aprender-core 0.30.0

Next-generation machine learning library in pure Rust
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
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141

// =========================================================================
// FALSIFY-TN: validated-tensor-v1.yaml contract (aprender autograd Tensor)
//
// Five-Whys (PMAT-354):
//   Why 1: aprender had no inline FALSIFY-TN-* tests for autograd Tensor
//   Why 2: Tensor tests exist but lack contract-mapped FALSIFY naming
//   Why 3: no mapping from validated-tensor-v1.yaml to inline test names
//   Why 4: aprender predates the inline FALSIFY convention
//   Why 5: Tensor add/matmul was "obviously correct" (basic linear algebra)
//
// References:
//   - provable-contracts/contracts/validated-tensor-v1.yaml
// =========================================================================

use super::Tensor;

/// FALSIFY-TN-001: Tensor addition is element-wise and commutative
#[test]
fn falsify_tn_001_add_commutative() {
    let a = Tensor::new(&[1.0, 2.0, 3.0], &[3]);
    let b = Tensor::new(&[4.0, 5.0, 6.0], &[3]);

    let ab = a.add(&b);
    let ba = b.add(&a);

    for i in 0..3 {
        assert!(
            (ab.data()[i] - ba.data()[i]).abs() < 1e-6,
            "FALSIFIED TN-001: add not commutative at [{i}]: {} != {}",
            ab.data()[i],
            ba.data()[i]
        );
    }
    // Check actual values
    assert!(
        (ab.data()[0] - 5.0).abs() < 1e-6,
        "FALSIFIED TN-001: 1+4={}, expected 5",
        ab.data()[0]
    );
}

/// FALSIFY-TN-002: Tensor shape preserved after add
#[test]
fn falsify_tn_002_add_preserves_shape() {
    let a = Tensor::new(&[1.0; 6], &[2, 3]);
    let b = Tensor::new(&[2.0; 6], &[2, 3]);

    let c = a.add(&b);
    assert_eq!(
        c.shape(),
        &[2, 3],
        "FALSIFIED TN-002: add changed shape to {:?}",
        c.shape()
    );
}

/// FALSIFY-TN-003: Tensor matmul produces correct shape
#[test]
fn falsify_tn_003_matmul_shape() {
    let a = Tensor::new(&[1.0; 6], &[2, 3]); // 2x3
    let b = Tensor::new(&[1.0; 12], &[3, 4]); // 3x4

    let c = a.matmul(&b);
    assert_eq!(
        c.shape(),
        &[2, 4],
        "FALSIFIED TN-003: matmul [2,3]x[3,4] produced shape {:?}, expected [2,4]",
        c.shape()
    );
}

/// FALSIFY-TN-004: Tensor mul_scalar scales all elements
#[test]
fn falsify_tn_004_mul_scalar() {
    let a = Tensor::new(&[1.0, 2.0, 3.0, 4.0], &[4]);
    let b = a.mul_scalar(2.5);

    for (i, &v) in b.data().iter().enumerate() {
        let expected = (i as f32 + 1.0) * 2.5;
        assert!(
            (v - expected).abs() < 1e-6,
            "FALSIFIED TN-004: mul_scalar[{i}]={v}, expected {expected}"
        );
    }
}

mod tensor_proptest_falsify {
    use super::*;
    use proptest::prelude::*;

    // FALSIFY-TN-001-prop: Addition commutativity for random tensors
    proptest! {
        #![proptest_config(ProptestConfig::with_cases(50))]

        #[test]
        fn falsify_tn_001_prop_add_commutative(
            seed in 0..1000u32,
            n in 1..=16usize,
        ) {
            let a_data: Vec<f32> = (0..n)
                .map(|i| ((i as f32 + seed as f32) * 0.37).sin() * 10.0)
                .collect();
            let b_data: Vec<f32> = (0..n)
                .map(|i| ((i as f32 + seed as f32 + 100.0) * 0.53).sin() * 10.0)
                .collect();
            let a = Tensor::new(&a_data, &[n]);
            let b = Tensor::new(&b_data, &[n]);

            let ab = a.add(&b);
            let ba = b.add(&a);

            for i in 0..n {
                prop_assert!(
                    (ab.data()[i] - ba.data()[i]).abs() < 1e-5,
                    "FALSIFIED TN-001-prop: add not commutative at [{}]",
                    i
                );
            }
        }
    }

    // FALSIFY-TN-004-prop: Scalar multiplication for random values
    proptest! {
        #![proptest_config(ProptestConfig::with_cases(50))]

        #[test]
        fn falsify_tn_004_prop_mul_scalar(
            val in -100.0f32..100.0,
            scalar in -10.0f32..10.0,
        ) {
            let a = Tensor::new(&[val], &[1]);
            let b = a.mul_scalar(scalar);
            let expected = val * scalar;
            prop_assert!(
                (b.data()[0] - expected).abs() < 1e-3,
                "FALSIFIED TN-004-prop: {}*{}={}, expected {}",
                val, scalar, b.data()[0], expected
            );
        }
    }
}