chantera / biaffineparser Goto Github PK

Can you tell more about the format in which input has to be given? It has not been mentioned in the Readme and I am having an issue using this repo.

Hi , When I run the test , the following error occurs：
file: <class 'str'>
model_file: <class 'chainer_model.DeepBiaffine'>
2019-01-29 11:03:57.087 CST bfc81b [trace] Orthogonal pretrainer loss: 4.45e-29
2019-01-29 11:03:58.607 CST bfc81b [trace] Orthogonal pretrainer loss: 4.24e-29
2019-01-29 11:03:58.623 CST bfc81b [error] Exception occurred during execution:
Traceback (most recent call last):
File "/home/lxiao/workspace/BiaffineParser/src/teras/teras/app/init.py", line 58, in run
self._process()
File "/home/lxiao/workspace/BiaffineParser/src/teras/teras/app/init.py", line 102, in _process
self._command(**kwargs)
File "parser.py", line 221, in test
_load_test_model(model, model_file, device_id=gpu)
File "parser.py", line 180, in _load_test_model
chainer.serializers.load_npz(file, model)
File "/home/lxiao/anaconda3/lib/python3.6/site-packages/chainer/serializers/npz.py", line 140, in load_npz
d.load(obj)
File "/home/lxiao/anaconda3/lib/python3.6/site-packages/chainer/serializer.py", line 83, in load
obj.serialize(self)
File "/home/lxiao/anaconda3/lib/python3.6/site-packages/chainer/link.py", line 795, in serialize
d[name].serialize(serializer[name])
File "/home/lxiao/anaconda3/lib/python3.6/site-packages/chainer/link.py", line 551, in serialize
data = serializer(name, param.data)
File "/home/lxiao/anaconda3/lib/python3.6/site-packages/chainer/serializers/npz.py", line 113, in call
numpy.copyto(value, dataset)
ValueError: could not broadcast input array from shape (63) into shape (64)
2019-01-29 11:03:58.624 CST bfc81b [info] LOG End with ACCESSID=[bfc81b] UNIQUEID=[UNIQID] ACCESSTIME=[2019 -01-29 11:03:16.919340 CST] PROCESSTIME=[41.704970000]

PS: my test data as follows:

1	只有	只有	_	_	_	7	obj	_	_
2	是	是	_	_	_	2	obj	_	_
3	签署	签署	_	_	_	7	obj	_	_
4	协定	协定	_	_	_	3	obj	_	_
5	两国	两国	_	_	_	4	obj	_	_
6	的	的	_	_	_	3	obj	_	_
7	居民	居民	_	_	_	10	obj	_	_
8	才	才	_	_	_	10	obj	_	_
9	会	会	_	_	_	10	obj	_	_
10	使用	使用	_	_	_	0	root	_	_
11	该	该	_	_	_	12	obj	_	_
12	协定	协定	_	_	_	14	obj	_	_
13	的	的	_	_	_	12	obj	_	_
14	条款	条款	_	_	_	10	obj	_	_
15	。	。	_	_	_	10	obj	_	_

2019-02-18 20:48:24.230 CST e2989e [info] saving the context to model/20190218-e2989e.context ...
0%| | 0/3997 [00:00<?, ?it/s]2019-02-18 20:48:24.748 CST e2989e [error] Exception occurred during execution:
Traceback (most recent call last):
File "/home/liuxiao/workspace/dependency-parser/biaffineparser/src/teras/teras/app/app.py", line 60, in run
self._process()
File "/home/liuxiao/workspace/dependency-parser/biaffineparser/src/teras/teras/app/app.py", line 104, in _process
self._command(**kwargs)
File "src/main.py", line 73, in train
trainer.fit(train_dataset, test_dataset, n_epoch, batch_size)
File "/home/liuxiao/workspace/dependency-parser/biaffineparser/src/teras/teras/training/trainer.py", line 107, in fit
main_loop()
File "/home/liuxiao/workspace/dependency-parser/biaffineparser/src/teras/teras/training/trainer.py", line 98, in main_loop
self._process(forward, train_dataset, lossfun, convert, batch_size, epoch_logs)
File "/home/liuxiao/workspace/dependency-parser/biaffineparser/src/teras/teras/training/trainer.py", line 145, in _process
ys = forward(*xs)
File "/home/liuxiao/miniconda3/lib/python3.6/site-packages/chainer/link.py", line 242, in call
out = forward(*args, **kwargs)
File "/home/liuxiao/workspace/dependency-parser/biaffineparser/src/models.py", line 122, in forward
self._hs = self.encoder(words, pretrained_words, postags)
File "/home/liuxiao/miniconda3/lib/python3.6/site-packages/chainer/link.py", line 242, in call
out = forward(*args, **kwargs)
File "/home/liuxiao/workspace/dependency-parser/biaffineparser/src/models.py", line 264, in forward
rs = self._concat_embeds(self.embeds(*xs))
File "/home/liuxiao/workspace/dependency-parser/biaffineparser/src/models.py", line 274, in _concat_embeds
rs_words = [rs_word_seq + rs_pre_seq for rs_word_seq, rs_pre_seq in zip(rs_words, rs_words_pretrained)]
File "/home/liuxiao/workspace/dependency-parser/biaffineparser/src/models.py", line 274, in
rs_words = [rs_word_seq + rs_pre_seq for rs_word_seq, rs_pre_seq in zip(rs_words, rs_words_pretrained)]
File "/home/liuxiao/miniconda3/lib/python3.6/site-packages/chainer/functions/math/basic_math.py", line 229, in add
return Add().apply((lhs, rhs))[0]
File "/home/liuxiao/miniconda3/lib/python3.6/site-packages/chainer/function_node.py", line 263, in apply
outputs = self.forward(in_data)
File "/home/liuxiao/miniconda3/lib/python3.6/site-packages/chainer/functions/math/basic_math.py", line 156, in forward
y = utils.force_array(x[0] + x[1])
ValueError: operands could not be broadcast together with shapes (37,100) (37,300)
When using a pre-trained word vector file with a dimension of 300, the following error is exposed. Also modified to word_embed_size=300 in dataset.py
Would you share the word vector file you used in your code? Thank you so much!

作者你好，
请问，你设计的解析器每秒大概能解析多少个句子。

I met a problem when I ran this model implemented by pytorch. You said the pytorch model weight initialization and GPU computation have not been completed yet. Will you update this code later?

Chinese & English. can you reproduce the results of the corresponding paper? thanks very much