5-star rating: 2.5★ ⭐⭐½
Tier: Basic
Composite score: 51.3 / 100
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)

Available Variants

This repository contains multiple deployment formats. Pick the one that matches your use case:

Variant	Description
`lora/`	WindyStandard — our proprietary production baseline. Stable, reliable, optimized for GPU inference.
`lora-ct2-int8/`	WindyStandard · CPU INT8 — CTranslate2 quantized version of WindyStandard. ~25% of the size, 2–4× faster on CPU, no measurable quality loss.

Quick usage

Transformers (PyTorch):

python

from transformers import MarianMTModel, MarianTokenizer
tokenizer = MarianTokenizer.from_pretrained("WindyWord/translate-guw-en", subfolder="lora")
model = MarianMTModel.from_pretrained("WindyWord/translate-guw-en", subfolder="lora")

CTranslate2 (fast CPU inference):

python

import ctranslate2
translator = ctranslate2.Translator("path/to/translate-guw-en/lora-ct2-int8")

Commercial Use

The WindyWord.ai platform provides:

Mobile apps (iOS, Android — coming soon)
Real-time voice-to-text-to-translation pipeline
API access with premium model quality
Offline deployment support

Visit windyword.ai for apps and commercial API access.

Provenance & License

Weights derived from the OPUS-MT project (Helsinki-NLP/opus-mt-guw-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.

Licensed CC-BY-4.0 — attribution preserved as required.

Certified by Opus 4.6 Opus-Claw (Dr. C) on Veron-1 (RTX 5090). Patient file: clinic record