Edit model card

opus-mt-tc-big-zle-zlw

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

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. >>ces<<

Usage

A short example code:

from transformers import MarianMTModel, MarianTokenizer

src_text = [
    ">>pol<< Это метафора.",
    ">>pol<< Что вы делали?"
]

model_name = "pytorch-models/opus-mt-tc-big-zle-zlw"
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:
#     To metafora.
#     Co robiliście?

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-zlw")
print(pipe(">>pol<< Это метафора."))

# expected output: To metafora.

Benchmarks

langpair testset chr-F BLEU #sent #words
bel-pol tatoeba-test-v2021-08-07 0.65517 47.1 287 1706
rus-ces tatoeba-test-v2021-08-07 0.69695 53.4 2934 16831
rus-pol tatoeba-test-v2021-08-07 0.72176 53.7 3543 21505
ukr-ces tatoeba-test-v2021-08-07 0.73149 58.0 1787 8550
ukr-pol tatoeba-test-v2021-08-07 0.74649 57.0 2519 13201
bel-ces flores101-devtest 0.41248 11.1 1012 22101
bel-pol flores101-devtest 0.42240 10.2 1012 22520
rus-ces flores101-devtest 0.50971 23.1 1012 22101
rus-pol flores101-devtest 0.48672 18.4 1012 22520
ukr-ces flores101-devtest 0.52482 25.1 1012 22101
ukr-pol flores101-devtest 0.48790 18.8 1012 22520
rus-ces newstest2012 0.45834 18.8 3003 65456
rus-ces newstest2013 0.52364 26.0 3000 57250

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

  • transformers version: 4.16.2
  • OPUS-MT git hash: 1bdabf7
  • port time: Thu Mar 24 00:50:29 EET 2022
  • port machine: LM0-400-22516.local
Downloads last month
5
Hosted inference API
This model can be loaded on the Inference API on-demand.

Evaluation results