listen-windy-lingua-am-ct2

WindyWord.ai STT — Amharic Lingua (CPU INT8 (CTranslate2))

Part of the Windstorm Labs open model catalogue. Scores, licence and attribution for every model: https://windytranslate.com/models/listen-windy-lingua-am-ct2

Licence under review: the upstream model b1n1yam/shook-medium-amharic-2k does not declare a licence on Hugging Face. Do not rely on this repository for commercial use until this note is removed.

Transcribes Amharic speech (Afro-Asiatic > Semitic > South Semitic).

Note: Quality caveat: ported from legacy upload that audited at 119% WER on FLEURS Amharic. Not retired pending a better community fine-tune; use with caution.

Quality

• FLEURS WER: 118.9% (50-sample audit)
• CER: 1.1203
• Band: unusable-gap
• Source: our 50-sample FLEURS screening (not a publication result)

About this variant

This is the ct2-int8 deployment format of our Amharic Lingua STT model. Load it via the ct2-int8/ subfolder.

Part of the WindyWord.ai STT fleet — covering 35+ languages that commercial speech-to-text APIs underserve, with proper dialect / script disclosures where they matter.

Usage

from transformers import WhisperForConditionalGeneration, WhisperProcessor
processor = WhisperProcessor.from_pretrained("WindyWord/listen-windy-lingua-am-ct2", subfolder="ct2-int8")
model = WhisperForConditionalGeneration.from_pretrained("WindyWord/listen-windy-lingua-am-ct2", subfolder="ct2-int8")

Apps

The Windy Word apps are built on this model family.

---

Provenance & License

CTranslate2 INT8 conversion of WindyWord/listen-windy-lingua-am, whose weights derive from b1n1yam/shook-medium-amharic-2k (community fine-tune by b1n1yam). Lineage explicit in commit history.

Licence under review: the upstream model b1n1yam/shook-medium-amharic-2k does not declare a licence on Hugging Face. Do not rely on this repository for commercial use until this note is removed.