Hindi speech recognition model listen-windy-lingua-hi-ct2
Open speech-to-text model published by Windstorm Labs on Hugging Face.
- Task
- Speech-to-text (automatic speech recognition)
- Language
- Hindi
- Lineage
- Whisper community fine-tune
- Formats in the repo
ct2-int8- Hugging Face
- WindyWord/listen-windy-lingua-hi-ct2
- Last updated on HF
- 2026-10-01
Quality
Not benchmarked by this site yet. Speech models have no published word-error rates here; any figures in the Hugging Face card were not re-measured by us.
From the model card: Outputs Hindi audio as Latin-script Hinglish, NOT Devanagari. FLEURS-Devanagari WER ≈100% is a script mismatch, not a quality failure. Useful for code-switched / chat / SMS contexts. For Devanagari output, use a separate model (not yet shipped).
How to run
pip install faster-whisper soundfileimport soundfile as sf
from huggingface_hub import snapshot_download
from faster_whisper import WhisperModel
path = snapshot_download("WindyWord/listen-windy-lingua-hi-ct2", allow_patterns=["ct2-int8/*"])
model = WhisperModel(f"{path}/ct2-int8", device="cpu", compute_type="int8")
audio, sr = sf.read("audio.wav", dtype="float32") # 16 kHz mono
segments, info = model.transcribe(audio)
print("".join(s.text for s in segments))Tested: Pattern run on 2026-10-01 against the ct2-int8 build of WindyWord/listen-windy-nano with faster-whisper 1.2.1 on CPU. Not run on every model individually.
Licence and attribution
- Licence
- Apache-2.0
- Upstream
- Oriserve/Whisper-Hindi2Hinglish-Swift
- Upstream licence
- Apache-2.0
- Attribution
- Based on Oriserve/Whisper-Hindi2Hinglish-Swift, a community fine-tune of OpenAI Whisper. Upstream licence not stated.
Copy this into your NOTICE file (or equivalent) when you ship the model:
This product uses the model WindyWord/listen-windy-lingua-hi-ct2
(https://huggingface.co/WindyWord/listen-windy-lingua-hi-ct2), published by Windstorm Labs (https://windstormlabs.com).
Based on Oriserve/Whisper-Hindi2Hinglish-Swift, a community fine-tune of OpenAI Whisper. Upstream licence not stated.
Imported by Windstorm Labs from an upstream community fine-tune; see the model card.
Licensed under the Apache License, Version 2.0: https://www.apache.org/licenses/LICENSE-2.0This is a convenience, not legal advice. See Licensing.