Parsing/home

From Nordic Language Processing Laboratory
Revision as of 09:57, 29 November 2017 by Aarons (talk | contribs)
Jump to: navigation, search

Background

An experimentation environment for data-driven dependency parsing is maintained for NLPL under the coordination of Uppsala University (UU). Initially, the software and data are commissioned on the Norwegian Abel supercluster.

Using the Uppsala Parser

  • Log into Abel
  • Activate the NLPL module repository:
module use -a /projects/nlpl/software/modulefiles/
  • Load the most recent version of the uuparser module:
module load uuparser

Train a parsing model

To train a set of parsing models on treebanks from Universal Dependencies (v2.1):

uuparser --include [languages to include denoted by their ISO id] --outdir my-output-directory

for example

uuparser --include "sv en ru" --outdir ~/experiments

will train separate models for UD Swedish, English and Russian and store the results in the experiments folder in your home directory. Model selection is included in the training process by default; that is, at each epoch the current model is evaluated on the UD dev data, and at the end of training the best performing model for each language is selected.

Predicting with a pre-trained parsing model

To predict on UD test data with the models trained above:

uuparser --include "sv en ru" --outdir ~/experiments --predict

Contact: Aaron Smith, Uppsala University, firstname.lastname@lingfil.uu.se