metadata

language:
  - pt
  - ru
  - uk
  - zle
tags:
  - translation
  - opus-mt-tc
license: cc-by-4.0
model-index:
  - name: opus-mt-tc-big-zle-pt
    results:
      - task:
          name: Translation rus-por
          type: translation
          args: rus-por
        dataset:
          name: flores101-devtest
          type: flores_101
          args: rus por devtest
        metrics:
          - name: BLEU
            type: bleu
            value: 31.9
      - task:
          name: Translation ukr-por
          type: translation
          args: ukr-por
        dataset:
          name: flores101-devtest
          type: flores_101
          args: ukr por devtest
        metrics:
          - name: BLEU
            type: bleu
            value: 33.6
      - task:
          name: Translation rus-por
          type: translation
          args: rus-por
        dataset:
          name: tatoeba-test-v2021-08-07
          type: tatoeba_mt
          args: rus-por
        metrics:
          - name: BLEU
            type: bleu
            value: 42.8
      - task:
          name: Translation ukr-por
          type: translation
          args: ukr-por
        dataset:
          name: tatoeba-test-v2021-08-07
          type: tatoeba_mt
          args: ukr-por
        metrics:
          - name: BLEU
            type: bleu
            value: 45.2

opus-mt-tc-big-zle-pt

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

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): rus ukr
target language(s): por
model: transformer-big
data: opusTCv20210807 (source)
tokenization: SentencePiece (spm32k,spm32k)
original model: opusTCv20210807_transformer-big_2022-03-23.zip
more information released models: OPUS-MT zle-por README

Usage

A short example code:

from transformers import MarianMTModel, MarianTokenizer

src_text = [
    ">>por<< Я маленькая.",
    ">>por<< Я войду первым."
]

model_name = "pytorch-models/opus-mt-tc-big-zle-pt"
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:
#     Sou pequena.
#     Eu entro primeiro.

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-pt")
print(pipe(">>por<< Я маленькая."))

# expected output: Sou pequena.

Benchmarks

test set translations: opusTCv20210807_transformer-big_2022-03-23.test.txt
test set scores: opusTCv20210807_transformer-big_2022-03-23.eval.txt
benchmark results: benchmark_results.txt
benchmark output: benchmark_translations.zip

langpair	testset	chr-F	BLEU	#sent	#words
rus-por	tatoeba-test-v2021-08-07	0.63749	42.8	10000	74713
ukr-por	tatoeba-test-v2021-08-07	0.65288	45.2	3372	21315
bel-por	flores101-devtest	0.48481	16.2	1012	26519
rus-por	flores101-devtest	0.58567	31.9	1012	26519
ukr-por	flores101-devtest	0.59378	33.6	1012	26519

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: Wed Mar 23 23:45:22 EET 2022
port machine: LM0-400-22516.local