How to use Silero VAD in iOS and how to convert the Onnx to CoreMLmodel

[Kamalesh7605]

Hi everyone, I'm working on integrating Voice Activity Detection (VAD) technology using the Silero model in my iOS app. I would appreciate any guidance on how to implement the Silero model in an iOS environment, including best practices and relevant resources. Additionally, I need assistance with converting the ONNX model to a Core ML model. Any help would be greatly appreciated!

[snakers4]

Hi,

Silero model in an iOS environment

Probably using ONNX runtime is the easiest option.

converting the ONNX model to a Core ML model

Typically, such things are what we do commercially for companies for a fee.

[Kamalesh7605]

Hi, I’m looking for guidance on how to integrate an ONNX model into an iOS app. If there is any official documentation or detailed resources available for iOS integration using ONNX Runtime, could you please share them with me? Thanks!

[luizlacerdam]

Hello @snakers4 , I'm looking for the same integration. Could you give any more details? It is study porpuses.
Ty!

Hi,

Silero model in an iOS environment

Probably using ONNX runtime is the easiest option.

converting the ONNX model to a Core ML model

Typically, such things are what we do commercially for companies for a fee.

[lumpidu]

If you are testing the ONNX runtime on iOS, it would be interesting to know the runtime memory requirements. You can use Instruments within XCode to see, how these grow over time. You can also tune in a limited way how memory is freed and used inside the memory pool.

In my experience, the benefit of CoreML is foremost in reducing the runtime memory requirements, as parts of the code can run on the neural engine. If you want to let your model run e.g. in audio unit extensions, this is very important.

[rootux]

@snakers4 Hey. We would love to get that commercial company service from you.
How can I contact you?