synor/sdk/flutter/example/example.dart

import 'dart:io';

import 'package:synor_compute/synor_compute.dart';

/// Example usage of Synor Compute SDK for Flutter/Dart
void main() async {
  // Initialize client with API key
  final client = SynorCompute(
    apiKey: Platform.environment['SYNOR_API_KEY'] ?? 'your-api-key',
    // Optional: customize defaults
    defaultProcessor: ProcessorType.auto,
    defaultPrecision: Precision.fp32,
    defaultPriority: Priority.normal,
  );

  try {
    // Check service health
    final isHealthy = await client.healthCheck();
    print('Service healthy: $isHealthy\n');

    // Example 1: Matrix multiplication
    await matrixMultiplicationExample(client);

    // Example 2: Tensor operations
    await tensorOperationsExample(client);

    // Example 3: LLM inference
    await llmInferenceExample(client);

    // Example 4: Streaming inference
    await streamingInferenceExample(client);

    // Example 5: Pricing and usage
    await pricingExample(client);

    // Example 6: List available models
    await modelRegistryExample(client);

    // Example 7: Training a model
    await trainingExample(client);

    // Example 8: Custom model upload
    await customModelExample(client);
  } finally {
    // Always dispose client to release resources
    client.dispose();
  }
}

/// Matrix multiplication example
Future<void> matrixMultiplicationExample(SynorCompute client) async {
  print('=== Matrix Multiplication ===');

  // Create random matrices
  final a = Tensor.rand([256, 512]);
  final b = Tensor.rand([512, 256]);

  print('A: ${a.shape}');
  print('B: ${b.shape}');

  // Perform multiplication on GPU with FP16 precision
  final result = await client.matmul(
    a,
    b,
    options: MatMulOptions(
      precision: Precision.fp16,
      processor: ProcessorType.gpu,
      priority: Priority.high,
    ),
  );

  if (result.isSuccess) {
    print('Result: ${result.result!.shape}');
    print('Execution time: ${result.executionTimeMs}ms');
    print('Cost: \$${result.cost?.toStringAsFixed(6)}');
    print('Processor: ${result.processor?.value}');
  } else {
    print('Error: ${result.error}');
  }
  print('');
}

/// Local tensor operations example
Future<void> tensorOperationsExample(SynorCompute client) async {
  print('=== Tensor Operations ===');

  // Create tensors
  final x = Tensor.randn([100], mean: 0.0, std: 1.0);
  print('Random normal tensor: mean=${x.mean().toStringAsFixed(4)}, '
      'std=${x.std().toStringAsFixed(4)}');

  // Create identity matrix
  final eye = Tensor.eye(4);
  print('Identity matrix:\n${eye.toNestedList()}');

  // Create linspace
  final linspace = Tensor.linspace(0, 10, 5);
  print('Linspace [0, 10, 5]: ${linspace.toNestedList()}');

  // Reshape operations
  final matrix = Tensor.arange(0, 12).reshape([3, 4]);
  print('Reshaped [0..12] to [3,4]:\n${matrix.toNestedList()}');

  // Transpose
  final transposed = matrix.transpose();
  print('Transposed to ${transposed.shape}');

  // Activations
  final input = Tensor(shape: [5], data: [-2.0, -1.0, 0.0, 1.0, 2.0]);
  print('ReLU of $input: ${input.relu().toNestedList()}');
  print('Sigmoid of $input: ${input.sigmoid().toNestedList()}');

  // Softmax
  final logits = Tensor(shape: [4], data: [1.0, 2.0, 3.0, 4.0]);
  print('Softmax of $logits: ${logits.softmax().toNestedList()}');

  print('');
}

/// LLM inference example
Future<void> llmInferenceExample(SynorCompute client) async {
  print('=== LLM Inference ===');

  final result = await client.inference(
    'llama-3-70b',
    'What is the capital of France? Answer in one word.',
    options: InferenceOptions(
      maxTokens: 10,
      temperature: 0.1,
      processor: ProcessorType.lpu, // Use LPU for LLM
    ),
  );

  if (result.isSuccess) {
    print('Response: ${result.result}');
    print('Time: ${result.executionTimeMs}ms');
  } else {
    print('Error: ${result.error}');
  }
  print('');
}

/// Streaming inference example
Future<void> streamingInferenceExample(SynorCompute client) async {
  print('=== Streaming Inference ===');
  print('Response: ');

  await for (final token in client.inferenceStream(
    'llama-3-70b',
    'Write a short poem about distributed computing.',
    options: InferenceOptions(
      maxTokens: 100,
      temperature: 0.7,
    ),
  )) {
    stdout.write(token);
  }

  print('\n');
}

/// Pricing and usage example
Future<void> pricingExample(SynorCompute client) async {
  print('=== Pricing Information ===');

  final pricing = await client.getPricing();

  print('Current spot prices:');
  for (final p in pricing) {
    print('  ${p.processor.value.toUpperCase().padRight(8)}: '
        '\$${p.pricePerSecond.toStringAsFixed(6)}/sec, '
        '${p.availableUnits} units available, '
        '${p.utilizationPercent.toStringAsFixed(1)}% utilized');
  }

  print('');

  // Get usage stats
  final usage = await client.getUsage();
  print('Usage Statistics:');
  print('  Total jobs: ${usage.totalJobs}');
  print('  Completed: ${usage.completedJobs}');
  print('  Failed: ${usage.failedJobs}');
  print('  Total compute time: ${usage.totalComputeSeconds.toStringAsFixed(2)}s');
  print('  Total cost: \$${usage.totalCost.toStringAsFixed(4)}');
  print('');
}

/// Model registry example - list available models
Future<void> modelRegistryExample(SynorCompute client) async {
  print('=== Model Registry ===');

  // List all available models
  final allModels = await client.listModels();
  print('Total available models: ${allModels.length}');

  // List only LLMs
  final llms = await client.listModels(category: ModelCategory.llm);
  print('\nAvailable LLMs:');
  for (final model in llms.take(5)) {
    print('  ${model.id.padRight(20)} ${model.formattedParameters.padRight(8)} '
        '${model.name}');
  }

  // Search for a specific model
  final searchResults = await client.searchModels('llama');
  print('\nSearch "llama": ${searchResults.length} results');

  // Get specific model info
  final modelInfo = await client.getModel('llama-3-70b');
  print('\nModel details for ${modelInfo.name}:');
  print('  Parameters: ${modelInfo.formattedParameters}');
  print('  Context length: ${modelInfo.contextLength}');
  print('  Format: ${modelInfo.format.value}');
  print('  Recommended processor: ${modelInfo.recommendedProcessor.value}');
  print('  License: ${modelInfo.license}');

  // List embedding models
  final embeddings = await client.listModels(category: ModelCategory.embedding);
  print('\nAvailable embedding models:');
  for (final model in embeddings) {
    print('  ${model.id} - ${model.name}');
  }

  // List image generation models
  final imageGen =
      await client.listModels(category: ModelCategory.imageGeneration);
  print('\nAvailable image generation models:');
  for (final model in imageGen) {
    print('  ${model.id} - ${model.name}');
  }

  print('');
}

/// Training example - train/fine-tune a model
Future<void> trainingExample(SynorCompute client) async {
  print('=== Model Training ===');

  // Example: Fine-tune Llama 3 8B on custom dataset
  print('Fine-tuning llama-3-8b on custom dataset...');

  // Note: In practice, you'd upload your dataset first:
  // final datasetCid = await client.uploadTensor(datasetTensor);

  final result = await client.fineTune(
    baseModel: 'llama-3-8b', // Use model alias
    datasetCid: 'QmYourDatasetCID', // Your uploaded dataset
    outputAlias: 'my-custom-llama', // Optional: alias for trained model
    options: TrainingOptions(
      framework: MlFramework.pytorch,
      epochs: 3,
      batchSize: 8,
      learningRate: 0.00002,
      optimizer: 'adamw',
      hyperparameters: {
        'weight_decay': 0.01,
        'warmup_steps': 100,
        'gradient_accumulation_steps': 4,
      },
      checkpointEvery: 500, // Save checkpoint every 500 steps
      processor: ProcessorType.gpu,
      priority: Priority.high,
    ),
  );

  if (result.isSuccess) {
    final training = result.result!;
    print('Training completed!');
    print('  New model CID: ${training.modelCid}');
    print('  Final loss: ${training.finalLoss.toStringAsFixed(4)}');
    print('  Duration: ${training.durationMs / 1000}s');
    print('  Cost: \$${training.cost.toStringAsFixed(4)}');
    print('  Metrics: ${training.metrics}');

    // Now use your trained model for inference
    print('\nUsing trained model for inference:');
    final inference = await client.inference(
      training.modelCid, // Use the CID of your trained model
      'Hello, how are you?',
      options: InferenceOptions(maxTokens: 50),
    );
    print('Response: ${inference.result}');
  } else {
    print('Training failed: ${result.error}');
  }

  print('');

  // Example: Streaming training progress
  print('Training with streaming progress...');
  await for (final progress in client.trainStream(
    modelCid: 'llama-3-8b',
    datasetCid: 'QmYourDatasetCID',
    options: TrainingOptions(epochs: 1, batchSize: 16),
  )) {
    // Update UI with progress
    stdout.write('\r${progress.progressText} - '
        '${progress.samplesPerSecond} samples/s');
  }
  print('\nTraining complete!');

  print('');
}

/// Custom model upload example
Future<void> customModelExample(SynorCompute client) async {
  print('=== Custom Model Upload ===');

  // Example: Upload a custom ONNX model
  // In practice, you'd read this from a file:
  // final modelBytes = await File('my_model.onnx').readAsBytes();

  // For demonstration, we'll show the API structure
  print('To upload your own Python-trained model:');
  print('''
  1. Train your model in Python:

     import torch
     model = MyModel()
     # ... train model ...
     torch.onnx.export(model, dummy_input, "my_model.onnx")

  2. Upload to Synor Compute:

     final modelBytes = await File('my_model.onnx').readAsBytes();
     final result = await client.uploadModel(
       modelBytes,
       ModelUploadOptions(
         name: 'my-custom-model',
         description: 'My custom trained model',
         category: ModelCategory.custom,
         format: ModelFormat.onnx,
         alias: 'my-model',  // Optional shortcut name
         isPublic: false,    // Keep private
         license: 'Proprietary',
       ),
     );
     print('Uploaded! CID: \${result.cid}');

  3. Use for inference:

     final result = await client.inference(
       result.cid,  // or 'my-model' if you set an alias
       'Your input data',
     );
''');

  // Supported model formats
  print('Supported model formats:');
  for (final format in ModelFormat.values) {
    print('  - ${format.value}');
  }

  // Supported categories
  print('\nSupported model categories:');
  for (final category in ModelCategory.values) {
    print('  - ${category.value}');
  }

  print('');
}