Edit model card

OPUS Tatoeba English-Polish

Update: The model is currently not functional. Please refer to the original checkpoint in the Tatoeba repository for a working version

This model was obtained by running the script convert_marian_to_pytorch.py with the flag -m eng-pol. The original models were trained by J�rg Tiedemann using the MarianNMT library. See all available MarianMTModel models on the profile of the Helsinki NLP group.

  • source language name: English

  • target language name: Polish

  • OPUS readme: README.md

  • model: transformer

  • source language code: en

  • target language code: pl

  • dataset: opus

  • release date: 2021-02-19

  • pre-processing: normalization + SentencePiece (spm32k,spm32k)

  • download original weights: opus-2021-02-19.zip

  • Training data:

    • eng-pol: Tatoeba-train (59742979)
  • Validation data:

    • eng-pol: Tatoeba-dev, 44146
    • total-size-shuffled: 44145
    • devset-selected: top 5000 lines of Tatoeba-dev.src.shuffled!
  • Test data:

    • Tatoeba-test.eng-pol: 10000/64925
  • test set translations file: test.txt

  • test set scores file: eval.txt

  • BLEU-scores

    Test set score
    Tatoeba-test.eng-pol 47.5
  • chr-F-scores

    Test set score
    Tatoeba-test.eng-pol 0.673

System Info:

  • hf_name: eng-pol
  • source_languages: en
  • target_languages: pl
  • opus_readme_url: https://object.pouta.csc.fi/Tatoeba-MT-models/eng-pol/opus-2021-02-19.zip/README.md
  • original_repo: Tatoeba-Challenge
  • tags: ['translation']
  • languages: ['en', 'pl']
  • src_constituents: ['eng']
  • tgt_constituents: ['pol']
  • src_multilingual: False
  • tgt_multilingual: False
  • helsinki_git_sha: 70b0a9621f054ef1d8ea81f7d55595d7f64d19ff
  • transformers_git_sha: 7c6cd0ac28f1b760ccb4d6e4761f13185d05d90b
  • port_machine: databox
  • port_time: 2021-10-18-15:11
Downloads last month
138
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for gsarti/opus-mt-tc-en-pl

Finetunes
1 model