CoreML export of Silero VAD v6.2.1 for Apple platforms. The package includes a drop-in 32 ms streaming model and a faster 256 ms aggregate variant.
Benchmarked on an Apple Silicon Mac using a 50-file VAD set with 2,445.9 seconds of audio, onset 0.25, offset 0.20, min speech 0.25s, min silence 0.10s.
The 32 ms variant preserves the existing streaming interface and is the safest drop-in replacement. The 256 ms variant is substantially faster and more recall-oriented, but uses coarser probability frames.
1import CoreML
2
3let config = MLModelConfiguration()
4config.computeUnits = .all
5let modelURL = Bundle.main.url(forResource: "silero_vad", withExtension: "mlmodelc")!
6let model = try MLModel(contentsOf: modelURL, configuration: config)