Cara Deploy Gemma 4 di Android & iOS (Panduan AI Mobile)

// build.gradle.kts
dependencies {
    implementation("com.google.ai.edge:ai-edge-sdk:0.3.0")
}

// Di Activity atau ViewModel
import com.google.ai.edge.InferenceSession
import com.google.ai.edge.ModelConfig

class GemmaViewModel : ViewModel() {
    private var session: InferenceSession? = null

    fun initModel(context: Context) {
        val config = ModelConfig.builder()
            .setModelPath("gemma-4-e2b-it.task")
            .setMaxTokens(1024)
            .setTemperature(0.7f)
            .build()

        session = InferenceSession.create(context, config)
    }

    fun generateResponse(prompt: String): String {
        return session?.generateResponse(prompt) ?: "Model belum dimuat"
    }
}

// Cek apakah AICore tersedia
val aiCoreAvailable = AICore.isAvailable(context)

if (aiCoreAvailable) {
    // AICore menangani manajemen dan optimasi model
    val session = AICore.createSession(
        model = "gemma-4-e2b-it",
        options = AICore.Options.builder()
            .setAccelerator(AICore.Accelerator.GPU)
            .build()
    )

    val response = session.generate("Jelaskan fotosintesis secara sederhana")
}

dependencies {
    implementation("com.google.mediapipe:tasks-genai:0.10.20")
}

// Inisialisasi LLM
val options = LlmInference.LlmInferenceOptions.builder()
    .setModelPath("/data/local/tmp/gemma-4-e2b-it.bin")
    .setMaxTokens(1024)
    .setTopK(40)
    .setTemperature(0.7f)
    .setRandomSeed(42)
    .build()

val llmInference = LlmInference.createFromOptions(context, options)

// Generate teks
val response = llmInference.generateResponse("Apa itu machine learning?")

// Stream respons
llmInference.generateResponseAsync(prompt) { partialResult, done ->
    // Update UI dengan setiap token
    textView.append(partialResult)
}

import LiteRT

class GemmaModel {
    private var interpreter: Interpreter?

    func loadModel() throws {
        guard let modelPath = Bundle.main.path(
            forResource: "gemma-4-e2b-it",
            ofType: "tflite"
        ) else {
            throw GemmaError.modelNotFound
        }

        var options = Interpreter.Options()
        options.threadCount = 4

        // Gunakan GPU delegate untuk akselerasi
        let gpuDelegate = MetalDelegate()
        interpreter = try Interpreter(
            modelPath: modelPath,
            options: options,
            delegates: [gpuDelegate]
        )
    }

    func generate(prompt: String) throws -> String {
        // Tokenisasi input
        let tokens = tokenize(prompt)

        // Jalankan inferensi
        try interpreter?.allocateTensors()
        try interpreter?.copy(tokens, toInputAt: 0)
        try interpreter?.invoke()

        // Decode output
        let output = try interpreter?.output(at: 0)
        return decode(output)
    }
}

import MediaPipeTasksGenAI

let options = LlmInference.Options()
options.modelPath = Bundle.main.path(
    forResource: "gemma-4-e2b-it",
    ofType: "bin"
)!
options.maxTokens = 1024
options.temperature = 0.7

let llm = try LlmInference(options: options)
let response = try llm.generateResponse(inputText: "Halo!")

// Praktik baik: release model saat tidak dipakai
override fun onPause() {
    super.onPause()
    session?.close()
}

override fun onResume() {
    super.onResume()
    if (session == null) initModel()
}