HPD-Agent.Providers.Audio.Silero
0.9.6
dotnet add package HPD-Agent.Providers.Audio.Silero --version 0.9.6
NuGet\Install-Package HPD-Agent.Providers.Audio.Silero -Version 0.9.6
<PackageReference Include="HPD-Agent.Providers.Audio.Silero" Version="0.9.6" />
<PackageVersion Include="HPD-Agent.Providers.Audio.Silero" Version="0.9.6" />
<PackageReference Include="HPD-Agent.Providers.Audio.Silero" />
paket add HPD-Agent.Providers.Audio.Silero --version 0.9.6
#r "nuget: HPD-Agent.Providers.Audio.Silero, 0.9.6"
#:package HPD-Agent.Providers.Audio.Silero@0.9.6
#addin nuget:?package=HPD-Agent.Providers.Audio.Silero&version=0.9.6
#tool nuget:?package=HPD-Agent.Providers.Audio.Silero&version=0.9.6
HPD-Agent.Providers.Audio.Silero
Digest-pinned Silero v6.2 voice activity source provider for HPD Agent Audio.
Install
dotnet add package HPD-Agent.Audio
dotnet add package HPD-Agent.Providers.Audio.Silero
Use When
Use this package for local 8 kHz or 16 kHz streaming Voice Activity inference.
The package never downloads a model while constructing a session. Provision the
official MIT-licensed v6.2 ONNX artifact explicitly, verify its SHA-256, and pass
its local path through SileroVadOptions.
The qualified ONNX Runtime 1.23.0 CPU assets cover linux-arm64, linux-x64,
osx-arm64, osx-x64, win-arm64, and win-x64. The provider validates and
prewarms the model before returning a source. Every audio session receives
isolated recurrent state while the immutable inference host is shared.
For repository qualification only, run
dotnet run --file eng/fetch-silero-vad-v6.2.cs. It pins
upstream commit be95df9152c0d7618fa1edfeb296fc3dae32376f and SHA-256
1a153a22f4509e292a94e67d6f9b85e8deb25b4988682b7e174c65279d8788e3.
Configure
using HPD.Agent;
using HPD.Agent.Providers.Audio.Silero;
var agent = await AgentBuilder.Create()
.WithSileroVoiceActivity(
modelPath: "/opt/models/silero_vad.onnx")
.BuildAsync();
The provider defaults to the digest above but accepts an explicit
modelSha256 for an independently verified artifact. Configuration fails
closed when the model is missing, the digest does not match, or the native
inference runtime cannot be loaded.
Silero performs local inference; it does not replace speech-to-text. Pair it with an STT provider through the voice-activity composition contracts when the application needs endpointing or turn detection.
| Product | Versions Compatible and additional computed target framework versions. |
|---|---|
| .NET | net10.0 is compatible. net10.0-android was computed. net10.0-browser was computed. net10.0-ios was computed. net10.0-maccatalyst was computed. net10.0-macos was computed. net10.0-tvos was computed. net10.0-windows was computed. net11.0 is compatible. |
-
net10.0
- HPD-Agent.Audio (>= 0.9.6)
- HPD-Agent.Framework (>= 0.9.6)
- Microsoft.ML.OnnxRuntime (>= 1.31.0)
-
net11.0
- HPD-Agent.Audio (>= 0.9.6)
- HPD-Agent.Framework (>= 0.9.6)
- Microsoft.ML.OnnxRuntime (>= 1.31.0)
NuGet packages
This package is not used by any NuGet packages.
GitHub repositories
This package is not used by any popular GitHub repositories.