xlm-roberta-base-ft-udpos28-lt Open Source Model - Supports Accurate Part-of-Speech Tagging for Lithuanian

Home

Xlm Roberta Base Ft Udpos28 Lt

Developed by wietsedv

Multilingual POS tagging model based on XLM-RoBERTa, specifically fine-tuned for Lithuanian

Sequence Labeling

Transformers

OtherOpen Source License:Apache-2.0 #Multilingual POS tagging #Lithuanian NLP #High-precision grammatical analysis

Downloads 23

Release Time : 3/2/2022

Model Overview

This model is a POS tagging model fine-tuned on Lithuanian data from Universal Dependencies v2.8 based on the XLM-RoBERTa base model, capable of accurately identifying POS tags in Lithuanian text.

Model Features

Multilingual support

Based on XLM-RoBERTa architecture, capable of handling multiple languages

High accuracy

Achieves 96.1% accuracy on Lithuanian test sets

Universal Dependencies standard

Follows the POS tagging standard of Universal Dependencies v2.8

Model Capabilities

POS tagging

Multilingual text processing

Natural language processing

Use Cases

Natural language processing

Text analysis

Perform POS tagging analysis on Lithuanian text

Accurately identifies nouns, verbs, adjectives, and other POS tags

Linguistic research

Used for Lithuanian grammar research and analysis

Provides accurate POS tagging data

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

This model addresses the part - of - speech tagging task. It is based on the XLM - RoBERTa base architecture and trained on Universal Dependencies v2.8 data for Lithuanian. It is part of a research effort to optimize cross - lingual transfer in POS tagging across over 100 languages.

🚀 Quick Start

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 Examples

Basic Usage

from transformers import AutoTokenizer, AutoModelForTokenClassification

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

📄 License

The model is released under the Apache - 2.0 license.

📚 Documentation

Model Information

Property	Details
Model Type	XLM - RoBERTa base fine - tuned for Universal Dependencies v2.8 POS tagging (Lithuanian)
Training Data	Universal Dependencies v2.8

Results

The model has been evaluated on multiple languages with the accuracy metric. Here are the results:

Language	Test accuracy
English	80.7
Dutch	80.6
German	76.0
Italian	77.8
French	75.5
Spanish	79.6
Russian	88.9
Swedish	81.6
Norwegian	76.3
Danish	78.9
Low Saxon	52.0
Akkadian	31.6
Armenian	84.1
Welsh	63.8
Old East Slavic	75.6
Albanian	76.8
Slovenian	81.4
Guajajara	26.7
Kurmanji	77.1
Turkish	74.9
Finnish	83.2
Indonesian	78.0
Ukrainian	88.1
Polish	86.3
Portuguese	81.6
Kazakh	83.1
Latin	78.7
Old French	56.1
Buryat	64.3
Kaapor	22.5
Korean	64.6
Estonian	81.5
Croatian	86.6
Gothic	22.6
Swiss German	48.1
Assyrian	14.6
North Sami	39.8
Naija	41.4
Latvian	89.0
Chinese	34.4
Tagalog	73.0
Bambara	26.4
Lithuanian	96.1
Galician	81.1
Vietnamese	65.3
Greek	81.8
Catalan	76.2
Czech	86.5
Erzya	48.7
Bhojpuri	50.9
Thai	54.5
Marathi	82.8
Basque	75.6
Slovak	88.5
Kiche	33.5
Yoruba	24.6
Warlpiri	44.1
Tamil	79.1
Maltese	25.5
Ancient Greek	65.8
Icelandic	80.7
Mbya Guarani	32.2
Urdu	59.1
Romanian	78.6
Persian	72.8
Apurina	42.0
Japanese	22.9
Hungarian	76.9
Hindi	62.2
Classical Chinese	15.8
Komi Permyak	48.3
Faroese	77.3
Sanskrit	41.0
Livvi	67.2
Arabic	73.9
Wolof	28.0
Bulgarian	85.9
Akuntsu	26.0
Makurap	17.8
Kangri	50.6
Breton	60.3
Telugu	85.0
Cantonese	39.1
Old Church Slavonic	51.6
Karelian	71.3
Upper Sorbian	75.7
South Levantine Arabic	67.0
Komi Zyrian	43.0
Irish	60.1
Nayini	46.2
Munduruku	18.8
Manx	33.3
Skolt Sami	37.3
Afrikaans	76.4
Old Turkish	37.1
Tupinamba	34.1
Belarusian	89.1
Serbian	87.7
Moksha	46.3
Western Armenian	75.4
Scottish Gaelic	56.2
Khunsari	39.2
Hebrew	83.3
Uyghur	76.6
Chukchi	35.4

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