      |      (2): nn.Sequencer @ nn.Recursor @ nn.Dropout(0.5, busy)                                                                                                03:55
      |      (3): nn.Sequencer @ nn.FastLSTM(200 -> 200)                                                                                                            03:58
      |      (4): nn.Sequencer @ nn.Recursor @ nn.Dropout(0.5, busy)
      |    }
    (  bwd  ): nn.Sequential {
      |      [input -> (1) -> (2) -> (3) -> output]
      |      (1): nn.ReverseTable
      |      (2): nn.Sequential {
      |        [input -> (1) -> (2) -> (3) -> (4) -> output]
      |        (1): nn.Sequencer @ nn.FastLSTM(200 -> 200)
      |        (2): nn.Sequencer @ nn.Recursor @ nn.Dropout(0.5, busy)
      |        (3): nn.Sequencer @ nn.FastLSTM(200 -> 200)
      |        (4): nn.Sequencer @ nn.Recursor @ nn.Dropout(0.5, busy)
      |      }
      |      (3): nn.ReverseTable
      |    }
    ( merge ): nn.Sequential {
      |      [input -> (1) -> (2) -> output]
      |      (1): nn.ZipTable
      |      (2): nn.Sequencer @ nn.Recursor @ nn.JoinTable
      |    }
  }
  (5): nn.Sequencer @ nn.Recursor @ nn.Sequential {
    [input -> (1) -> (2) -> output]
    (1): nn.Linear(400 -> 9663)
    (2): nn.LogSoftMax
  }
}

I would be very thankful for any help you can provide to solve this issue.