metadata
language:
- arn
- es
license: cc-by-4.0
tags:
- translation
- nllb
- mapudungun
- low-resource-mt
- morphology
base_model: facebook/nllb-200-nllb-200-3.3b
mapudungun-nllb-3.3B-arn-es-standard-bpe
Fine-tuned NLLB-200 3.3B for Mapudungun→Spanish translation using the Standard BPE tokenization condition.
NLLB-200's default SentencePiece BPE, no modification.
Part of the paper: Bringing Mapudungun into the Modern MT Ecosystem: Morphology-Aware Tokenization for NLLB-200 Fine-Tuning (AmericasNLP 2026 @ ACL).
Usage
from transformers import pipeline
pipe = pipeline(
"translation",
model="byumatrixlab/mapudungun-nllb-3.3B-arn-es-standard-bpe",
src_lang="arn_Latn",
tgt_lang="spa_Latn",
)
print(pipe("your text here", max_length=256))
Citation
@inproceedings{thompson2026mapudungun,
title = {Bringing {Mapudungun} into the Modern {MT} Ecosystem: Morphology-Aware Tokenization for {NLLB}-200 Fine-Tuning},
author = {Thompson, Isaac},
booktitle = {Proceedings of the 5th Workshop on NLP for Indigenous Languages of the Americas (AmericasNLP 2026)},
year = {2026},
}