Created
December 9, 2025 10:21
-
-
Save monday8am/aeb12c48548d6589106e2a78f57edd30 to your computer and use it in GitHub Desktop.
Simplified code for instantiate / use the Conversation API
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| override suspend fun initialize( | |
| modelConfig: ModelConfiguration, | |
| modelPath: String, | |
| ): Result<Unit> = | |
| withContext(dispatcher) { | |
| val engineConfig = | |
| EngineConfig( | |
| modelPath = modelPath, | |
| backend = if (modelConfig.hardwareAcceleration == HardwareBackend.GPU_SUPPORTED) | |
| Backend.GPU else Backend.CPU, | |
| visionBackend = null, // Text-only inference | |
| audioBackend = null, // Text-only inference | |
| maxNumTokens = modelConfig.contextLength, | |
| ) | |
| val engine = Engine(engineConfig) | |
| engine.initialize() | |
| // Configure conversation with tools for native tool calling | |
| val conversationConfig = | |
| ConversationConfig( | |
| systemMessage = Message.of("You are Qwen, created by Alibaba Cloud. You are a helpful assistant."), | |
| tools = tools, // Native LiteRT-LM tools with @Tool annotations | |
| samplerConfig = | |
| SamplerConfig( | |
| topK = modelConfig.defaultTopK, | |
| topP = modelConfig.defaultTopP.toDouble(), | |
| temperature = modelConfig.defaultTemperature.toDouble(), | |
| ), | |
| ) | |
| val conversation = engine.createConversation(conversationConfig) | |
| } | |
| override fun promptStreaming(prompt: String): Flow<String> { | |
| val userMessage = Message.of(prompt) | |
| var startTime = 0L | |
| return instance.conversation | |
| .sendMessageAsync(userMessage) | |
| .map { message -> | |
| message.contents.filterIsInstance<Content.Text>().joinToString("") { it.text } | |
| }.filter { it.isNotEmpty() } | |
| .onStart { | |
| startTime = System.currentTimeMillis() | |
| Logger.i("LocalInferenceEngine") { "Streaming inference started." } | |
| }.onCompletion { | |
| val duration = System.currentTimeMillis() - startTime | |
| Logger.i("LocalInferenceEngine") { "✅ Streaming inference complete: ${duration}ms" } | |
| }.flowOn(dispatcher) | |
| } |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment