WindstormLabs/windy-turbo-ggml

Windy Turbo — ggml q5_1 (mobile build)

Windstorm Labs is the origin of record for this artifact. It is created here and promoted downstream; it is never edited anywhere else.

whisper.cpp / whisper.rn build of Windy Turbo, quantized to q5_1 so it runs on phones and tablets.

File	ggml-windy-turbo-q5_1.bin
Size	624.1 MB
SHA-256	d9ba2ed792b7fa1a89cf1f9b27ce819769f6520170397c77092af2d40a43b4c6
Quantization	q5_1
Languages	multilingual
Certified WER	5.56% (grade A)

Lineage

Stage	Value
Upstream base	openai/whisper-large-v3-turbo (809M params, MIT)
Windy fine-tune	LoRA, merged — Clinic patient windy-turbo
Converted from	WindyWord/listen-windy-turbo @ fa5e36f738624092d6da1524ee2db69404203458
Converter	whisper.cpp @ 2ca53bb45e38, models/convert-h5-to-ggml.py -> whisper-quantize q5_1
Built on	Veron-1-5090, 2026-08-01

Promotion path

WindstormLabs (here, origin) -> WindyWord (product) -> WindyWordMobileApp (mobile).

Downstream copies are push-only. Fix weights here, never downstream.

Honesty note

Distinct weights from upstream (different SHA-256). On short clean English clips the smaller models can be indistinguishable from stock Whisper — a formal bake-off has not been run. Do not claim quality gains without one.

Clinic patient file: sneakyfree/Windy-Clinic -> stt-models/windy-turbo.json

Licence

Upstream base is mit on Hugging Face, verified against the source repo we converted from. Note OpenAI's Whisper GitHub repo states MIT while its HF model cards state Apache-2.0; we follow the HF distribution point, since that is what we actually pulled. Windy weights are a derivative and are distributed under the same terms, with attribution upstream.