Character-level part-of-speech tagger of Slovene language

Character-level part-of-speech tagger of Slovene language

CLARIN.SI data & tools

Avtorji: Belej, Primož ; Robnik-Šikonja, Marko and Krek, Simon

Identifikator vnosa: http://hdl.handle.net/11356/1211

URL projekta: https://github.com/PrimozBelej/SloTagger

Dokumentirano v: https://repozitorij.uni-lj.si/IzpisGradiva.php?id=105266&lang=eng

Datum objave: 2019-03-02

Vrsta: toolService

Jezik(i): Slovenian

Opis: Part-of-speech tagger for Slovene language implemented using convolutional and LSTM neural networks. Tagger uses character-level representation of sentences. The tagger has been trained on the ssj500k 2.1 corpus, http://hdl.handle.net/11356/1181.

Izdajatelj: Faculty of Computer and Information Science, University of Ljubljana

Centre for Language Resources and Technologies, University of Ljubljana

Ključne besede: part-of-speech tagging neural networks

Zbirke: CLARIN.SI data & tools

Prikaži polni zapis vnosa

Datoteke v tem vnosu

Prenesi vse datoteke v vnosu (140.75 MB)

To je vnos

Publicly Available

z licenco:
GNU General Public Licence, version 3

Ime: README.md
Velikost: 2.89 KB
Format: Neznano
Opis: Markdown document containing brief description and examples of usage.
MD5: 6ce03d3464c49b5d7069a3b8fcee8661

Prenesi datoteko

Ime: characterlist
Velikost: 327 bajtov
Format: Neznano
Opis: List of all character that model is aware of. This list is generated during the training process and must be present in the working directory while using the tagger.
MD5: c31ca9602090336659e201b9acd71b71

Prenesi datoteko

Ime: config.py
Velikost: 38 bajtov
Format: Neznano
Opis: Configuration file.
MD5: 22ea18a7b4467079ae1dc4e010d1ee6d

Prenesi datoteko

Ime: en_sl_tag
Velikost: 27.65 KB
Format: Neznano
Opis: Pairs of English and Slovene tags from Multext east specification for Slovene language.
MD5: 13e8cc353a9e5a5b4474b49263dd64b7

Prenesi datoteko

Ime: model.json
Velikost: 14.83 KB
Format: Neznano
Opis: Keras model configuration. This file is generated during the training process.
MD5: 7058c188044db19fd1900e4cd667345c

Prenesi datoteko

Ime: neuralmodel.py
Velikost: 4.69 KB
Format: Neznano
Opis: Python library containing implementations of neural networks used in language model.
MD5: 44e913dc9f264609215926bcbed31a5a

Prenesi datoteko

Ime: pos_embeddings
Velikost: 236.46 KB
Format: Neznano
Opis: Pairs of Multext East tags and their vector space embeddings.
MD5: 7b90f745984463f1df090140a8174af2

Prenesi datoteko

Ime: poslib.py
Velikost: 11.49 KB
Format: Neznano
Opis: Python library containing implementations of POS tag transformations.
MD5: 0e8bd56dd114eb825b631035da759118

Prenesi datoteko

Ime: tag.py
Velikost: 8.47 KB
Format: Neznano
Opis: Python module with command line interface used for tagging text in plain text files and XML/TEI files. Basic examples are available in README.md.
MD5: f042274b8804e7811bcbce043b29940b

Prenesi datoteko

Ime: teiutils.py
Velikost: 2.17 KB
Format: Neznano
Opis: Python library implementing helper functions for working with XML files.
MD5: 3c86ae3c5b02f16bceba7aacab980a79

Prenesi datoteko

Ime: test.py
Velikost: 1.47 KB
Format: Neznano
Opis: Python module for testing tagging accuracy given predicted sentences and validation sentences. Module has command line interface.
MD5: 008b9bddfef8529392d122dd92a046b9

Prenesi datoteko

Ime: train.py
Velikost: 2.99 KB
Format: Neznano
Opis: Python module with command line interface for training new language models. Basic example of use is in README.md.
MD5: 75d3b843f37c26cfed11b3a3303eb4dc

Prenesi datoteko

Ime: txtutils.py
Velikost: 761 bajtov
Format: Neznano
Opis: Python library containing helper functions for working with text files.
MD5: 6cf6adb0961947a10145cda0373ab832

Prenesi datoteko

Ime: model_weights.h5
Velikost: 140.44 MB
Format: Neznano
Opis: Keras model trained on ssj500k.
MD5: 05337dd1a9ccfad2e90d09a3af6b9acc

Prenesi datoteko