Derived from Helsinki-NLP/opus-mt-gil-fi, 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 LoRA fine-tune trained on parallel corpus data and merged into the base weights.


Method	`windy-hallmark`
Tensors modified	36 of 254
Max absolute weight delta	6.530e-05
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.

A single transformers build in safetensors format, loadable directly:

python

from transformers import AutoTokenizer, AutoModelForSeq2SeqLM

tok = AutoTokenizer.from_pretrained("WindyTranslate/translate-gil-fi")
model = AutoModelForSeq2SeqLM.from_pretrained("WindyTranslate/translate-gil-fi")

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: gil to fi.
No quality benchmark has been run on this pair. 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.

WindyTranslate/translate-gil-fi

translate-gil-fi

Attribution

What was changed

Contents

Limitations

WindyTranslate/
translate-gil-fi