Back to all models
translation mask_token:
Query this model
🔥 This model is currently loaded and running on the Inference API. ⚠️ This model could not be loaded by the inference API. ⚠️ This model can be loaded on the Inference API on-demand.
JSON Output
API endpoint  

⚡️ Upgrade your account to access the Inference API

Share Copied link to clipboard

Monthly model downloads

Helsinki-NLP/opus-mt-sla-sla Helsinki-NLP/opus-mt-sla-sla
21 downloads
last 30 days

pytorch

tf

Contributed by

Language Technology Research Group at the University of Helsinki university
1 team member · 1325 models

How to use this model directly from the 🤗/transformers library:

			
Copy to clipboard
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM tokenizer = AutoTokenizer.from_pretrained("Helsinki-NLP/opus-mt-sla-sla") model = AutoModelForSeq2SeqLM.from_pretrained("Helsinki-NLP/opus-mt-sla-sla")
Uploaded in S3

sla-sla

  • source group: Slavic languages

  • target group: Slavic languages

  • OPUS readme: sla-sla

  • model: transformer

  • source language(s): bel bel_Latn bos_Latn bul bul_Latn ces dsb hrv hsb mkd orv_Cyrl pol rus slv srp_Cyrl srp_Latn ukr

  • target language(s): bel bel_Latn bos_Latn bul bul_Latn ces dsb hrv hsb mkd orv_Cyrl pol rus slv srp_Cyrl srp_Latn ukr

  • model: transformer

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

  • a sentence initial language token is required in the form of >>id<< (id = valid target language ID)

  • download original weights: opus-2020-07-27.zip

  • test set translations: opus-2020-07-27.test.txt

  • test set scores: opus-2020-07-27.eval.txt

Benchmarks

testset BLEU chr-F
newstest2012-cesrus.ces.rus 15.9 0.437
newstest2012-rusces.rus.ces 13.6 0.403
newstest2013-cesrus.ces.rus 19.8 0.473
newstest2013-rusces.rus.ces 17.9 0.449
Tatoeba-test.bel-bul.bel.bul 100.0 1.000
Tatoeba-test.bel-ces.bel.ces 33.5 0.630
Tatoeba-test.bel-hbs.bel.hbs 45.4 0.644
Tatoeba-test.bel-mkd.bel.mkd 19.3 0.531
Tatoeba-test.bel-pol.bel.pol 46.9 0.681
Tatoeba-test.bel-rus.bel.rus 58.5 0.767
Tatoeba-test.bel-ukr.bel.ukr 55.1 0.743
Tatoeba-test.bul-bel.bul.bel 10.7 0.423
Tatoeba-test.bul-ces.bul.ces 36.9 0.585
Tatoeba-test.bul-hbs.bul.hbs 53.7 0.807
Tatoeba-test.bul-mkd.bul.mkd 31.9 0.715
Tatoeba-test.bul-pol.bul.pol 38.6 0.607
Tatoeba-test.bul-rus.bul.rus 44.8 0.655
Tatoeba-test.bul-ukr.bul.ukr 49.9 0.691
Tatoeba-test.ces-bel.ces.bel 30.9 0.585
Tatoeba-test.ces-bul.ces.bul 75.8 0.859
Tatoeba-test.ces-hbs.ces.hbs 50.0 0.661
Tatoeba-test.ces-hsb.ces.hsb 7.9 0.246
Tatoeba-test.ces-mkd.ces.mkd 24.6 0.569
Tatoeba-test.ces-pol.ces.pol 44.3 0.652
Tatoeba-test.ces-rus.ces.rus 50.8 0.690
Tatoeba-test.ces-slv.ces.slv 4.9 0.240
Tatoeba-test.ces-ukr.ces.ukr 52.9 0.687
Tatoeba-test.dsb-pol.dsb.pol 16.3 0.367
Tatoeba-test.dsb-rus.dsb.rus 12.7 0.245
Tatoeba-test.hbs-bel.hbs.bel 32.9 0.531
Tatoeba-test.hbs-bul.hbs.bul 100.0 1.000
Tatoeba-test.hbs-ces.hbs.ces 40.3 0.626
Tatoeba-test.hbs-mkd.hbs.mkd 19.3 0.535
Tatoeba-test.hbs-pol.hbs.pol 45.0 0.650
Tatoeba-test.hbs-rus.hbs.rus 53.5 0.709
Tatoeba-test.hbs-ukr.hbs.ukr 50.7 0.684
Tatoeba-test.hsb-ces.hsb.ces 17.9 0.366
Tatoeba-test.mkd-bel.mkd.bel 23.6 0.548
Tatoeba-test.mkd-bul.mkd.bul 54.2 0.833
Tatoeba-test.mkd-ces.mkd.ces 12.1 0.371
Tatoeba-test.mkd-hbs.mkd.hbs 19.3 0.577
Tatoeba-test.mkd-pol.mkd.pol 53.7 0.833
Tatoeba-test.mkd-rus.mkd.rus 34.2 0.745
Tatoeba-test.mkd-ukr.mkd.ukr 42.7 0.708
Tatoeba-test.multi.multi 48.5 0.672
Tatoeba-test.orv-pol.orv.pol 10.1 0.355
Tatoeba-test.orv-rus.orv.rus 10.6 0.275
Tatoeba-test.orv-ukr.orv.ukr 7.5 0.230
Tatoeba-test.pol-bel.pol.bel 29.8 0.533
Tatoeba-test.pol-bul.pol.bul 36.8 0.578
Tatoeba-test.pol-ces.pol.ces 43.6 0.626
Tatoeba-test.pol-dsb.pol.dsb 0.9 0.097
Tatoeba-test.pol-hbs.pol.hbs 42.4 0.644
Tatoeba-test.pol-mkd.pol.mkd 19.3 0.535
Tatoeba-test.pol-orv.pol.orv 0.7 0.109
Tatoeba-test.pol-rus.pol.rus 49.6 0.680
Tatoeba-test.pol-slv.pol.slv 7.3 0.262
Tatoeba-test.pol-ukr.pol.ukr 46.8 0.664
Tatoeba-test.rus-bel.rus.bel 34.4 0.577
Tatoeba-test.rus-bul.rus.bul 45.5 0.657
Tatoeba-test.rus-ces.rus.ces 48.0 0.659
Tatoeba-test.rus-dsb.rus.dsb 10.7 0.029
Tatoeba-test.rus-hbs.rus.hbs 44.6 0.655
Tatoeba-test.rus-mkd.rus.mkd 34.9 0.617
Tatoeba-test.rus-orv.rus.orv 0.1 0.073
Tatoeba-test.rus-pol.rus.pol 45.2 0.659
Tatoeba-test.rus-slv.rus.slv 30.4 0.476
Tatoeba-test.rus-ukr.rus.ukr 57.6 0.751
Tatoeba-test.slv-ces.slv.ces 42.5 0.604
Tatoeba-test.slv-pol.slv.pol 39.6 0.601
Tatoeba-test.slv-rus.slv.rus 47.2 0.638
Tatoeba-test.slv-ukr.slv.ukr 36.4 0.549
Tatoeba-test.ukr-bel.ukr.bel 36.9 0.597
Tatoeba-test.ukr-bul.ukr.bul 56.4 0.733
Tatoeba-test.ukr-ces.ukr.ces 52.1 0.686
Tatoeba-test.ukr-hbs.ukr.hbs 47.1 0.670
Tatoeba-test.ukr-mkd.ukr.mkd 20.8 0.548
Tatoeba-test.ukr-orv.ukr.orv 0.2 0.058
Tatoeba-test.ukr-pol.ukr.pol 50.1 0.695
Tatoeba-test.ukr-rus.ukr.rus 63.9 0.790
Tatoeba-test.ukr-slv.ukr.slv 14.5 0.288

System Info:

  • hf_name: sla-sla

  • source_languages: sla

  • target_languages: sla

  • opus_readme_url: https://github.com/Helsinki-NLP/Tatoeba-Challenge/tree/master/models/sla-sla/README.md

  • original_repo: Tatoeba-Challenge

  • tags: ['translation']

  • languages: ['be', 'hr', 'mk', 'cs', 'ru', 'pl', 'bg', 'uk', 'sl', 'sla']

  • src_constituents: {'bel', 'hrv', 'orv_Cyrl', 'mkd', 'bel_Latn', 'srp_Latn', 'bul_Latn', 'ces', 'bos_Latn', 'csb_Latn', 'dsb', 'hsb', 'rus', 'srp_Cyrl', 'pol', 'rue', 'bul', 'ukr', 'slv'}

  • tgt_constituents: {'bel', 'hrv', 'orv_Cyrl', 'mkd', 'bel_Latn', 'srp_Latn', 'bul_Latn', 'ces', 'bos_Latn', 'csb_Latn', 'dsb', 'hsb', 'rus', 'srp_Cyrl', 'pol', 'rue', 'bul', 'ukr', 'slv'}

  • src_multilingual: True

  • tgt_multilingual: True

  • prepro: normalization + SentencePiece (spm32k,spm32k)

  • url_model: https://object.pouta.csc.fi/Tatoeba-MT-models/sla-sla/opus-2020-07-27.zip

  • url_test_set: https://object.pouta.csc.fi/Tatoeba-MT-models/sla-sla/opus-2020-07-27.test.txt

  • src_alpha3: sla

  • tgt_alpha3: sla

  • short_pair: sla-sla

  • chrF2_score: 0.672

  • bleu: 48.5

  • brevity_penalty: 1.0

  • ref_len: 59320.0

  • src_name: Slavic languages

  • tgt_name: Slavic languages

  • train_date: 2020-07-27

  • src_alpha2: sla

  • tgt_alpha2: sla

  • prefer_old: False

  • long_pair: sla-sla

  • helsinki_git_sha: 480fcbe0ee1bf4774bcbe6226ad9f58e63f6c535

  • transformers_git_sha: 2207e5d8cb224e954a7cba69fa4ac2309e9ff30b

  • port_machine: brutasse

  • port_time: 2020-08-21-14:41