Skip to content

EthenEthenEthen

Open Source Model Profile · Helsinki-NLP

opus-mt-en-xh

opus-mt-en-xh is a Helsinki-NLP Marian model for English-to-Xhosa translation with a reported 37.9 BLEU on JW300.

Publisher
Helsinki-NLP
Task
translation
Model type
marian
License
apache-2.0
Library
transformers
Publication status
Accepted · not indexed

Model overview

opus-mt-en-xh is published by Helsinki-NLP as a Marian translation model. The captured configuration identifies MarianMTModel. According to the model card, it translates English to Xhosa using a transformer-align setup on OPUS data, and card data records apache-2.0.

Recorded capabilities

English-to-Xhosa translation

According to the model card, the source language is English and the target language is Xhosa, with hub tags recording en and xh.

Marian transformer-align

The captured configuration identifies MarianMTModel, and the model card describes a transformer-align model with normalization and SentencePiece preprocessing.

Reported JW300 scores

According to the model card, the JW300 English-Xhosa test set reports 37.9 BLEU and 0.652 chr-F with published test translations and scores.

Use cases in the source record

  • English-to-Xhosa translation work using the publisher's documented OPUS-trained Marian setup.
  • Benchmark comparison on the JW300 English-Xhosa test set using the reported BLEU and chr-F figures.

Limitations and unknowns

  • No parameter count was extracted from this record.
  • No context-window value was extracted from this record.
  • No evaluation results beyond the publisher-reported JW300 BLEU and chr-F figures were extracted.
  • Provider state is historical snapshot data and should be refreshed before being presented as current.

Source and provenance

Source: Helsinki-NLP/opus-mt-en-xh

Captured: Unknown. Processed: 2026-09-07T19:34:31.430367+00:00.

opus-mt-en-xh source languages: en target languages: xh OPUS readme: en-xh dataset: opus model: transformer-align pre-processing: normalization + SentencePiece download original weights: opus-2020-01-08.zip test set translations: opus-2020-01-08.test.txt test set scores: opus-2020-01-08.eval.txt Benchmarks testset BLEU chr-F JW300.en.xh 37.9 0.652

F001F002F003F004F005F006F008F009F010