numr 0.5.1

High-performance numerical computing with multi-backend GPU acceleration (CPU/CUDA/WebGPU)
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

//! Scalar fallbacks for convolution operations

use crate::ops::conv_common::{Conv1dParams, Conv2dParams};

/// Scalar conv1d for f32
#[inline]
pub unsafe fn conv1d_scalar_f32(
    input: *const f32,
    weight: *const f32,
    bias: Option<*const f32>,
    output: *mut f32,
    params: Conv1dParams,
) {
    crate::runtime::cpu::kernels::conv::conv1d_kernel(input, weight, bias, output, params);
}

/// Scalar conv1d for f64
#[inline]
pub unsafe fn conv1d_scalar_f64(
    input: *const f64,
    weight: *const f64,
    bias: Option<*const f64>,
    output: *mut f64,
    params: Conv1dParams,
) {
    crate::runtime::cpu::kernels::conv::conv1d_kernel(input, weight, bias, output, params);
}

/// Scalar conv2d for f32
#[inline]
pub unsafe fn conv2d_scalar_f32(
    input: *const f32,
    weight: *const f32,
    bias: Option<*const f32>,
    output: *mut f32,
    params: Conv2dParams,
) {
    crate::runtime::cpu::kernels::conv::conv2d_kernel(input, weight, bias, output, params);
}

/// Scalar conv2d for f64
#[inline]
pub unsafe fn conv2d_scalar_f64(
    input: *const f64,
    weight: *const f64,
    bias: Option<*const f64>,
    output: *mut f64,
    params: Conv2dParams,
) {
    crate::runtime::cpu::kernels::conv::conv2d_kernel(input, weight, bias, output, params);
}

/// Scalar depthwise conv2d for f32
#[inline]
pub unsafe fn depthwise_conv2d_scalar_f32(
    input: *const f32,
    weight: *const f32,
    bias: Option<*const f32>,
    output: *mut f32,
    params: Conv2dParams,
) {
    crate::runtime::cpu::kernels::conv::depthwise_conv2d_kernel(
        input, weight, bias, output, params,
    );
}

/// Scalar depthwise conv2d for f64
#[inline]
pub unsafe fn depthwise_conv2d_scalar_f64(
    input: *const f64,
    weight: *const f64,
    bias: Option<*const f64>,
    output: *mut f64,
    params: Conv2dParams,
) {
    crate::runtime::cpu::kernels::conv::depthwise_conv2d_kernel(
        input, weight, bias, output, params,
    );
}