metadata

language:
  - br
  - cel
  - cy
  - en
  - ga
tags:
  - translation
  - opus-mt-tc
license: cc-by-4.0
model-index:
  - name: opus-mt-tc-big-cel-en
    results:
      - task:
          name: Translation cym-eng
          type: translation
          args: cym-eng
        dataset:
          name: flores101-devtest
          type: flores_101
          args: cym eng devtest
        metrics:
          - name: BLEU
            type: bleu
            value: 50.2
      - task:
          name: Translation gle-eng
          type: translation
          args: gle-eng
        dataset:
          name: flores101-devtest
          type: flores_101
          args: gle eng devtest
        metrics:
          - name: BLEU
            type: bleu
            value: 37.4
      - task:
          name: Translation bre-eng
          type: translation
          args: bre-eng
        dataset:
          name: tatoeba-test-v2021-08-07
          type: tatoeba_mt
          args: bre-eng
        metrics:
          - name: BLEU
            type: bleu
            value: 36.1
      - task:
          name: Translation cym-eng
          type: translation
          args: cym-eng
        dataset:
          name: tatoeba-test-v2021-08-07
          type: tatoeba_mt
          args: cym-eng
        metrics:
          - name: BLEU
            type: bleu
            value: 53.6
      - task:
          name: Translation gle-eng
          type: translation
          args: gle-eng
        dataset:
          name: tatoeba-test-v2021-08-07
          type: tatoeba_mt
          args: gle-eng
        metrics:
          - name: BLEU
            type: bleu
            value: 57.7

opus-mt-tc-big-cel-en

Neural machine translation model for translating from Celtic languages (cel) 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: 2022-03-13
source language(s): bre cym gle
target language(s): eng
model: transformer-big
data: opusTCv20210807+bt (source)
tokenization: SentencePiece (spm32k,spm32k)
original model: opusTCv20210807+bt_transformer-big_2022-03-13.zip
more information released models: OPUS-MT cel-eng README

Usage

A short example code:

from transformers import MarianMTModel, MarianTokenizer

src_text = [
    "A-du emaoc’h?",
    "Ta'n ushtey glen."
]

model_name = "pytorch-models/opus-mt-tc-big-cel-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:
#     Is that you?
#     Ta'n ushtey glen.

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-cel-en")
print(pipe("A-du emaoc’h?"))

# expected output: Is that you?

Benchmarks

test set translations: opusTCv20210807+bt_transformer-big_2022-03-13.test.txt
test set scores: opusTCv20210807+bt_transformer-big_2022-03-13.eval.txt
benchmark results: benchmark_results.txt
benchmark output: benchmark_translations.zip

langpair	testset	chr-F	BLEU	#sent	#words
bre-eng	tatoeba-test-v2021-08-07	0.53712	36.1	383	2065
cym-eng	tatoeba-test-v2021-08-07	0.69239	53.6	818	5563
gle-eng	tatoeba-test-v2021-08-07	0.72087	57.7	1913	11190
cym-eng	flores101-devtest	0.71379	50.2	1012	24721
gle-eng	flores101-devtest	0.63946	37.4	1012	24721

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 18:36:25 EEST 2022
port machine: LM0-400-22516.local