pretrained-multilingual-parsers

A set of pretrained multilingual (Malt)parsers

They were trained using MaltParser 1.7, MaltOptimizer (a customized version) and Universal Dependency Treebanks v2.0 (std version)

Models

Models using CPOSTAG (universal tags) information

bilingual models

monolingual models (used as baseline for comparison)

Models using POSTAG (specific tags) information

bilingual models

monolingual models

Note: To execute these models, in the target conll first copy the column POSTAG into the column CPOSTAG first

Execution

java -jar -Xmx2000m maltparser-1.7.2/maltparser-1.7.2.jar - MODEL_MCO -i PATH_INPUT_CONLL -o PATH_OUTPUT_CONLL -m parse

Example:

java -jar -Xmx2000m maltparser-1.7.2/maltparser-1.7.2.jar -c es_fr-liblinear_google_universal_maltoptimizer -i es-universal-test.conll -o es-universal-test-output.conll -m parse

References

Vilares, D., Gómez-Rodríguez, C. & Alonso, M. A. (2016, August). One model, two languages: training bilingual parsers with harmonized treebanks. In The 54th Annual Meeting of the Association for Computational Linguistics (p. 425-431).