--- license: apache-2.0 tags: - translation - marian - windyword - french-based-creoles - haitian-creole - mauritian-creole - seychellois-creole - réunion-creole - english language: - cpf - en library_name: transformers pipeline_tag: translation base_model: Helsinki-NLP/opus-mt-cpf-en --- # WindyWord.ai Translation — French-based Creoles → English > Part of the Windstorm Labs open model catalogue. Scores, licence and attribution for every model: https://windytranslate.com/models/translate-cpf-en > > Canonical copy: https://huggingface.co/WindyTranslate/translate-cpf-en **Translates French-based Creoles (Haitian Creole, Mauritian Creole, Seychellois Creole, Réunion Creole) → English.** ## Quality No quality score is published in this repository. Current screening scores, where measured, are on the catalogue page linked above. ## Available Variants Deployment formats in this repository (subfolders): | Variant | Description | |---|---| | `lora/` | **WindyStandard** — production baseline. Transformers format for GPU inference. | | `lora-ct2-int8/` | **WindyStandard · CPU INT8** — CTranslate2 INT8 quantization of WindyStandard for CPU inference. | | `herm0/` | **WindyEnhanced** — further fine-tuned on the OPUS-100, Tatoeba and WikiMatrix parallel corpora. | | `herm0-ct2-int8/` | **WindyEnhanced · CPU INT8** — CTranslate2 INT8 quantization of WindyEnhanced. | ### Quick usage **Transformers (PyTorch):** ```python from transformers import MarianMTModel, MarianTokenizer tokenizer = MarianTokenizer.from_pretrained("WindyWord/translate-cpf-en", subfolder="lora") model = MarianMTModel.from_pretrained("WindyWord/translate-cpf-en", subfolder="lora") ``` **CTranslate2 (fast CPU inference):** ```python import ctranslate2 translator = ctranslate2.Translator("path/to/translate-cpf-en/lora-ct2-int8") ``` ## Apps The [Windy Word](https://windyword.ai) apps are built on this model family. --- ## Provenance & License Weights derived from [Helsinki-NLP/opus-mt-cpf-en](https://huggingface.co/Helsinki-NLP/opus-mt-cpf-en) (OPUS-MT, Helsinki-NLP, University of Helsinki), licensed Apache-2.0. Windy variants are released under the same licence. Licence text: [`LICENSE`](LICENSE). Attribution: [`NOTICE.md`](NOTICE.md).