crash-sv/scribe-translate-ct2

Opus-MT for Scribe SV (CTranslate2)

CTranslate2 conversions of fourteen Helsinki-NLP Opus-MT models. They are used by Scribe SV, a Windows dictation and translation utility, and are published here so the application can download them on demand instead of shipping them inside the installer. The application downloads only the folders of the language pairs it needs.

Nothing was retrained or fine-tuned: these are format conversions of the original models.

Contents

folder	direction	source model	license
opus-mt-tc-big-zle-en-ct2	East Slavic (ru, uk, be) → English	Helsinki-NLP/opus-mt-tc-big-zle-en	CC-BY-4.0
opus-mt-tc-big-en-zle-ct2	English → East Slavic (ru, uk, be)	Helsinki-NLP/opus-mt-tc-big-en-zle	CC-BY-4.0
opus-mt-tc-big-cat_oci_spa-en-ct2	Catalan, Occitan, Spanish → English	Helsinki-NLP/opus-mt-tc-big-cat_oci_spa-en	CC-BY-4.0
opus-mt-tc-big-en-es-ct2	English → Spanish	Helsinki-NLP/opus-mt-tc-big-en-es	CC-BY-4.0
opus-mt-tc-big-fr-en-ct2	French → English	Helsinki-NLP/opus-mt-tc-big-fr-en	CC-BY-4.0
opus-mt-tc-big-en-fr-ct2	English → French	Helsinki-NLP/opus-mt-tc-big-en-fr	CC-BY-4.0
opus-mt-de-en-ct2	German → English	Helsinki-NLP/opus-mt-de-en	Apache-2.0
opus-mt-en-de-ct2	English → German	Helsinki-NLP/opus-mt-en-de	CC-BY-4.0
opus-mt-tc-big-it-en-ct2	Italian → English	Helsinki-NLP/opus-mt-tc-big-it-en	CC-BY-4.0
opus-mt-tc-big-en-it-ct2	English → Italian	Helsinki-NLP/opus-mt-tc-big-en-it	CC-BY-4.0
opus-mt-roa-en-ct2	Romance languages → English	Helsinki-NLP/opus-mt-roa-en	Apache-2.0
opus-mt-tc-big-en-pt-ct2	English → Portuguese	Helsinki-NLP/opus-mt-tc-big-en-pt	CC-BY-4.0
opus-mt-zh-en-ct2	Chinese → English	Helsinki-NLP/opus-mt-zh-en	CC-BY-4.0
opus-mt-en-zh-ct2	English → Chinese	Helsinki-NLP/opus-mt-en-zh	Apache-2.0

Each folder holds a complete CTranslate2 model with its SentencePiece and tokenizer files, and the text of its license.

Language token

Three models are multi-target. When translating into their languages, the source sentence must start with a target-language token:

model	token
opus-mt-tc-big-en-zle (Russian)	>>rus<<
opus-mt-tc-big-en-pt (Portuguese)	>>por<<
opus-mt-en-zh (Simplified Chinese)	>>cmn_Hans<<

>>rus<< The build failed.

Citation

Attribution to Helsinki-NLP is required when using these models.

• Tiedemann, J., Thottingal, S. OPUS-MT — Building open translation services for the World. EAMT 2020.
• Tiedemann, J. The Tatoeba Translation Challenge — Realistic Data Sets for Low Resource and Multilingual MT. WMT 2020.

Attribution and license

OPUS-MT translation models, CTranslate2 (CC-BY-4.0 group)

• Author: Helsinki-NLP, Language Technology Research Group, University of Helsinki
• License: CC-BY-4.0 — https://creativecommons.org/licenses/by/4.0/
• Upstream: https://huggingface.co/Helsinki-NLP/opus-mt-tc-big-zle-en
• Upstream: https://huggingface.co/Helsinki-NLP/opus-mt-tc-big-en-zle
• Upstream: https://huggingface.co/Helsinki-NLP/opus-mt-tc-big-cat_oci_spa-en
• Upstream: https://huggingface.co/Helsinki-NLP/opus-mt-tc-big-en-es
• Upstream: https://huggingface.co/Helsinki-NLP/opus-mt-tc-big-fr-en
• Upstream: https://huggingface.co/Helsinki-NLP/opus-mt-tc-big-en-fr
• Upstream: https://huggingface.co/Helsinki-NLP/opus-mt-en-de
• Upstream: https://huggingface.co/Helsinki-NLP/opus-mt-tc-big-it-en
• Upstream: https://huggingface.co/Helsinki-NLP/opus-mt-tc-big-en-it
• Upstream: https://huggingface.co/Helsinki-NLP/opus-mt-tc-big-en-pt
• Upstream: https://huggingface.co/Helsinki-NLP/opus-mt-zh-en
• Folders: opus-mt-en-de-ct2, opus-mt-tc-big-cat_oci_spa-en-ct2, opus-mt-tc-big-en-es-ct2, opus-mt-tc-big-en-fr-ct2, opus-mt-tc-big-en-it-ct2, opus-mt-tc-big-en-pt-ct2, opus-mt-tc-big-en-zle-ct2, opus-mt-tc-big-fr-en-ct2, opus-mt-tc-big-it-en-ct2, opus-mt-tc-big-zle-en-ct2, opus-mt-zh-en-ct2
• Changes: CTranslate2 conversion of the original models; no retraining.

License text (LICENSE-CC-BY-4.0.txt) lies in each of its folders.

OPUS-MT translation models, CTranslate2 (Apache-2.0 group)

• Author: Helsinki-NLP, Language Technology Research Group, University of Helsinki
• License: Apache-2.0 — https://www.apache.org/licenses/LICENSE-2.0
• Upstream: https://huggingface.co/Helsinki-NLP/opus-mt-de-en
• Upstream: https://huggingface.co/Helsinki-NLP/opus-mt-roa-en
• Upstream: https://huggingface.co/Helsinki-NLP/opus-mt-en-zh
• Folders: opus-mt-de-en-ct2, opus-mt-en-zh-ct2, opus-mt-roa-en-ct2
• Changes: CTranslate2 conversion of the original models; no retraining.

License text (LICENSE-Apache-2.0.txt) lies in each of its folders.

This republication keeps the original licenses.