VoxaTrace/com.musicmuni.voxatrace.ai/ModelLoader

ModelLoader

actual object ModelLoader

Android implementation of ModelLoader.

Loads models from downloaded files via ModelDownloader. After VT.initialize(), models are downloaded in background and loaded on demand.

expect object ModelLoader

Platform-specific model loader.

Provides access to bundled ONNX models for VAD and pitch detection. Method names align with public VADBackend enum values.

actual object ModelLoader

iOS implementation of ModelLoader.

Loads models from downloaded files via ModelDownloader. After VT.initialize(), models are downloaded in background and loaded on demand.

actual object ModelLoader

JVM implementation of ModelLoader.

Mirrors the Android implementation, but JVM-native: there is no Android Context, so the backing models directory is resolved by ModelDownloader under <user.home>/.voxatrace/models.

has* checks return false (rather than throwing) when a model file is absent, so callers can degrade gracefully. load* delegates to ModelDownloader.awaitModel, which downloads on demand if needed.

Functions

actual fun configure(context: Any?)

expect fun configure(context: Any?)

Configure the model loader. Must be called before loading any models.

actual fun configure(context: Any?)

actual fun configure(context: Any?)

hasSingingRealtimeVAD

actual fun hasSingingRealtimeVAD(): Boolean

expect fun hasSingingRealtimeVAD(): Boolean

Check if Singing Realtime VAD model is available.

actual fun hasSingingRealtimeVAD(): Boolean

actual fun hasSingingRealtimeVAD(): Boolean

actual fun hasSpeechVAD(): Boolean

expect fun hasSpeechVAD(): Boolean

Check if Speech VAD model is available.

actual fun hasSpeechVAD(): Boolean

actual fun hasSpeechVAD(): Boolean

actual fun hasSwiftF0(): Boolean

expect fun hasSwiftF0(): Boolean

Check if SwiftF0 model is available.

actual fun hasSwiftF0(): Boolean

actual fun hasSwiftF0(): Boolean

loadSingingRealtimeVAD

actual fun loadSingingRealtimeVAD(): ByteArray

expect fun loadSingingRealtimeVAD(): ByteArray

Load the Singing Realtime VAD model (SwiftF0). Use with VADBackend.SINGING_REALTIME.

actual fun loadSingingRealtimeVAD(): ByteArray

actual fun loadSingingRealtimeVAD(): ByteArray

actual fun loadSpeechVAD(): ByteArray

expect fun loadSpeechVAD(): ByteArray

Load the Speech VAD model (Silero). Use with VADBackend.SPEECH.

actual fun loadSpeechVAD(): ByteArray

actual fun loadSpeechVAD(): ByteArray

actual fun loadSwiftF0(): ByteArray

expect fun loadSwiftF0(): ByteArray

Load the SwiftF0 pitch detection model. Use with PitchDetection for pitch detection.

actual fun loadSwiftF0(): ByteArray

actual fun loadSwiftF0(): ByteArray