xnn 0.2.0

A lightweight ML framework with GPU-first architecture
Documentation 
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

//! Tests for `Tensor::constant` operation.

use std::f32::consts::PI;

use approx::assert_relative_eq;
use xnn::{Context, Tensor};

#[test]
fn test_constant_broadcast_f32() {
    let ctx = Context::try_default().unwrap();
    let t = Tensor::<f32>::constant(&ctx, &[4], &[42.0]).unwrap();
    assert_eq!(t.dimensions(), &[4]);
    for val in &t.to_vec().unwrap() {
        assert_relative_eq!(*val, 42.0, epsilon = 1e-4);
    }
}

#[test]
fn test_constant_broadcast_i32() {
    let ctx = Context::try_default().unwrap();
    let t = Tensor::<i32>::constant(&ctx, &[4], &[42]).unwrap();
    assert_eq!(t.dimensions(), &[4]);
    assert_eq!(t.to_vec().unwrap(), vec![42, 42, 42, 42]);
}

#[test]
fn test_constant_broadcast_u32() {
    let ctx = Context::try_default().unwrap();
    let t = Tensor::<u32>::constant(&ctx, &[4], &[42]).unwrap();
    assert_eq!(t.dimensions(), &[4]);
    assert_eq!(t.to_vec().unwrap(), vec![42, 42, 42, 42]);
}

#[test]
fn test_constant_broadcast_2d() {
    let ctx = Context::try_default().unwrap();
    let t = Tensor::<f32>::constant(&ctx, &[2, 3], &[1.5]).unwrap();
    assert_eq!(t.dimensions(), &[2, 3]);
    for val in &t.to_vec().unwrap() {
        assert_relative_eq!(*val, 1.5, epsilon = 1e-4);
    }
}

#[test]
fn test_constant_exact_fill() {
    let ctx = Context::try_default().unwrap();
    let data = vec![1.0, 2.0, 3.0, 4.0];
    let t = Tensor::<f32>::constant(&ctx, &[4], &data).unwrap();
    assert_eq!(t.dimensions(), &[4]);
    for (a, b) in t.to_vec().unwrap().iter().zip(data.iter()) {
        assert_relative_eq!(a, b, epsilon = 1e-4);
    }
}

#[test]
fn test_constant_exact_fill_2d() {
    let ctx = Context::try_default().unwrap();
    let data = vec![1.0, 2.0, 3.0, 4.0, 5.0, 6.0];
    let t = Tensor::<f32>::constant(&ctx, &[2, 3], &data).unwrap();
    assert_eq!(t.dimensions(), &[2, 3]);
    for (a, b) in t.to_vec().unwrap().iter().zip(data.iter()) {
        assert_relative_eq!(a, b, epsilon = 1e-4);
    }
}

#[test]
fn test_constant_non_aligned() {
    let ctx = Context::try_default().unwrap();
    let t = Tensor::<f32>::constant(&ctx, &[5], &[PI]).unwrap();
    assert_eq!(t.dimensions(), &[5]);
    for val in &t.to_vec().unwrap() {
        assert_relative_eq!(*val, PI, epsilon = 1e-4);
    }
}

#[test]
fn test_constant_large() {
    let ctx = Context::try_default().unwrap();
    let len = 4096 * 4096;
    let t = Tensor::<f32>::constant(&ctx, &[len], &[PI]).unwrap();
    assert_eq!(t.dimensions(), &[len]);
    for val in &t.to_vec().unwrap() {
        assert_relative_eq!(*val, PI, epsilon = 1e-4);
    }
}

#[test]
fn test_constant_scalar() {
    let ctx = Context::try_default().unwrap();
    let t = Tensor::<f32>::constant(&ctx, &[], &[42.0]).unwrap();
    assert_eq!(t.dimensions(), &[] as &[usize]);
    assert_relative_eq!(t.to_vec().unwrap()[0], 42.0, epsilon = 1e-4);
}

#[test]
fn test_constant_empty_value_error() {
    let ctx = Context::try_default().unwrap();
    let result = Tensor::<f32>::constant(&ctx, &[4], &[]);
    assert!(result.is_err());
}

#[test]
fn test_constant_zero_dimension_error() {
    let ctx = Context::try_default().unwrap();
    let result = Tensor::<f32>::constant(&ctx, &[0], &[1.0]);
    assert!(result.is_err());
}

#[test]
fn test_constant_length_mismatch_error() {
    let ctx = Context::try_default().unwrap();
    let result = Tensor::<f32>::constant(&ctx, &[4], &[1.0, 2.0, 3.0]);
    assert!(result.is_err());
}