LlamaEngineBackendHooks extension

Low-level hooks that TextToSpeechEngine, DecisionEngine and backend integrations use on a LlamaEngine.

Applications should use those engines instead. These are extension members, not instance members, so a subclass of LlamaEngine cannot override them: fake a backend that implements the matching Backend* interface instead.

on

Properties

backendDecisionCapabilities → Future<BackendDecisionCapabilities>

Available on LlamaEngine, provided by the LlamaEngineBackendHooks extension

Returns decision-model support for the loaded model.
no setter
backendTextToSpeechCapabilities → Future<BackendTextToSpeechCapabilities>

Available on LlamaEngine, provided by the LlamaEngineBackendHooks extension

Returns backend-native text-to-speech capabilities for the loaded model.
no setter
contextHandle → int?

Available on LlamaEngine, provided by the LlamaEngineBackendHooks extension

Internal context handle.
no setter
modelHandle → int?

Available on LlamaEngine, provided by the LlamaEngineBackendHooks extension

Internal model handle.
no setter

Methods

cancelTextToSpeechBackend() → void

Available on LlamaEngine, provided by the LlamaEngineBackendHooks extension

Cancels backend-native synthesis started by TextToSpeechEngine.
freeDecisionHeadBackend(int headHandle) → Future<void>

Available on LlamaEngine, provided by the LlamaEngineBackendHooks extension

Frees the decision head headHandle.
loadDecisionHeadBackend(String headPath, {String? configPath}) → Future<BackendDecisionHeadInfo>

Available on LlamaEngine, provided by the LlamaEngineBackendHooks extension

Loads the decision head at headPath for the loaded model.
runDecisionBackend(int headHandle, List<BackendDecisionSequence> sequences) → Future<List<BackendDecisionOutput>>

Available on LlamaEngine, provided by the LlamaEngineBackendHooks extension

Runs sequences through the decision head headHandle.
synthesizeTextToSpeechBackend(BackendTextToSpeechRequest request, {void onProgress(BackendTextToSpeechProgress progress)?}) → Future<BackendTextToSpeechResult>

Available on LlamaEngine, provided by the LlamaEngineBackendHooks extension

Runs backend-native text-to-speech for TextToSpeechEngine.