translate-gaa-es

Machine translation model, gaa to es, published by Windstorm Labs.

Attribution

Derived from Helsinki-NLP/opus-mt-gaa-es, 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	lora-fog
Tensors modified	36 of 254
Max absolute weight delta	6.354e-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.

Contents

A single transformers build in safetensors format, loadable directly:

from transformers import AutoTokenizer, AutoModelForSeq2SeqLM

tok = AutoTokenizer.from_pretrained("WindyTranslate/translate-gaa-es")
model = AutoModelForSeq2SeqLM.from_pretrained("WindyTranslate/translate-gaa-es")

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: gaa to es.
• 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.