UDPipe 2 is a Python prototype, capable of performing tagging, lemmatization and syntactic analysis of CoNLL-U input. It took part in several competitions, reaching excellent results in all of them:

• one of the winners of the CoNLL 2018 Shared Task
• one of the winners of the SIGMORPHON 2019 Shared Task 2
• the winner of EvaLatin 2020

Compared to UDPipe 1, it is Python-only, it does not perform tokenization, and the models require more computation power.

REST Service

The UDPipe 2 models are currently available from the LINDAT UDPipe REST Service. Apart from the web interface, you can use the udpipe2_client.py script to process your files through the service.

Getting UDPipe 2

You can get UDPipe 2 sources from the udpipe-2 branch of the UDPipe repository. The sources can be used to both train a new model and to run a local REST server for inference.

Models

The available models are described on a separate page.

Acknowledgements

This work has been supported by the Ministry of Education, Youth and Sports of the Czech Republic, Project No. LM2018101 LINDAT/CLARIAH-CZ.

Publications

• Milan Straka (2018): UDPipe 2.0 Prototype at CoNLL 2018 UD Shared Task. In: Proceedings of CoNLL 2018: The SIGNLL Conference on Computational Natural Language Learning, pp. 197-207, Association for Computational Linguistics, Stroudsburg, PA, USA, ISBN 978-1-948087-72-8
• Milan Straka, Jana Straková, Jan Hajič (2019): UDPipe at SIGMORPHON 2019: Contextualized Embeddings, Regularization with Morphological Categories, Corpora Merging. In: Proceedings of the 16th SIGMORPHON Workshop on Computational Research in Phonetics, Phonology, and Morphology, pp. 95-103, Association for Computational Linguistics, Stroudsburg, PA, USA, ISBN 978-1-950737-36-9
• Milan Straka, Jana Straková, Jan Hajič (2019): Evaluating Contextualized Embeddings on 54 Languages in POS Tagging, Lemmatization and Dependency Parsing. In: ArXiv.org Computing Research Repository, ISSN 2331-8422, 1904.02099

Bibtex for Referencing

@InProceedings{straka-2018-udpipe,
    title = "{UDP}ipe 2.0 Prototype at {C}o{NLL} 2018 {UD} Shared Task",
    author = "Straka, Milan",
    booktitle = "Proceedings of the {C}o{NLL} 2018 Shared Task: Multilingual Parsing from Raw Text to Universal Dependencies",
    month = oct,
    year = "2018",
    address = "Brussels, Belgium",
    publisher = "Association for Computational Linguistics",
    url = "https://www.aclweb.org/anthology/K18-2020",
    doi = "10.18653/v1/K18-2020",
    pages = "197--207",
}