TypeError: iteration over a 0-d tensor

[TreyLawrence]

Hello! I'm running into an issue with training the Penn Treeback model:
python main.py --data data/penn --dropouti 0.4 --dropoutl 0.29 --dropouth 0.225 --seed 28 --batch_size 12 --lr 20.0 --epoch 1000 --nhid 960 --nhidlast 620 --emsize 280 --n_experts 15 --save PTB --single_gpu

Traceback (most recent call last):
  File "main.py", line 260, in <module>
    train()
  File "main.py", line 202, in train
    hidden[s_id] = repackage_hidden(hidden[s_id])
  File "/mos/utils.py", line 12, in repackage_hidden
    return tuple(repackage_hidden(v) for v in h)
  File "/mos/utils.py", line 12, in <genexpr>
    return tuple(repackage_hidden(v) for v in h)
  File "/mos/utils.py", line 12, in repackage_hidden
    return tuple(repackage_hidden(v) for v in h)
  File "/mos/utils.py", line 12, in <genexpr>
    return tuple(repackage_hidden(v) for v in h)
  File "/mos/utils.py", line 12, in repackage_hidden
    return tuple(repackage_hidden(v) for v in h)
  File "/mos/utils.py", line 12, in <genexpr>
    return tuple(repackage_hidden(v) for v in h)
  File "/mos/utils.py", line 12, in repackage_hidden
    return tuple(repackage_hidden(v) for v in h)
  File "/mos/utils.py", line 12, in <genexpr>
    return tuple(repackage_hidden(v) for v in h)
  File "/mos/utils.py", line 12, in repackage_hidden
    return tuple(repackage_hidden(v) for v in h)
  File "/mos/utils.py", line 12, in <genexpr>
    return tuple(repackage_hidden(v) for v in h)
  File "/mos/utils.py", line 12, in repackage_hidden
    return tuple(repackage_hidden(v) for v in h)
  File "/usr/local/lib/python3.5/dist-packages/torch/tensor.py", line 360, in __iter__
    raise TypeError('iteration over a 0-d tensor')
TypeError: iteration over a 0-d tensor

[TreyLawrence]

This is with python3. If I try with python2 I get:

Traceback (most recent call last):
  File "main.py", line 116, in <module>
    corpus = data.Corpus(args.data)
  File "/spell/data.py", line 30, in __init__
    self.train = self.tokenize(os.path.join(path, 'train.txt'))
  File "/spell/data.py", line 38, in tokenize
    with open(path, 'r', encoding='utf-8') as f:
TypeError: 'encoding' is an invalid keyword argument for this function

[zihangdai]

What Pytorch version you are using?
print(torch.__version__)

[TreyLawrence]

Thanks for the quick reply!
I'm using 0.4.0

[kimiyoung]

You will need pytorch 0.2.0 to run our code and replicate the results.

[guillaume-chevalier]

I'd like an update is possible

[octavian-ganea]

I'm also interested. Is it possible to easily fix the code to work with the latest pytorch version ? Thanks!

[badripatro]

Hi ,
I am getting same error

File "train/train_GD.py", line 272, in train
ques_hidden = repackage_hidden(ques_hidden, batch_size)
File "/home/badri/gpuserver150/data150/workspace_project/visual_dialog/visDial_pytorch_vae_ale/misc/utils.py", line 22, in repackage_hidden
return tuple(repackage_hidden(v, batch_size) for v in h)
File "/home/badri/gpuserver150/data150/workspace_project/visual_dialog/visDial_pytorch_vae_ale/misc/utils.py", line 22, in
return tuple(repackage_hidden(v, batch_size) for v in h)
File "/home/badri/gpuserver150/data150/workspace_project/visual_dialog/visDial_pytorch_vae_ale/misc/utils.py", line 22, in repackage_hidden
return tuple(repackage_hidden(v, batch_size) for v in h)
File "/home/badri/gpuserver150/data150/workspace_project/visual_dialog/visDial_pytorch_vae_ale/misc/utils.py", line 22, in
return tuple(repackage_hidden(v, batch_size) for v in h)
File "/home/badri/gpuserver150/data150/workspace_project/visual_dialog/visDial_pytorch_vae_ale/misc/utils.py", line 22, in repackage_hidden
return tuple(repackage_hidden(v, batch_size) for v in h)
File "/home/badri/gpuserver150/data150/workspace_project/visual_dialog/visDial_pytorch_vae_ale/misc/utils.py", line 22, in
return tuple(repackage_hidden(v, batch_size) for v in h)
File "/home/badri/gpuserver150/data150/workspace_project/visual_dialog/visDial_pytorch_vae_ale/misc/utils.py", line 22, in repackage_hidden
return tuple(repackage_hidden(v, batch_size) for v in h)
File "/home/badri/gpuserver150/data150/workspace_project/visual_dialog/visDial_pytorch_vae_ale/misc/utils.py", line 22, in
return tuple(repackage_hidden(v, batch_size) for v in h)
File "/home/badri/gpuserver150/data150/workspace_project/visual_dialog/visDial_pytorch_vae_ale/misc/utils.py", line 22, in repackage_hidden
return tuple(repackage_hidden(v, batch_size) for v in h)
File "/home/badri/anaconda3/lib/python3.6/site-packages/torch/tensor.py", line 381, in iter
raise TypeError('iteration over a 0-d tensor')
TypeError: iteration over a 0-d tensor
Exception ignored in: <bound method _DataLoaderIter.del of <torch.utils.data.dataloader._DataLoaderIter object at 0x7f16c0be6ac8>>

---

I need in pytorch version 0.4 to run my code.

[huu4ontocord]

Has this been fixed? I think the problem is the function never finds a Variable and it recurses all the way to the single element in a tensor.

[huu4ontocord]

I think the answer is to use isinstanceof instead of type in repackage_hidden

if isinstance(h, Variable):

[kimiyoung]

I think this should be fairly easy to fix, but sorry that we don't have time for this at this point. You are welcome to initiate a pull request.

[kimiyoung]

resolved by #9

[Harrite77]

Execuse me, I got the same error today, could you tell me any method to solve it? thank you!

[c04s84]

if isinstance(h, Variable):

this solved my problem. thank you so much!