opus-mt-tc-big-ces_slk-en

Neural machine translation model for translating from Czech and Slovak (ces+slk) to English (en).

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: 2022-03-17
source language(s): ces
target language(s): eng
model: transformer-big
data: opusTCv20210807+bt (source)
tokenization: SentencePiece (spm32k,spm32k)
original model: opusTCv20210807+bt_transformer-big_2022-03-17.zip
more information released models: OPUS-MT ces+slk-eng README

Usage

A short example code:

from transformers import MarianMTModel, MarianTokenizer

src_text = [
    "Podívej se na své kalhoty! Zapni si je na zip.",
    "Mrzí mě, že Tom odchází."
]

model_name = "pytorch-models/opus-mt-tc-big-ces_slk-en"
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:
#     Look at your pants, zip them up.
#     I'm sorry Tom's leaving.

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-ces_slk-en")
print(pipe("Podívej se na své kalhoty! Zapni si je na zip."))

# expected output: Look at your pants, zip them up.

Benchmarks

test set translations: opusTCv20210807+bt_transformer-big_2022-03-17.test.txt
test set scores: opusTCv20210807+bt_transformer-big_2022-03-17.eval.txt
benchmark results: benchmark_results.txt
benchmark output: benchmark_translations.zip

langpair	testset	chr-F	BLEU	#sent	#words
ces-eng	tatoeba-test-v2021-08-07	0.72120	57.7	13824	105010
ces-eng	flores101-devtest	0.66511	41.2	1012	24721
slk-eng	flores101-devtest	0.66084	40.1	1012	24721
ces-eng	multi30k_test_2016_flickr	0.62216	38.6	1000	12955
ces-eng	multi30k_test_2018_flickr	0.61838	37.9	1071	14689
ces-eng	newssyscomb2009	0.56380	29.9	502	11818
ces-eng	news-test2008	0.54071	26.2	2051	49380
ces-eng	newstest2009	0.55871	28.8	2525	65399
ces-eng	newstest2010	0.57634	30.3	2489	61711
ces-eng	newstest2011	0.57002	30.3	3003	74681
ces-eng	newstest2012	0.56564	29.4	3003	72812
ces-eng	newstest2013	0.58723	33.1	3000	64505
ces-eng	newstest2014	0.64192	38.3	3003	68065
ces-eng	newstest2015	0.58688	33.6	2656	53569
ces-eng	newstest2016	0.61544	36.8	2999	64670
ces-eng	newstest2017	0.58085	32.3	3005	61721
ces-eng	newstest2018	0.58627	33.0	2983	63495

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: 3405783
port time: Wed Apr 13 18:42:25 EEST 2022
port machine: LM0-400-22516.local

Helsinki-NLP
/

opus-mt-tc-big-ces_slk-en