Opus-mt-uk-bg Open-source Machine Translation Model - Free and Precise Translation from Ukrainian to Bulgarian

Home

Opus Mt Uk Bg

Developed by Helsinki-NLP

This is a Transformer-based machine translation model from Ukrainian to Bulgarian, developed by the Helsinki-NLP team.

Machine Translation

Transformers

OtherOpen Source License:Apache-2.0 #Slavic language translation #Transformer architecture #Standardized preprocessing

Downloads 49

Release Time : 3/2/2022

Model Overview

This model is specifically designed for translating Ukrainian text into Bulgarian, employing standardized preprocessing and SentencePiece tokenization techniques.

Model Features

High-quality translation

Achieved 55.7 BLEU score and 0.734 chr-F score on the Tatoeba test set

Standardized preprocessing

Utilizes standardized processing + SentencePiece tokenization (spm32k) preprocessing pipeline

Dedicated language pair support

Specialized in Ukrainian to Bulgarian translation tasks

Model Capabilities

Text translation

Cross-language conversion

Use Cases

Language services

Document translation

Translate Ukrainian documents into Bulgarian

Achieved 55.7 BLEU score on the Tatoeba test set

Cross-language communication

Facilitate communication between Ukrainian and Bulgarian speakers

🚀 ukr-bul

This project focuses on the translation from Ukrainian to Bulgarian. It provides a specific model and related evaluation metrics for this language pair.

🚀 Quick Start

This project is mainly used for Ukrainian to Bulgarian translation. You can download the original weights and test set translations through the following links:

Download original weights: opus-2020-06-17.zip
Test set translations: opus-2020-06-17.test.txt
Test set scores: opus-2020-06-17.eval.txt

✨ Features

Translation Direction: It supports translation from Ukrainian to Bulgarian.
Model Type: The transformer-align model is used.
Pre - processing: Normalization and SentencePiece (spm32k, spm32k) are applied.

📚 Documentation

Language Pair Details

Source Group: Ukrainian
Target Group: Bulgarian

Model Information

Property	Details
Model Type	transformer-align
Pre - processing	normalization + SentencePiece (spm32k,spm32k)

Benchmarks

testset	BLEU	chr-F
Tatoeba-test.ukr.bul	55.7	0.734

System Info

hf_name: ukr-bul
source_languages: ukr
target_languages: bul
opus_readme_url: https://github.com/Helsinki-NLP/Tatoeba-Challenge/tree/master/models/ukr-bul/README.md
original_repo: Tatoeba-Challenge
tags: ['translation']
languages: ['uk', 'bg']
src_constituents: {'ukr'}
tgt_constituents: {'bul', 'bul_Latn'}
src_multilingual: False
tgt_multilingual: False
prepro: normalization + SentencePiece (spm32k,spm32k)
url_model: https://object.pouta.csc.fi/Tatoeba-MT-models/ukr-bul/opus-2020-06-17.zip
url_test_set: https://object.pouta.csc.fi/Tatoeba-MT-models/ukr-bul/opus-2020-06-17.test.txt
src_alpha3: ukr
tgt_alpha3: bul
short_pair: uk-bg
chrF2_score: 0.7340000000000001
bleu: 55.7
brevity_penalty: 0.976
ref_len: 5181.0
src_name: Ukrainian
tgt_name: Bulgarian
train_date: 2020-06-17
src_alpha2: uk
tgt_alpha2: bg
prefer_old: False
long_pair: ukr-bul
helsinki_git_sha: 480fcbe0ee1bf4774bcbe6226ad9f58e63f6c535
transformers_git_sha: 2207e5d8cb224e954a7cba69fa4ac2309e9ff30b
port_machine: brutasse
port_time: 2020-08-21-14:41

📄 License

This project is licensed under the Apache - 2.0 license.

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