translation opus-mt-tc

opus-mt-tc-big-zle-zle

Neural machine translation model for translating from East Slavic languages (zle) to East Slavic languages (zle).

This model is part of the OPUS-MT project, an effort to make neural machine translation models widely available and accessible for many languages in the world. All models are originally trained using the amazing framework of Marian NMT, an efficient NMT implementation written in pure C++. The models have been converted to pyTorch using the transformers library by huggingface. Training data is taken from OPUS and training pipelines use the procedures of OPUS-MT-train.

@inproceedings{tiedemann-thottingal-2020-opus,
    title = "{OPUS}-{MT} {--} Building open translation services for the World",
    author = {Tiedemann, J{\"o}rg  and Thottingal, Santhosh},
    booktitle = "Proceedings of the 22nd Annual Conference of the European Association for Machine Translation",
    month = nov,
    year = "2020",
    address = "Lisboa, Portugal",
    publisher = "European Association for Machine Translation",
    url = "https://aclanthology.org/2020.eamt-1.61",
    pages = "479--480",
}

@inproceedings{tiedemann-2020-tatoeba,
    title = "The Tatoeba Translation Challenge {--} Realistic Data Sets for Low Resource and Multilingual {MT}",
    author = {Tiedemann, J{\"o}rg},
    booktitle = "Proceedings of the Fifth Conference on Machine Translation",
    month = nov,
    year = "2020",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2020.wmt-1.139",
    pages = "1174--1182",
}

Model info

This is a multilingual translation model with multiple target languages. A sentence initial language token is required in the form of >>id<< (id = valid target language ID), e.g. >>bel<<

Usage

A short example code:

from transformers import MarianMTModel, MarianTokenizer

src_text = [
    ">>ukr<< Кот мёртвый.",
    ">>bel<< Джон живе в Нью-Йорку."
]

model_name = "pytorch-models/opus-mt-tc-big-zle-zle"
tokenizer = MarianTokenizer.from_pretrained(model_name)
model = MarianMTModel.from_pretrained(model_name)
translated = model.generate(**tokenizer(src_text, return_tensors="pt", padding=True))

for t in translated:
    print( tokenizer.decode(t, skip_special_tokens=True) )

# expected output:
#     Кіт мертвий.
#     Джон жыве ў Нью-Йорку.

You can also use OPUS-MT models with the transformers pipelines, for example:

from transformers import pipeline
pipe = pipeline("translation", model="Helsinki-NLP/opus-mt-tc-big-zle-zle")
print(pipe(">>ukr<< Кот мёртвый."))

# expected output: Кіт мертвий.

Benchmarks

langpair testset chr-F BLEU #sent #words
bel-rus tatoeba-test-v2021-08-07 0.82526 68.6 2500 18895
bel-ukr tatoeba-test-v2021-08-07 0.81036 65.5 2355 15179
rus-bel tatoeba-test-v2021-08-07 0.66943 50.3 2500 18756
rus-ukr tatoeba-test-v2021-08-07 0.83639 70.1 10000 60212
ukr-bel tatoeba-test-v2021-08-07 0.75368 58.9 2355 15175
ukr-rus tatoeba-test-v2021-08-07 0.86806 75.7 10000 60387
bel-rus flores101-devtest 0.47960 14.5 1012 23295
bel-ukr flores101-devtest 0.47335 12.8 1012 22810
rus-ukr flores101-devtest 0.55287 25.5 1012 22810
ukr-rus flores101-devtest 0.56224 28.3 1012 23295

Acknowledgements

The work is supported by the European Language Grid as pilot project 2866, by the FoTran project, funded by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 771113), and the MeMAD project, funded by the European Union’s Horizon 2020 Research and Innovation Programme under grant agreement No 780069. We are also grateful for the generous computational resources and IT infrastructure provided by CSC -- IT Center for Science, Finland.

Model conversion info