matten/ops/
tensor_ops.rs

1//! Element-wise binary operators for borrowed tensor pairs (RFC-006).
2//!
3//! `*` is element-wise multiplication; matrix multiplication is explicit and
4//! arrives in RFC-010 / M6.
5
6use crate::Tensor;
7use crate::ops::broadcast::apply_binary;
8use std::ops::{Add, Div, Mul, Sub};
9
10impl Add for &Tensor {
11    type Output = Tensor;
12    /// Element-wise addition with NumPy-style broadcasting.
13    ///
14    /// # Panics
15    ///
16    /// Panics with `"matten broadcast error in add: ..."` if the shapes are
17    /// incompatible.
18    ///
19    /// ```
20    /// use matten::Tensor;
21    /// let a = Tensor::new(vec![1.0, 2.0, 3.0, 4.0], &[2, 2]);
22    /// let b = Tensor::ones(&[2, 2]);
23    /// let c = &a + &b;
24    /// assert_eq!(c.as_slice(), &[2.0, 3.0, 4.0, 5.0]);
25    /// ```
26    fn add(self, rhs: &Tensor) -> Tensor {
27        apply_binary(self, rhs, "add", |a, b| a + b)
28    }
29}
30
31impl Sub for &Tensor {
32    type Output = Tensor;
33    /// Element-wise subtraction with broadcasting.
34    ///
35    /// # Panics
36    ///
37    /// Panics on incompatible shapes.
38    ///
39    /// ```
40    /// use matten::Tensor;
41    /// let a = Tensor::new(vec![5.0, 4.0, 3.0, 2.0], &[2, 2]);
42    /// let b = Tensor::ones(&[2, 2]);
43    /// let c = &a - &b;
44    /// assert_eq!(c.as_slice(), &[4.0, 3.0, 2.0, 1.0]);
45    /// ```
46    fn sub(self, rhs: &Tensor) -> Tensor {
47        apply_binary(self, rhs, "sub", |a, b| a - b)
48    }
49}
50
51impl Mul for &Tensor {
52    type Output = Tensor;
53    /// Element-wise multiplication with broadcasting (`*` is **not** matrix
54    /// multiply; use `matmul` for that).
55    ///
56    /// # Panics
57    ///
58    /// Panics on incompatible shapes.
59    ///
60    /// ```
61    /// use matten::Tensor;
62    /// let a = Tensor::new(vec![1.0, 2.0, 3.0, 4.0], &[2, 2]);
63    /// let b = Tensor::full(&[2, 2], 2.0);
64    /// let c = &a * &b;
65    /// assert_eq!(c.as_slice(), &[2.0, 4.0, 6.0, 8.0]);
66    /// ```
67    fn mul(self, rhs: &Tensor) -> Tensor {
68        apply_binary(self, rhs, "mul", |a, b| a * b)
69    }
70}
71
72impl Div for &Tensor {
73    type Output = Tensor;
74    /// Element-wise division with broadcasting. Division by zero follows IEEE 754
75    /// `f64` behavior (yields `inf`, `-inf`, or `NaN`); no error is produced.
76    ///
77    /// # Panics
78    ///
79    /// Panics on incompatible shapes.
80    ///
81    /// ```
82    /// use matten::Tensor;
83    /// let a = Tensor::new(vec![4.0, 9.0], &[2]);
84    /// let b = Tensor::new(vec![2.0, 3.0], &[2]);
85    /// let c = &a / &b;
86    /// assert_eq!(c.as_slice(), &[2.0, 3.0]);
87    /// ```
88    fn div(self, rhs: &Tensor) -> Tensor {
89        apply_binary(self, rhs, "div", |a, b| a / b)
90    }
91}
matten/ops/tensor_ops.rs

matten/ops/
tensor_ops.rs
