Add model card
Browse files
README.md
ADDED
|
@@ -0,0 +1,80 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: cc-by-4.0
|
| 3 |
+
tags:
|
| 4 |
+
- translation
|
| 5 |
+
- marian
|
| 6 |
+
- windyword
|
| 7 |
+
- turkic
|
| 8 |
+
- turkish
|
| 9 |
+
- azerbaijani
|
| 10 |
+
- kazakh
|
| 11 |
+
- uzbek
|
| 12 |
+
- kyrgyz
|
| 13 |
+
- turkmen
|
| 14 |
+
- english
|
| 15 |
+
language:
|
| 16 |
+
- trk
|
| 17 |
+
- en
|
| 18 |
+
library_name: transformers
|
| 19 |
+
pipeline_tag: translation
|
| 20 |
+
---
|
| 21 |
+
|
| 22 |
+
# WindyWord.ai Translation — Turkic → English
|
| 23 |
+
|
| 24 |
+
**Translates Turkic (Turkish, Azerbaijani, Kazakh, Uzbek, Kyrgyz, Turkmen) → English.**
|
| 25 |
+
|
| 26 |
+
**Quality Rating: ⭐⭐½ (2.5★ Basic)**
|
| 27 |
+
|
| 28 |
+
Part of the [WindyWord.ai](https://windyword.ai) translation fleet — 1,800+ proprietary language pairs.
|
| 29 |
+
|
| 30 |
+
## Quality & Pricing Tier
|
| 31 |
+
|
| 32 |
+
- **5-star rating:** 2.5★ ⭐⭐½
|
| 33 |
+
- **Tier:** Basic
|
| 34 |
+
- **Composite score:** 51.1 / 100
|
| 35 |
+
- **Rated via:** Grand Rounds v2 — an 8-test stress battery (paragraphs, multi-paragraph, native input, domain stress, edge cases, round-trip fidelity, speed, and consistency checks)
|
| 36 |
+
|
| 37 |
+
## Available Variants
|
| 38 |
+
|
| 39 |
+
This repository contains multiple deployment formats. Pick the one that matches your use case:
|
| 40 |
+
|
| 41 |
+
| Variant | Description |
|
| 42 |
+
|---|---|
|
| 43 |
+
| `lora/` | **WindyStandard** — our proprietary production baseline. Stable, reliable, optimized for GPU inference. |
|
| 44 |
+
| `lora-ct2-int8/` | **WindyStandard · CPU INT8** — CTranslate2 quantized version of WindyStandard. ~25% of the size, 2–4× faster on CPU, no measurable quality loss. |
|
| 45 |
+
|
| 46 |
+
### Quick usage
|
| 47 |
+
|
| 48 |
+
**Transformers (PyTorch):**
|
| 49 |
+
```python
|
| 50 |
+
from transformers import MarianMTModel, MarianTokenizer
|
| 51 |
+
tokenizer = MarianTokenizer.from_pretrained("WindyWord/translate-trk-en", subfolder="lora")
|
| 52 |
+
model = MarianMTModel.from_pretrained("WindyWord/translate-trk-en", subfolder="lora")
|
| 53 |
+
```
|
| 54 |
+
|
| 55 |
+
**CTranslate2 (fast CPU inference):**
|
| 56 |
+
```python
|
| 57 |
+
import ctranslate2
|
| 58 |
+
translator = ctranslate2.Translator("path/to/translate-trk-en/lora-ct2-int8")
|
| 59 |
+
```
|
| 60 |
+
|
| 61 |
+
## Commercial Use
|
| 62 |
+
|
| 63 |
+
The WindyWord.ai platform provides:
|
| 64 |
+
- **Mobile apps** (iOS, Android — coming soon)
|
| 65 |
+
- **Real-time voice-to-text-to-translation** pipeline
|
| 66 |
+
- **API access** with premium model quality
|
| 67 |
+
- **Offline deployment** support
|
| 68 |
+
|
| 69 |
+
Visit [windyword.ai](https://windyword.ai) for apps and commercial API access.
|
| 70 |
+
|
| 71 |
+
---
|
| 72 |
+
|
| 73 |
+
## Provenance & License
|
| 74 |
+
|
| 75 |
+
Weights derived from the OPUS-MT project ([Helsinki-NLP/opus-mt-trk-en](https://huggingface.co/Helsinki-NLP/opus-mt-trk-en)) under CC-BY-4.0. WindyStandard, WindyEnhanced, and WindyScripture variants are proprietary to WindyWord.ai, independently trained and quality-certified via our Grand Rounds v2 test battery.
|
| 76 |
+
|
| 77 |
+
Licensed CC-BY-4.0 — attribution preserved as required.
|
| 78 |
+
|
| 79 |
+
*Certified by Opus 4.6 Opus-Claw (Dr. C) on Veron-1 (RTX 5090).*
|
| 80 |
+
*Patient file: [clinic record](https://github.com/sneakyfree/Windy-Clinic/blob/main/translation-pairs/trk-en.json)*
|