tiedeman's picture
Initial commit
e1fb2f3
metadata
language:
  - da
  - nb
  - sv
  - tr
tags:
  - translation
  - opus-mt-tc
license: cc-by-4.0
model-index:
  - name: opus-mt-tc-big-gmq-tr
    results:
      - task:
          name: Translation dan-tur
          type: translation
          args: dan-tur
        dataset:
          name: flores101-devtest
          type: flores_101
          args: dan tur devtest
        metrics:
          - name: BLEU
            type: bleu
            value: 23.4
          - name: chr-F
            type: chrf
            value: 0.56363
      - task:
          name: Translation nob-tur
          type: translation
          args: nob-tur
        dataset:
          name: flores101-devtest
          type: flores_101
          args: nob tur devtest
        metrics:
          - name: BLEU
            type: bleu
            value: 19
          - name: chr-F
            type: chrf
            value: 0.52696
      - task:
          name: Translation swe-tur
          type: translation
          args: swe-tur
        dataset:
          name: flores101-devtest
          type: flores_101
          args: swe tur devtest
        metrics:
          - name: BLEU
            type: bleu
            value: 22.2
          - name: chr-F
            type: chrf
            value: 0.54996
      - task:
          name: Translation dan-tur
          type: translation
          args: dan-tur
        dataset:
          name: tatoeba-test-v2021-08-07
          type: tatoeba_mt
          args: dan-tur
        metrics:
          - name: BLEU
            type: bleu
            value: 43
          - name: chr-F
            type: chrf
            value: 0.6783
      - task:
          name: Translation swe-tur
          type: translation
          args: swe-tur
        dataset:
          name: tatoeba-test-v2021-08-07
          type: tatoeba_mt
          args: swe-tur
        metrics:
          - name: BLEU
            type: bleu
            value: 34.2
          - name: chr-F
            type: chrf
            value: 0.63653

opus-mt-tc-big-gmq-tr

Table of Contents

Model Details

Neural machine translation model for translating from North Germanic languages (gmq) to Turkish (tr).

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. Model Description:

Uses

This model can be used for translation and text-to-text generation.

Risks, Limitations and Biases

CONTENT WARNING: Readers should be aware that the model is trained on various public data sets that may contain content that is disturbing, offensive, and can propagate historical and current stereotypes.

Significant research has explored bias and fairness issues with language models (see, e.g., Sheng et al. (2021) and Bender et al. (2021)).

How to Get Started With the Model

A short example code:

from transformers import MarianMTModel, MarianTokenizer

src_text = [
    "Aftensmaden dufter lækkert.",
    "Vi ser vad som händer tillsammans."
]

model_name = "pytorch-models/opus-mt-tc-big-gmq-tr"
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:
#     Akşam yemeği çok güzel kokuyor.
#     Birlikte bakalım neler olacak.

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-tr")
print(pipe("Aftensmaden dufter lækkert."))

# expected output: Akşam yemeği çok güzel kokuyor.

Training

Evaluation

langpair testset chr-F BLEU #sent #words
dan-tur tatoeba-test-v2021-08-07 0.67830 43.0 758 3436
swe-tur tatoeba-test-v2021-08-07 0.63653 34.2 203 1008
dan-tur flores101-devtest 0.56363 23.4 1012 20253
nob-tur flores101-devtest 0.52696 19.0 1012 20253
swe-tur flores101-devtest 0.54996 22.2 1012 20253

Citation Information

@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",
}

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: 8b9f0b0
  • port time: Sat Aug 13 00:05:52 EEST 2022
  • port machine: LM0-400-22516.local