python - SRL with nlpnet not works with some sentences -
I installed and downloaded but it works only with a few sentences I have Ubuntu 13.04 32bit and Python 2.7 with VirtualAv I use .4.
Here I have posted two examples, first seems to work, second number:
Example 1 (to work): Python nlpnet -tag.py srl / path / to / nlpnet-data / This paragraph is the first sentence. This is the second sentence of the paragraph. This is the first sentence of the paragraph. Sentence A1: Article A0: First V: Sentence is the second sentence of this paragraph. Sentence A1: Article A0: Second V: Sentence Example 2 (does not work):
This sentence took from
Python nlpnet-tag.py srl / path / to / nlpnet-data / $ 1.4 billion robot spacecraft has to face six years of travel to explore Jupiter and its 16 known moons. The 1.4 billion dollar robot spacecraft is a six-year journey in which Jupiter and its 16 known moons can be detected. I also tested with python and tokenization functions but in the second example, calling arg_structures , gives an empty list. Did I miss something? ... if you can provide a solution even when you do not use nlpnet, it will be appreciated.
Thank you
PS: I also want to see if someone helps me fix it. / P>
Elric, author of the package, this:
The problem here is That trained models were trained for Portuguese, it would not work in English.
When you receive a blank arg_structures list, it means that the network does not recognize a token as a prediction.
Python nlpnet -tag.py srl / path / to / nlpnet-data / This paragraph is the first sentence. This is the second sentence of the paragraph. This is the first sentence of the paragraph. Sentence A1: Article A0: First V: Sentence is the second sentence of this paragraph. Sentence A1: Article A0: Second V: Sentence Example 2 (does not work):
This sentence took from
Python nlpnet-tag.py srl / path / to / nlpnet-data / $ 1.4 billion robot spacecraft has to face six years of travel to explore Jupiter and its 16 known moons. The 1.4 billion dollar robot spacecraft is a six-year journey in which Jupiter and its 16 known moons can be detected. I also tested with python and tokenization functions but in the second example, calling arg_structures , gives an empty list. Did I miss something? ... if you can provide a solution even when you do not use nlpnet, it will be appreciated.
Thank you
PS: I also want to see if someone helps me fix it. / P>
Elric, author of the package, this:
The problem here is That trained models were trained for Portuguese, it would not work in English.
When you receive a blank arg_structures list, it means that the network does not recognize a token as a prediction.
Comments
Post a Comment