WindstormLabs/windy-edge-ggml

Windy Edge — 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 Edge, quantized to q5_1 so it runs on phones and tablets.

File	ggml-windy-edge-q5_1.bin
Size	584.6 MB
SHA-256	6b225f4bcd02a6aedc4b3a810cd34a0028350b0550f0e794d144cdee61353583
Quantization	q5_1
Languages	ENGLISH ONLY
Certified WER	7.31% (grade A)

Lineage

Stage	Value
Upstream base	distil-whisper/distil-large-v3 (756M params, MIT)
Windy fine-tune	LoRA, merged — Clinic patient windy-edge
Converted from	WindyWord/listen-windy-edge @ 9566640871ed5f9fc2c4cf8001d9064a8def3b8b
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-edge.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.