Xlm Roberta Base Ft Udpos28 Sv

Developed by wietsedv

A multilingual POS tagging model based on XLM-RoBERTa, fine-tuned on Universal Dependencies v2.8, supporting POS tagging tasks in Swedish and multiple other languages.

Sequence Labeling

Transformers

OtherOpen Source License:Apache-2.0 #Multilingual POS Tagging #High Accuracy #Universal Dependencies Parsing

Downloads 72

Release Time : 3/2/2022

Model Overview

This model is a multilingual POS tagging model based on the XLM-RoBERTa architecture, fine-tuned on Universal Dependencies v2.8, with special optimization for Swedish.

Model Features

Multilingual Support

Supports POS tagging tasks for over 90 languages, including major European and some Asian languages

High Accuracy

Achieves 97.9% accuracy on Swedish test sets, with most European languages exceeding 85% accuracy

Based on XLM-RoBERTa

Utilizes the powerful XLM-RoBERTa pre-trained model as a foundation, offering strong cross-lingual transfer capabilities

Model Capabilities

POS Tagging

Multilingual Text Processing

Cross-lingual Transfer Learning

Use Cases

Natural Language Processing

Multilingual Text Analysis

Performs POS tagging on multilingual texts, supporting subsequent syntactic analysis and semantic understanding

Achieves 97.9% accuracy on Swedish

Linguistic Research

Used for comparing grammatical structures and POS distributions across different languages

Supports POS tagging for over 90 languages

language:

sv license: apache-2.0 library_name: transformers tags:
part-of-speech
token-classification datasets:
universal_dependencies metrics:
accuracy

model-index:

name: xlm-roberta-base-ft-udpos28-sv results:
- task: type: token-classification name: Part-of-Speech Tagging dataset: type: universal_dependencies name: Universal Dependencies v2.8 metrics:
  - type: accuracy name: English Test accuracy value: 92.3
  - type: accuracy name: Dutch Test accuracy value: 90.0
  - type: accuracy name: German Test accuracy value: 91.1
  - type: accuracy name: Italian Test accuracy value: 88.0
  - type: accuracy name: French Test accuracy value: 88.2
  - type: accuracy name: Spanish Test accuracy value: 91.1
  - type: accuracy name: Russian Test accuracy value: 91.4
  - type: accuracy name: Swedish Test accuracy value: 97.9
  - type: accuracy name: Norwegian Test accuracy value: 89.7
  - type: accuracy name: Danish Test accuracy value: 92.9
  - type: accuracy name: Low Saxon Test accuracy value: 57.4
  - type: accuracy name: Akkadian Test accuracy value: 40.4
  - type: accuracy name: Armenian Test accuracy value: 87.5
  - type: accuracy name: Welsh Test accuracy value: 69.6
  - type: accuracy name: Old East Slavic Test accuracy value: 76.2
  - type: accuracy name: Albanian Test accuracy value: 80.3
  - type: accuracy name: Slovenian Test accuracy value: 81.0
  - type: accuracy name: Guajajara Test accuracy value: 35.1
  - type: accuracy name: Kurmanji Test accuracy value: 77.3
  - type: accuracy name: Turkish Test accuracy value: 79.2
  - type: accuracy name: Finnish Test accuracy value: 87.0
  - type: accuracy name: Indonesian Test accuracy value: 84.2
  - type: accuracy name: Ukrainian Test accuracy value: 90.4
  - type: accuracy name: Polish Test accuracy value: 88.9
  - type: accuracy name: Portuguese Test accuracy value: 90.1
  - type: accuracy name: Kazakh Test accuracy value: 83.4
  - type: accuracy name: Latin Test accuracy value: 79.1
  - type: accuracy name: Old French Test accuracy value: 62.6
  - type: accuracy name: Buryat Test accuracy value: 63.0
  - type: accuracy name: Kaapor Test accuracy value: 20.8
  - type: accuracy name: Korean Test accuracy value: 64.3
  - type: accuracy name: Estonian Test accuracy value: 89.6
  - type: accuracy name: Croatian Test accuracy value: 90.8
  - type: accuracy name: Gothic Test accuracy value: 26.0
  - type: accuracy name: Swiss German Test accuracy value: 51.8
  - type: accuracy name: Assyrian Test accuracy value: 17.2
  - type: accuracy name: North Sami Test accuracy value: 45.4
  - type: accuracy name: Naija Test accuracy value: 48.1
  - type: accuracy name: Latvian Test accuracy value: 87.1
  - type: accuracy name: Chinese Test accuracy value: 48.5
  - type: accuracy name: Tagalog Test accuracy value: 72.3
  - type: accuracy name: Bambara Test accuracy value: 31.8
  - type: accuracy name: Lithuanian Test accuracy value: 86.2
  - type: accuracy name: Galician Test accuracy value: 88.1
  - type: accuracy name: Vietnamese Test accuracy value: 66.3
  - type: accuracy name: Greek Test accuracy value: 88.1
  - type: accuracy name: Catalan Test accuracy value: 90.1
  - type: accuracy name: Czech Test accuracy value: 90.1
  - type: accuracy name: Erzya Test accuracy value: 50.8
  - type: accuracy name: Bhojpuri Test accuracy value: 51.7
  - type: accuracy name: Thai Test accuracy value: 66.4
  - type: accuracy name: Marathi Test accuracy value: 86.5
  - type: accuracy name: Basque Test accuracy value: 76.4
  - type: accuracy name: Slovak Test accuracy value: 90.5
  - type: accuracy name: Kiche Test accuracy value: 42.4
  - type: accuracy name: Yoruba Test accuracy value: 31.2
  - type: accuracy name: Warlpiri Test accuracy value: 42.5
  - type: accuracy name: Tamil Test accuracy value: 85.3
  - type: accuracy name: Maltese Test accuracy value: 30.6
  - type: accuracy name: Ancient Greek Test accuracy value: 63.0
  - type: accuracy name: Icelandic Test accuracy value: 85.3
  - type: accuracy name: Mbya Guarani Test accuracy value: 32.3
  - type: accuracy name: Urdu Test accuracy value: 67.6
  - type: accuracy name: Romanian Test accuracy value: 85.5
  - type: accuracy name: Persian Test accuracy value: 77.4
  - type: accuracy name: Apurina Test accuracy value: 47.4
  - type: accuracy name: Japanese Test accuracy value: 35.5
  - type: accuracy name: Hungarian Test accuracy value: 87.1
  - type: accuracy name: Hindi Test accuracy value: 75.1
  - type: accuracy name: Classical Chinese Test accuracy value: 30.8
  - type: accuracy name: Komi Permyak Test accuracy value: 52.4
  - type: accuracy name: Faroese Test accuracy value: 80.3
  - type: accuracy name: Sanskrit Test accuracy value: 40.7
  - type: accuracy name: Livvi Test accuracy value: 68.5
  - type: accuracy name: Arabic Test accuracy value: 82.0
  - type: accuracy name: Wolof Test accuracy value: 37.4
  - type: accuracy name: Bulgarian Test accuracy value: 92.9
  - type: accuracy name: Akuntsu Test accuracy value: 41.1
  - type: accuracy name: Makurap Test accuracy value: 22.6
  - type: accuracy name: Kangri Test accuracy value: 47.1
  - type: accuracy name: Breton Test accuracy value: 64.3
  - type: accuracy name: Telugu Test accuracy value: 84.9
  - type: accuracy name: Cantonese Test accuracy value: 48.8
  - type: accuracy name: Old Church Slavonic Test accuracy value: 51.1
  - type: accuracy name: Karelian Test accuracy value: 74.1
  - type: accuracy name: Upper Sorbian Test accuracy value: 77.5
  - type: accuracy name: South Levantine Arabic Test accuracy value: 69.6
  - type: accuracy name: Komi Zyrian Test accuracy value: 44.5
  - type: accuracy name: Irish Test accuracy value: 70.5
  - type: accuracy name: Nayini Test accuracy value: 44.9
  - type: accuracy name: Munduruku Test accuracy value: 24.3
  - type: accuracy name: Manx Test accuracy value: 34.1
  - type: accuracy name: Skolt Sami Test accuracy value: 42.0
  - type: accuracy name: Afrikaans Test accuracy value: 92.1
  - type: accuracy name: Old Turkish Test accuracy value: 40.3
  - type: accuracy name: Tupinamba Test accuracy value: 41.4
  - type: accuracy name: Belarusian Test accuracy value: 89.8
  - type: accuracy name: Serbian Test accuracy value: 91.5
  - type: accuracy name: Moksha Test accuracy value: 46.7
  - type: accuracy name: Western Armenian Test accuracy value: 80.3
  - type: accuracy name: Scottish Gaelic Test accuracy value: 60.4
  - type: accuracy name: Khunsari Test accuracy value: 45.9
  - type: accuracy name: Hebrew Test accuracy value: 87.5
  - type: accuracy name: Uyghur Test accuracy value: 76.9
  - type: accuracy name: Chukchi Test accuracy value: 35.9

XLM-RoBERTa base Universal Dependencies v2.8 POS tagging: Swedish

This model is part of our paper called:

Make the Best of Cross-lingual Transfer: Evidence from POS Tagging with over 100 Languages

Check the Space for more details.

Usage

from transformers import AutoTokenizer, AutoModelForTokenClassification

tokenizer = AutoTokenizer.from_pretrained("wietsedv/xlm-roberta-base-ft-udpos28-sv")
model = AutoModelForTokenClassification.from_pretrained("wietsedv/xlm-roberta-base-ft-udpos28-sv")

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご