@@ -10,22 +10,14 @@ and extended to support DM, PSD, EDS and AMR.
1010
1111### Install
1212
13- Create a Python virtual environment. For example, on Linux:
14-
15- virtualenv --python=/usr/bin/python3 venv
16- . venv/bin/activate # on bash
17- source venv/bin/activate.csh # on csh
13+ Install the latest code from GitHub:
1814
19- Install the latest code from GitHub (may be unstable):
20-
21- git clone https://github.com/danielhers/tupa --branch=mrp
22- cd tupa
23- pip install .
15+ pip install git+https://github.com/danielhers/tupa.git@mrp
2416
2517### Train the parser
2618
2719Having a directory with MRP graph files
28- (for example, [ the MRP sample ] ( http://svn.nlpl.eu/mrp/2019/public/sample .tgz ) ),
20+ (for example, [ the MRP UCCA data ] ( http://svn.nlpl.eu/mrp/2019/public/ucca .tgz ) ),
2921run:
3022
3123 python -m tupa -t <train_dir> -d <dev_dir> -m <model_filename>
@@ -49,15 +41,16 @@ Contributors
4941
5042Citation
5143--------
52- If you make use of this software, please cite [ the following paper] ( http://aclweb.org/anthology/P18-1035 ) :
53-
54- @InProceedings{hershcovich2018multitask,
55- author = {Hershcovich, Daniel and Abend, Omri and Rappoport, Ari},
56- title = {Multitask Parsing Across Semantic Representations},
57- booktitle = {Proc. of ACL},
58- year = {2018},
59- pages = {373--385},
60- url = {http://aclweb.org/anthology/P18-1035}
44+ If you make use of this software, please cite [ the following paper] ( https://www.aclweb.org/anthology/K19-2002 ) :
45+
46+ @InProceedings{hershcovich-arviv-2019-tupa,
47+ author = {Hershcovich, Daniel and Arviv, Ofir},
48+ title = {{TUPA} at {MRP} 2019: A Multi-Task Baseline Syste},
49+ booktitle = {Proc. of CoNLL MRP Shared Task},
50+ year = {2019},
51+ pages = {28--39},
52+ url = {https://www.aclweb.org/anthology/K19-2002},
53+ doi = {10.18653/v1/K19-2002}
6154 }
6255
6356
0 commit comments