translate-de-no
Part of the Windstorm Labs open model catalogue. Scores, licence and attribution for every model: https://windytranslate.com/models/translate-de-no
Machine translation model, de to no, published by Windstorm Labs.
Attribution
Derived from Helsinki-NLP/opus-mt-de-no, licensed apache-2.0.
Windstorm Labs did not train the original model. This notice provides the attribution the
licence requires.
What was changed
The weights in this repository have been modified from the original. A deliberate, documented and reproducible numerical transformation applied to the base weights. It is not a training run and is not described as one.
| Method | epsilon-noise-fog |
| Tensors modified | 72 of 254 |
| Max absolute weight delta | 5.215e-07 |
| Verified | tensor-by-tensor against the upstream original, 2026-07-26 |
The comparison is tensor-level rather than file-level: safetensors and PyTorch .bin
containers hash differently even when the tensors inside are identical, so a file-hash
mismatch would prove nothing.
Contents
A single transformers build in safetensors format, loadable directly:
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tok = AutoTokenizer.from_pretrained("WindyTranslate/translate-de-no")
model = AutoModelForSeq2SeqLM.from_pretrained("WindyTranslate/translate-de-no")
batch = tok(["<your text here>"], return_tensors="pt", padding=True)
print(tok.batch_decode(model.generate(**batch), skip_special_tokens=True))
Other builds of this pair, including CTranslate2 INT8, are not included in this repository.
Limitations
- A single language direction:
detono. - Not yet scored. No quality benchmark has been run on this model yet. The weights are verified to differ from the original; that is a statement about provenance, not about translation quality.
- Quality is not monotonic with model size or with the amount of fine-tuning applied.