Downloads · 30 days
0
WindyTranslate/translate-af-fi
translate-af-fi is a translation model from WindyTranslate. Use it when you need text moved from one language to another. It is set up for transformers. The card lists the license as apache-2.0.
Part of the Windstorm Labs open model catalogue. Scores, licence and attribution for every model: https://windytranslate.com/models/translate-af-fi
Downloads · 30 days
0
Access
Public
Updated Oct 1, 2026
Repo size
758 MB
Likes
0
Public
Click a slice to open those files.
.safetensors602 MB · 78%
From the Hugging Face model README
Part of the Windstorm Labs open model catalogue. Scores, licence and attribution for every model: https://windytranslate.com/models/translate-af-fi
Translates Afrikaans → Finnish.
| Benchmark | chrF++ | BLEU | Band |
|---|---|---|---|
| FLORES-200 dev, 48 sentences (Afrikaans → Finnish) | 41.05 | 11.75 | Usable |
Measured 2026-08-04; screening score, not a publication result. Bands: Excellent ≥60, Good 45–60, Usable 30–45, Limited 15–30, Not recommended <15 (chrF++).
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. |
Transformers (PyTorch):
from transformers import MarianMTModel, MarianTokenizer
tokenizer = MarianTokenizer.from_pretrained("WindyTranslate/translate-af-fi", subfolder="lora")
model = MarianMTModel.from_pretrained("WindyTranslate/translate-af-fi", subfolder="lora")
CTranslate2 (fast CPU inference):
import ctranslate2
translator = ctranslate2.Translator("path/to/translate-af-fi/lora-ct2-int8")
The Windy Word apps are built on this model family.
Weights derived from Helsinki-NLP/opus-mt-af-fi (OPUS-MT, Helsinki-NLP, University of Helsinki), licensed Apache-2.0. Windy variants are released under the same licence.