Windy TranslateEnglish to Umbundu translation model translate-en-umb
Open machine translation from English into Umbundu, published by Windstorm Labs on Hugging Face.
- Task
- Translation: English → Umbundu
- Languages
- English en, Umbundu umb
- Lineage
- OPUS-MT
- Format
- transformers (Marian), safetensors
- Hugging Face
- WindyTranslate/translate-en-umb
- Last updated on HF
- 2026-08-04
Quality
| Benchmark | chrF++ | BLEU | Band | Measured |
|---|---|---|---|---|
| FLORES-200 dev, 48 sentences | 21.16 | 0.67 | Limited | 2026-08-04 |
A screening score on 48 sentences, not a publication result. Bands: Excellent ≥ 60, Good 45–60, Usable 30–45, Limited 15–30, Not recommended < 15. Method and caveats.
Recorded output (sentence 1, first 110 characters)
- Source
- On Monday, scientists from the Stanford University School of Medicine announced the invention of a new diagnos…
- Reference
- K’eteke lyavali lyosemana, olonoño viosikola yavelapo yesaku yo Stanford vasapula etungo lyocina cimwe cikwati…
- This model
- Keteke likuavo, olonoño vio Kosikola ya velapo yo Gileada, via sandeka omakina yokaliye yi tukuiwa hati, omaki
How to run
Shell
pip install transformers sentencepiece torchPython
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
repo = "WindyTranslate/translate-en-umb"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForSeq2SeqLM.from_pretrained(repo)
batch = tok(["Where is the train station?"], return_tensors="pt", padding=True)
print(tok.batch_decode(model.generate(**batch), skip_special_tokens=True)[0])Tested: Pattern run on 2026-10-01 against WindyTranslate/translate-tc-big-en-fr with transformers 4.57.6 and 5.18.0 on CPU ("Où se trouve la gare ?"); multi-target token prefix run against translate-tc-bible-big-mul-deu_eng_nld. Not run on every model individually.
Licence and attribution
- Licence
- Apache-2.0
- Upstream
- Helsinki-NLP/opus-mt-en-umb
- Upstream licence
- Apache-2.0
- Attribution
- Based on OPUS-MT model Helsinki-NLP/opus-mt-en-umb by Helsinki-NLP (Language Technology Research Group, University of Helsinki). Licensed Apache-2.0.
Copy this into your NOTICE file (or equivalent) when you ship the model:
NOTICE
This product uses the model WindyTranslate/translate-en-umb
(https://huggingface.co/WindyTranslate/translate-en-umb), published by Windstorm Labs (https://windstormlabs.com).
Based on OPUS-MT model Helsinki-NLP/opus-mt-en-umb by Helsinki-NLP (Language Technology Research Group, University of Helsinki). Licensed Apache-2.0.
Windstorm Labs modified the upstream weights; the model card describes what was changed.
Licensed under the Apache License, Version 2.0: https://www.apache.org/licenses/LICENSE-2.0This is a convenience, not legal advice. See Licensing.
Related models
| Model | chrF++ (FLORES-200 dev) | Licence |
|---|---|---|
| Umbundu to English translate-umb-en | 18.8 Limited | Apache-2.0 |
| English to French translate-tc-big-en-fr | 69.6 Excellent | CC-BY-4.0 |
| English to Portuguese translate-tc-big-en-pt | 66.6 Excellent | CC-BY-4.0 |
| English to Bulgarian translate-tc-big-en-bg | 66.6 Excellent | CC-BY-4.0 |
| English to French translate-en-fr | 66.2 Excellent | Apache-2.0 |
| English to Danish translate-en-da | 65.6 Excellent | Apache-2.0 |
| English to Swedish translate-en-sv | 64.7 Excellent | Apache-2.0 |
| English to Romanian translate-tc-big-en-ro | 64.4 Excellent | CC-BY-4.0 |