burn_tensor/tensor/api/
int.rs

1use crate::check;
2use crate::check::TensorCheck;
3use crate::{
4    backend::Backend, cartesian_grid, Float, Int, Shape, Tensor, TensorData, TensorPrimitive,
5};
6
7use core::ops::Range;
8
9impl<B> Tensor<B, 1, Int>
10where
11    B: Backend,
12{
13    /// Returns a new integer tensor on the specified device.
14    ///
15    /// # Arguments
16    ///
17    /// * `range` - The range of values to generate.
18    /// * `device` - The device to create the tensor on.
19    pub fn arange(range: Range<i64>, device: &B::Device) -> Self {
20        Tensor::new(B::int_arange(range, device))
21    }
22
23    /// Returns a new integer tensor on the specified device.
24    ///
25    /// # Arguments
26    ///
27    /// * `range` - The range of values to generate.
28    /// * `step` - The step between each value.
29    pub fn arange_step(range: Range<i64>, step: usize, device: &B::Device) -> Self {
30        Tensor::new(B::int_arange_step(range, step, device))
31    }
32
33    /// Create a one hot tensor from an index tensor.
34    ///
35    /// # Arguments
36    ///
37    /// * `num_classes` - The number of classes to use in encoding.
38    ///
39    /// # Example
40    ///
41    /// ```rust
42    /// use burn_tensor::backend::Backend;
43    /// use burn_tensor::{Tensor, Int};
44    ///
45    /// fn example<B: Backend>() {
46    ///     let device = B::Device::default();
47    ///     let indices: Tensor<B, 1, Int> = Tensor::from_ints([0, 1, 2, 3], &device);
48    ///     let one_hot = indices.one_hot(4);
49    ///     println!("{}", one_hot.to_data());
50    ///     // [[1, 0, 0, 0], [0, 1, 0, 0], [0, 0, 1, 0], [0, 0, 0, 1]]
51    /// }
52    /// ```
53    pub fn one_hot(self, num_classes: usize) -> Tensor<B, 2, Int> {
54        check!(TensorCheck::one_hot_tensor(self.clone(), num_classes));
55        let [num_samples] = self.dims();
56        let indices = self.unsqueeze_dim(1);
57        let values = indices.ones_like();
58        Tensor::zeros([num_samples, num_classes], &indices.device()).scatter(1, indices, values)
59    }
60}
61
62impl<const D: usize, B> Tensor<B, D, Int>
63where
64    B: Backend,
65{
66    /// Create a tensor from integers (i32), placing it on a given device.
67    ///
68    /// # Example
69    ///
70    /// ```rust
71    /// use burn_tensor::backend::Backend;
72    /// use burn_tensor::{Tensor, Int};
73    ///
74    /// fn example<B: Backend>() {
75    ///     let device = B::Device::default();
76    ///     let _x: Tensor<B, 1, Int> = Tensor::from_ints([1, 2], &device);
77    ///     let _y: Tensor<B, 2, Int> = Tensor::from_ints([[1, 2], [3, 4]], &device);
78    /// }
79    /// ```
80    pub fn from_ints<A: Into<TensorData>>(ints: A, device: &B::Device) -> Self {
81        Self::from_data(ints.into().convert::<i32>(), device)
82    }
83
84    /// Returns a new tensor with the same shape and device as the current tensor and the data
85    /// cast to Float.
86    ///
87    /// # Example
88    ///
89    /// ```rust
90    /// use burn_tensor::backend::Backend;
91    /// use burn_tensor::{Int, Tensor};
92    ///
93    /// fn example<B: Backend>() {
94    ///     let device = Default::default();
95    ///     let int_tensor = Tensor::<B, 1, Int>::arange(0..5, &device);
96    ///     let float_tensor = int_tensor.float();
97    /// }
98    /// ```
99    pub fn float(self) -> Tensor<B, D, Float> {
100        Tensor::new(TensorPrimitive::Float(B::int_into_float(self.primitive)))
101    }
102
103    /// Generates a cartesian grid for the given tensor shape on the specified device.
104    /// The generated tensor is of dimension `D2 = D + 1`, where each element at dimension D contains the cartesian grid coordinates for that element.
105    ///
106    /// # Arguments
107    ///
108    /// * `shape` - The shape specifying the dimensions of the tensor.
109    /// * `device` - The device to create the tensor on.
110    ///
111    /// # Panics
112    ///
113    /// Panics if `D2` is not equal to `D+1`.
114    ///
115    /// # Examples
116    ///
117    /// ```rust
118    ///    use burn_tensor::Int;
119    ///    use burn_tensor::{backend::Backend, Shape, Tensor};
120    ///    fn example<B: Backend>() {
121    ///        let device = Default::default();
122    ///        let result: Tensor<B, 3, _> = Tensor::<B, 2, Int>::cartesian_grid([2, 3], &device);
123    ///        println!("{}", result);
124    ///    }
125    /// ```
126    pub fn cartesian_grid<S: Into<Shape>, const D2: usize>(
127        shape: S,
128        device: &B::Device,
129    ) -> Tensor<B, D2, Int> {
130        cartesian_grid::<B, S, D, D2>(shape, device)
131    }
132}
burn_tensor/tensor/api/int.rs

burn_tensor/tensor/api/
int.rs
