Edit model card
YAML Metadata Error: "language[4]" must be a string

opus-mt-tc-big-gmq-zle

Neural machine translation model for translating from North Germanic languages (gmq) 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. >>rus<<

Usage

A short example code:

from transformers import MarianMTModel, MarianTokenizer

src_text = [
    ">>bel<< Det er allerede torsdag i morgen.",
    ">>ukr<< Tom lekte katt och råtta med Mary."
]

model_name = "pytorch-models/opus-mt-tc-big-gmq-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:
#     Гэта ўжо чацвер заўтра.
#     Том грав кішку і щура з Марією.

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-gmq-zle")
print(pipe(">>bel<< Det er allerede torsdag i morgen."))

# expected output: Гэта ўжо чацвер заўтра.

Benchmarks

langpair testset chr-F BLEU #sent #words
dan-rus tatoeba-test-v2021-08-07 0.72627 53.9 1713 10480
nob-rus tatoeba-test-v2021-08-07 0.66881 45.8 1277 10659
swe-rus tatoeba-test-v2021-08-07 0.66248 45.9 1282 7659
dan-rus flores101-devtest 0.53271 25.6 1012 23295
dan-ukr flores101-devtest 0.54273 25.5 1012 22810
nob-rus flores101-devtest 0.50426 22.1 1012 23295
nob-ukr flores101-devtest 0.51156 21.6 1012 22810
swe-rus flores101-devtest 0.53226 25.8 1012 23295
swe-ukr flores101-devtest 0.54257 25.7 1012 22810

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 02:08:53 EET 2022
  • port machine: LM0-400-22516.local
Downloads last month
21
Hosted inference API
This model can be loaded on the Inference API on-demand.

Evaluation results