Print statement doesn't automatically flush stdout/stderr on python 3 about ncrfpp HOT 8 CLOSED

Hi @qiuwei thanks for the suggestion.

Can you give a more detail description about the flush operator? As I already used several sys.stdout.flush() to force the system print the logs.

from ncrfpp.

If I use ncrfpp with python3, the log is not written to the stdout immediately as with python 2.
That's default behavior at the moment if don't make any modifications to the code at the moment.

With the sample data, it's hard to notice though. On my own dataset, I noticed the log was flushed only after every 20ish epochs.

from ncrfpp.

Traceback (most recent call last):
File "main.py", line 438, in
train(data)
File "main.py", line 328, in train
batch_word, batch_features, batch_wordlen, batch_wordrecover, batch_char, batch_charlen, batch_charrecover, batch_label, mask = batchify_with_label(instance, data.HP_gpu)
File "main.py", line 236, in batchify_with_label
mask[idx, :seqlen] = torch.Tensor([1]*seqlen)
TypeError: mul() received an invalid combination of arguments - got (list), but expected one of:

• (Tensor other)
  didn't match because some of the arguments have invalid types: (list)
• (float other)
  didn't match because some of the arguments have invalid types: (list)

from ncrfpp.

Epoch: 0/1
Learning rate is set as: 0.015
Traceback (most recent call last):
File "main.py", line 438, in
train(data)
File "main.py", line 328, in train
batch_word, batch_features, batch_wordlen, batch_wordrecover, batch_char, batch_charlen, batch_charrecover, batch_label, mask = batchify_with_label(instance, data.HP_gpu)
File "main.py", line 236, in batchify_with_label
mask[idx, :seqlen] = torch.Tensor([1]*seqlen)
TypeError: mul() received an invalid combination of arguments - got (list), but expected one of:

• (Tensor other)
  didn't match because some of the arguments have invalid types: (list)
• (Number other)
  didn't match because some of the arguments have invalid types: (list)

from ncrfpp.

@nthoangcute please confirm you used the right PyTorch version. PyTorch 0.3

from ncrfpp.

@jiesutd opss! My pytorch is currently 0.5.0a0.

Traceback (most recent call last):
  File "main.py", line 440, in <module>
    train(data)
  File "main.py", line 284, in train
    model = SeqModel(data)
  File "/root/vietbt/NCRFpp/model/seqmodel.py", line 31, in __init__
    self.word_hidden = WordSequence(data)
  File "/root/vietbt/NCRFpp/model/wordsequence.py", line 71, in __init__
    self.lstm = self.lstm.cuda()
  File "/opt/conda/lib/python3.6/site-packages/torch/nn/modules/module.py", line 258, in cuda
    return self._apply(lambda t: t.cuda(device))
  File "/opt/conda/lib/python3.6/site-packages/torch/nn/modules/rnn.py", line 112, in _apply
    self.flatten_parameters()
  File "/opt/conda/lib/python3.6/site-packages/torch/nn/modules/rnn.py", line 105, in flatten_parameters
    self.batch_first, bool(self.bidirectional))
RuntimeError: CuDNN error: CUDNN_STATUS_INTERNAL_ERROR

Maybe I have to downgrade my pytorch version to 0.3

from ncrfpp.

@UPWHY please use the PyTorch 0.3 version.

from ncrfpp.

我现在用的Python是2.7，pytorch是0.3，但是还是报错：Traceback (most recent call last):
File "main.py", line 438, in
train(data)
File "main.py", line 328, in train
batch_word, batch_features, batch_wordlen, batch_wordrecover, batch_char, batch_charlen, batch_charrecover, batch_label, mask = batchify_with_label(instance, data.HP_gpu)
File "main.py", line 236, in batchify_with_label
mask[idx, :seqlen] = torch.Tensor([1]*seqlen)
TypeError: mul() received an invalid combination of arguments - got (list), but expected one of:

• (Tensor other)
  didn't match because some of the arguments have invalid types: (list)
• (Number other)
  didn't match because some of the arguments have invalid types: (list)

from ncrfpp.