aprender-serve 0.64.0

Pure Rust ML inference engine built from scratch - model serving for GGUF and safetensors
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27


/// IMP-189d: Real-world throughput measurement
#[test]
#[ignore = "Requires running llama.cpp server on port 8082"]
fn test_imp_189d_realworld_throughput() {
    // Simulate throughput measurements (tok/s)
    let throughput = vec![
        143.0, 145.0, 141.0, 144.0, 142.0, 146.0, 140.0, 143.5, 144.5, 141.5,
    ];

    let result = ThroughputResult::from_samples(&throughput);

    println!("\nIMP-189d: Real-World Throughput:");
    println!("  Mean: {:.2} tok/s", result.mean_toks);
    println!("  StdDev: {:.2}", result.std_dev);
    println!("  Variance: {:.2}", result.variance);
    println!("  CV: {:.4} ({:.1}%)", result.cv, result.cv * 100.0);
    println!("  Stable (5%): {}", result.is_stable(0.05));
    println!(
        "  QA-036: {}",
        if result.meets_qa036 { "PASS" } else { "FAIL" }
    );
}

// ================================================================================
// IMP-190: Benchmark Versioning (QA-037)
// Benchmark results versioned and reproducible