Complete SDK implementation for Flutter and Dart applications: lib/src/types.dart: - Precision, ProcessorType, Priority, JobStatus enums - SynorConfig for client configuration - MatMulOptions, Conv2dOptions, AttentionOptions, InferenceOptions - PricingInfo and UsageStats data classes - SynorException for error handling lib/src/tensor.dart: - Full Tensor class with shape, dtype, and data - Factory constructors: zeros, ones, rand, randn, eye, linspace, arange - Operations: reshape, transpose, flatten - Statistics: sum, mean, std, min, max, argmin, argmax - Element-wise: add, sub, mul, div, scalar ops - Activations: relu, sigmoid, tanh, softmax - JSON serialization with base64-encoded binary data lib/src/job.dart: - JobResult with status, result, timing, and cost - Job class with WebSocket streaming and HTTP polling - JobStatusUpdate for real-time progress tracking - JobBatch for parallel job management lib/src/client.dart: - SynorCompute main client - Operations: matmul, conv2d, attention, elementwise, reduce - LLM inference with streaming support - Tensor upload/download/delete - Job management: submit, cancel, list - Pricing and usage statistics Platform support: Android, iOS, Linux, macOS, Web, Windows
178 lines
4.9 KiB
Dart
178 lines
4.9 KiB
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);
|
|
} 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('');
|
|
}
|