fsmn-vad (ONNX, repackaged for vadonnx)
ONNX voice activity detection model packaged for use with
vadonnx.
- Upstream source: funasr/fsmn-vad-onnx (card: apache-2.0), model from FunASR (Alibaba DAMO Academy)
- License: FunASR Model Open Source License Agreement v1.1 for the weights. MIT is the FunASR code license, not the weights license.
- Sample rate: 16000 Hz
- Frame size: 160 samples
- Stateful: True
This repository redistributes the model in ONNX form together with a
signature.json describing its input/output wiring. All rights and the original
license belong to the upstream authors.
License terms
The FunASR Model Open Source License Agreement v1.1 (the license_link above)
covers FunASR model weights and their derivatives. In substance: you may use, copy,
modify and share the weights. You must attribute the source and author and keep the
model names. Unjustified denigration of FunASR forfeits the license. Violation
terminates it. Alibaba may revise the agreement, with automatic effect for continued
use. The upstream Hugging Face card summarises the license as apache-2.0.
Attribution: FSMN-VAD, FunASR, Alibaba DAMO
Academy, model fsmn-vad (speech_fsmn_vad_zh-cn-16k-common).
Usage
from vadonnx import load_vad
vad = load_vad("fsmn")
segments = vad.get_speech_segments(audio, sample_rate=16000)