Index out of range error

[RaviTej310]

I am trying to implement a different version of this simple recurrent network which accepts three image of 200x200 and predicts a fourth. I modified it as:

require 'rnn'
require 'image'

-- hyper-parameters 
batchSize = 1
rho = 3 -- sequence length
hiddenSize = 7
nIndex = 10
lr = 0.1


-- build simple recurrent neural network
local r = nn.Recurrent(
   hiddenSize, nn.LookupTable(nIndex, hiddenSize), 
   nn.Linear(hiddenSize, hiddenSize), nn.Sigmoid(), 
   rho
)

local rnn = nn.Sequential()
   :add(r)
   :add(nn.Linear(hiddenSize, nIndex))
   :add(nn.LogSoftMax())

-- wrap the non-recurrent module (Sequential) in Recursor.
-- This makes it a recurrent module
-- i.e. Recursor is an AbstractRecurrent instance
rnn = nn.Recursor(rnn, rho)

print(rnn)

-- build criterion

criterion = nn.ClassNLLCriterion()

offsets = {}
for i=1,batchSize do
   table.insert(offsets, math.ceil(math.random()*sequence:size(1)))
end
--print(offsets)
offsets = torch.LongTensor(offsets)
--print(offsets)

-- training
local iteration = 1
count=1
datasetname = '/home/srilatha/Desktop/Research_intern/Data_sets/Cropped_data_set'
j=1
dir_name=datasetname.."/"..j.."/"
image_size=200

while count>0 do
   -- 1. create a sequence of rho time-steps

   local inputs, targets = {}, {}
   for step=1,rho do
      filename=dir_name..step..".JPG"
      filename1=dir_name..(step+1)..".JPG" 
      img1=image.load(filename)
      img1=image.scale(img1,image_size,image_size)
      img1=torch.reshape(img1,1,image_size*image_size)
      inputs[step]=img1
      img2=image.load(filename1)
      img2=image.scale(img2,image_size,image_size)
      img2=torch.reshape(img2,1,image_size*image_size)
      targets[step] = img2
   end
   print(inputs[2][1][1])
   print(targets[1][1][1])

   -- 2. forward sequence through rnn

   rnn:zeroGradParameters() 
   rnn:forget() -- forget all past time-steps

   --print(rnn:forward(inputs[1][1]))
   local outputs, err = {}, 0
   for step=1,rho do
      outputs[step] = rnn:forward(inputs[step])
      print(outputs[step])
      err = err + criterion:forward(outputs[step], targets[step])
   end

   print(string.format("Iteration %d ; NLL err = %f ", iteration, err))

   -- 3. backward sequence through rnn (i.e. backprop through time)

   local gradOutputs, gradInputs = {}, {}
   for step=rho,1,-1 do -- reverse order of forward calls
      gradOutputs[step] = criterion:backward(outputs[step], targets[step])
      gradInputs[step] = rnn:backward(inputs[step], gradOutputs[step])
   end

   -- 4. update

   rnn:updateParameters(lr)

   iteration = iteration + 1
   count=count-1
end

But I'm getting an index out of range error:

/home/srilatha/torch/install/share/lua/5.1/nn/Container.lua:67: 
In 1 module of nn.Sequential:
In 1 module of nn.Sequential:
.../srilatha/torch/install/share/lua/5.1/nn/LookupTable.lua:73: index out of range at /tmp/luarocks_torch-scm-1-3482/torch7/lib/TH/generic/THTensorMath.c:156
stack traceback:
    [C]: in function 'index'
    .../srilatha/torch/install/share/lua/5.1/nn/LookupTable.lua:73: in function <.../srilatha/torch/install/share/lua/5.1/nn/LookupTable.lua:68>
    [C]: in function 'xpcall'
    /home/srilatha/torch/install/share/lua/5.1/nn/Container.lua:63: in function 'rethrowErrors'
    ...e/srilatha/torch/install/share/lua/5.1/nn/Sequential.lua:44: in function 'updateOutput'
    ...e/srilatha/torch/install/share/lua/5.1/rnn/Recurrent.lua:68: in function <...e/srilatha/torch/install/share/lua/5.1/rnn/Recurrent.lua:64>
    [C]: in function 'xpcall'
    /home/srilatha/torch/install/share/lua/5.1/nn/Container.lua:63: in function 'rethrowErrors'
    ...e/srilatha/torch/install/share/lua/5.1/nn/Sequential.lua:44: in function 'updateOutput'
    /home/srilatha/torch/install/share/lua/5.1/rnn/Recursor.lua:25: in function 'forward'
    /home/srilatha/Desktop/Research_intern/rnn_simple.lua:75: in main chunk
    [C]: in function 'dofile'
    [string "_RESULT={dofile('/home/srilatha/Desktop/Resea..."]:1: in main chunk
    [C]: in function 'xpcall'
    /home/srilatha/torch/install/share/lua/5.1/trepl/init.lua:651: in function </home/srilatha/torch/install/share/lua/5.1/trepl/init.lua:560>
    [string "require('trepl')()"]:1: in main chunk

What is going wrong and how can I fix this?

[nicholas-leonard]

@RaviTej310 You cannot use a LookupTable with images. You should instead use a Linear or SpatialConvolution. PR #280 is a very rough draft of how to use SpatialConvolution with Recurrence. It does not run, but the part about building the model should give you an idea.

[scalastic]

Sorry but I get the same error with LSTM on stock indices (not images) :

In 1 module of nn.Sequential:
...DEV/TOOLS/torch/install/share/lua/5.1/nn/LookupTable.lua:73: index out of range at /Users/jeanjerome/DEV/TOOLS/torch/pkg/torch/lib/TH/generic/THTensorMath.c:156
stack traceback:
    [C]: in function 'index'
    ...DEV/TOOLS/torch/install/share/lua/5.1/nn/LookupTable.lua:73: in function <...DEV/TOOLS/torch/install/share/lua/5.1/nn/LookupTable.lua:68>
    [C]: in function 'xpcall'
    ...e/DEV/TOOLS/torch/install/share/lua/5.1/nn/Container.lua:63: in function 'rethrowErrors'
    .../DEV/TOOLS/torch/install/share/lua/5.1/nn/Sequential.lua:44: in function 'forward'
    ...anjerome/DEV/PROJECTS/FINAPP/FIN_POC/POC-05/src/main.lua:348: in function 'opfunc'
    ...rome/DEV/TOOLS/torch/install/share/lua/5.1/optim/sgd.lua:44: in function 'optimMethod'
    ...anjerome/DEV/PROJECTS/FINAPP/FIN_POC/POC-05/src/main.lua:372: in function 'train'
    ...anjerome/DEV/PROJECTS/FINAPP/FIN_POC/POC-05/src/main.lua:478: in main chunk
    [C]: at 0x010b097cd0

Almost the same stack :(

The problem occurs when feeding the LookupTable on mini-batch.

Could someone points us to a relevant example (with at least dim-2d input Tensor not torch.randperm() please.......!!!) ?

Many thanks

[nicholas-leonard]

@scalastic Usually, index out of range error occurs when you have an invalid index. For example, with a lookup table of 100 indices, only indexes between 1 and 100 will be valid. Here is a 2d example:

-- 3 x 4 tensor of indices randomly sampled between 1 and 100
input = torch.LongTensor(3,4):random(1,100)
-- lookup table consisting of 100 embeddings of 5 units each:
lookup = nn.LookupTable(100, 5)
-- forward indices through
lookup:forward(input)
(1,.,.) = 
 -0.3788  0.7058  0.4782 -1.0016  0.3914
  0.6950 -1.1974  1.4737  0.0400  0.3055
 -0.4760 -0.7315  0.3597 -0.8322 -0.8582
  0.1845 -0.9815 -0.6764 -0.4099  2.3405

(2,.,.) = 
 -0.4075 -0.3032  0.4292  0.6594  1.1332
 -0.3993  1.6485 -0.2334 -0.4767 -1.1654
  0.2760  0.8907  0.4681  0.2643  0.7932
 -0.6268  0.6494 -1.2751  2.0526 -0.3467

(3,.,.) = 
 -0.2379 -0.5717 -0.7413  0.9472  0.6435
 -1.1906 -1.0219  1.1680 -0.3493 -0.0208
  0.1248  0.5194 -1.0458  1.8812  0.0879
 -1.4082 -0.2250 -0.4062 -0.8765 -0.2728
[torch.DoubleTensor of size 3x4x5]

Here is an example of a similar error (line 75 instead of 73 because 2d instead of 1d):

-- out of range (102 > 100)
input[1][1] = 102
lookup:forward(input)
/usr/local/share/lua/5.1/nn/LookupTable.lua:75: index out of range at /tmp/luarocks_torch-scm-1-9702/torch7/lib/TH/generic/THTensorMath.c:156
stack traceback:
    [C]: in function 'index'
    /usr/local/share/lua/5.1/nn/LookupTable.lua:75: in function 'forward'
    [string "_RESULT={lookup:forward(input)}"]:1: in main chunk
    [C]: in function 'xpcall'
    /usr/local/share/lua/5.1/trepl/init.lua:651: in function 'repl'
    /usr/local/lib/luarocks/rocks/trepl/scm-1/bin/th:199: in main chunk
    [C]: at 0x00405e60  

[scalastic]

@nicholas-leonard Thank you for your answer. I'm on the move right now, I'll watch all this in few days and I'll give you my feedback.