TypeError: forward() missing 2 required positional arguments: 'attn' and 'src_map'

[ymym3412]

Hi, I get some error when I did python VisServer.py .
My enviroment is this.
Ubuntu 16.04
Anaconda
Python 3.6

I installed Seq2Seq-Vis with conda.

git clone https://github.com/HendrikStrobelt/Seq2Seq-Vis.git
cd Seq2Seq-Vis
source setup_cpu.sh
cd ..
source Seq2Seq-Vis/setup_onmt_custom.sh

And, I trained summarization model with my own data.
I trained following command.

python preprocess.py -train_src data/newspaper/train_source.txt -train_tgt data/newspaper/train_target.txt -valid_src data/newspaper/dev_source.txt -valid_tgt data/newspaper/dev_target.txt -save_data data/newspaper/newspaper -src_seq_length 6705 -tgt_seq_length 30 -src_seq_length_trunc 300 -tgt_seq_length_trunc 30 -dynamic_dict -share_vocab -max_shard_size 134217728

python train.py -save_model models/newspapre -data data/newspaper/newspaper -copy_attn -global_attention dot -coverage_attn -word_vec_size 128 -rnn_size 512 -layers 1 -encoder_type brnn -epochs 9 -max_grad_norm 2 -dropout 0.5 -batch_size 32 -optim adagrad -learning_rate 0.15 -adagrad_accumulator_init 0.1 -seed 777 -gpuid 0

After training, I create state.h5 , encoder.faiss , decoder.faiss , context.faiss.
Then I do python VisServer.py , but I got following error.

(s2sv) ymym@ymym:~/workspace/OpenNMT-py$ python VisServer.py 
/home/ymym/anaconda3/envs/s2sv/lib/python3.6/site-packages/h5py/__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.
  from ._conv import register_converters as _register_converters
/home/ymym/anaconda3/envs/s2sv/lib/python3.6/site-packages/torch/nn/modules/rnn.py:38: UserWarning: dropout option adds dropout after all but last recurrent layer, so non-zero dropout expects num_layers greater than 1, but got dropout=0.5 and num_layers=1
  "num_layers={}".format(dropout, num_layers))
Loading model parameters.
partial: []
/home/ymym/anaconda3/envs/s2sv/lib/python3.6/site-packages/torchtext/data/field.py:321: UserWarning: volatile was removed and now has no effect. Use `with torch.no_grad():` instead.
  return Variable(arr, volatile=not train), lengths
/home/ymym/anaconda3/envs/s2sv/lib/python3.6/site-packages/torchtext/data/field.py:322: UserWarning: volatile was removed and now has no effect. Use `with torch.no_grad():` instead.
  return Variable(arr, volatile=not train)
/home/ymym/workspace/OpenNMT-py/onmt/translate/Translator.py:88: UserWarning: volatile was removed and now has no effect. Use `with torch.no_grad():` instead.
  def var(a): return Variable(a, volatile=True)
/home/ymym/workspace/OpenNMT-py/onmt/Models.py:629: UserWarning: volatile was removed and now has no effect. Use `with torch.no_grad():` instead.
  for e in self._all]
/home/ymym/workspace/OpenNMT-py/onmt/modules/GlobalAttention.py:178: UserWarning: Implicit dimension choice for softmax has been deprecated. Change the call to include dim=X as an argument.
  align_vectors = self.sm(align.view(batch*targetL, sourceL))
/home/ymym/workspace/OpenNMT-py/onmt/modules/CopyGenerator.py:91: UserWarning: Implicit dimension choice for softmax has been deprecated. Change the call to include dim=X as an argument.
  prob = F.softmax(logits)
Traceback (most recent call last):
  File "VisServer.py", line 401, in <module>
    main()
  File "VisServer.py", line 371, in main
    reply = model.translate(["This is a test ."], dump_data=True)
  File "VisServer.py", line 322, in translate
    partial=partial, attn_overwrite=attn_overwrite)
  File "/home/ymym/workspace/OpenNMT-py/onmt/translate/Translator.py", line 273, in translate_batch
    batch, resorted[0])
  File "/home/ymym/workspace/OpenNMT-py/onmt/translate/Translator.py", line 362, in _get_top_k
    out = self.model.generator.forward(dec)
TypeError: forward() missing 2 required positional arguments: 'attn' and 'src_map'

Thanks.

[HendrikStrobelt]

hi - thanks for reaching out. We will check what's wrong. Might take a little bit.. Just to double check, you only have the modified ONMT version in your conda environment ?

[ymym3412]

I check my ONMT enviroment. And I remenber little modifying ONMT code!
First time when I tried python VisServer.py command, I got following error.

(s2sv) ymym@ymym:~/workspace/OpenNMT-py$ python VisServer.py 
/home/ymym/anaconda3/envs/s2sv/lib/python3.6/site-packages/h5py/__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.
  from ._conv import register_converters as _register_converters
/home/ymym/anaconda3/envs/s2sv/lib/python3.6/site-packages/torch/nn/modules/rnn.py:38: UserWarning: dropout option adds dropout after all but last recurrent layer, so non-zero dropout expects num_layers greater than 1, but got dropout=0.5 and num_layers=1
  "num_layers={}".format(dropout, num_layers))
Loading model parameters.
partial: []
/home/ymym/anaconda3/envs/s2sv/lib/python3.6/site-packages/torchtext/data/field.py:321: UserWarning: volatile was removed and now has no effect. Use `with torch.no_grad():` instead.
  return Variable(arr, volatile=not train), lengths
/home/ymym/anaconda3/envs/s2sv/lib/python3.6/site-packages/torchtext/data/field.py:322: UserWarning: volatile was removed and now has no effect. Use `with torch.no_grad():` instead.
  return Variable(arr, volatile=not train)
/home/ymym/workspace/OpenNMT-py/onmt/translate/Translator.py:88: UserWarning: volatile was removed and now has no effect. Use `with torch.no_grad():` instead.
  def var(a): return Variable(a, volatile=True)
/home/ymym/workspace/OpenNMT-py/onmt/Models.py:629: UserWarning: volatile was removed and now has no effect. Use `with torch.no_grad():` instead.
  for e in self._all]
/home/ymym/workspace/OpenNMT-py/onmt/modules/GlobalAttention.py:178: UserWarning: Implicit dimension choice for softmax has been deprecated. Change the call to include dim=X as an argument.
  align_vectors = self.sm(align.view(batch*targetL, sourceL))
Traceback (most recent call last):
  File "VisServer.py", line 402, in <module>
    main()
  File "VisServer.py", line 372, in main
    reply = model.translate(["This is a test ."], dump_data=True)
  File "VisServer.py", line 322, in translate
    partial=partial, attn_overwrite=attn_overwrite)
  File "/home/ymym/workspace/OpenNMT-py/onmt/translate/Translator.py", line 167, in translate_batch
    attn_overwrite=new_attn)
  File "/home/ymym/anaconda3/envs/s2sv/lib/python3.6/site-packages/torch/nn/modules/module.py", line 491, in __call__
    result = self.forward(*input, **kwargs)
  File "/home/ymym/workspace/OpenNMT-py/onmt/Models.py", line 261, in forward
    attn_overwrite=attn_overwrite)
  File "/home/ymym/workspace/OpenNMT-py/onmt/Models.py", line 490, in _run_forward_pass
    context.transpose(0, 1))
ValueError: too many values to unpack (expected 2)

So, I fixed onmt/Models.py L489-L490 as

_, copy_attn = self.copy_attn(output, context.transpose(0, 1))

-->

_, copy_attn, _ = self.copy_attn(output, context.transpose(0, 1))

After modifying, I try python VisServer.py, but I got error message described in my first message.

Thanks.

[tangxiangru]

I meet the same err!

 Traceback (most recent call last):
  File "translate.py", line 36, in <module>
    main(opt)

  File "translate.py", line 24, in main
    attn_debug=opt.attn_debug)

  File "/home1/txr/OpenNMT-py-cpu/onmt/translate/translator.py", line 219, in translate
    batch_data = self.translate_batch(batch, data, fast=self.fast)

  File "/home1/txr/OpenNMT-py-cpu/onmt/translate/translator.py", line 320, in translate_batch
    return self._translate_batch(batch, data)

  File "/home1/txr/OpenNMT-py-cpu/onmt/translate/translator.py", line 623, in _translate_batch
    ret["gold_score"] = self._run_target(batch, data)

  File "/home1/txr/OpenNMT-py-cpu/onmt/translate/translator.py", line 669, in _run_target
    out = self.model.generator.forward(dec)

TypeError: forward() missing 2 required positional arguments: 'attn' and 'src_map'

[tangxiangru]

So anyone solved it?

[dattatreya303]

I am facing the same issue. Did someone fix this somehow for seq2seq-vis to work with copy-attn as well?