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.

Publications: OPUS-MT – Building open translation services for the World and The Tatoeba Translation Challenge – Realistic Data Sets for Low Resource and Multilingual MT (Please, cite if you use this model.)

@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

Release: 2022-03-23
source language(s): dan isl nob nor swe
target language(s): rus ukr
valid target language labels: >>rus<< >>ukr<<
model: transformer-big
data: opusTCv20210807+pbt (source)
tokenization: SentencePiece (spm32k,spm32k)
original model: opusTCv20210807+pbt_transformer-big_2022-03-23.zip
more information released models: OPUS-MT gmq-zle README
more information about the model: MarianMT

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

test set translations: opusTCv20210807+pbt_transformer-big_2022-03-23.test.txt
test set scores: opusTCv20210807+pbt_transformer-big_2022-03-23.eval.txt
benchmark results: benchmark_results.txt
benchmark output: benchmark_translations.zip

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

Helsinki-NLP
/

opus-mt-tc-big-gmq-zle