concision_core/activate/impls/
impl_nonlinear.rs

1/*
2    Appellation: sigmoid <mod>
3    Contrib: FL03 <jo3mccain@icloud.com>
4*/
5use crate::activate::{ReLU, Sigmoid, Softmax, Tanh};
6
7use ndarray::{Array, ArrayBase, Data, Dimension, ScalarOperand};
8use num_traits::Float;
9
10impl<A, S, D> ReLU for ArrayBase<S, D>
11where
12    A: Copy + core::cmp::PartialOrd + num::Zero,
13    S: Data<Elem = A>,
14    D: Dimension,
15{
16    type Output = Array<A, D>;
17
18    fn relu(&self) -> Self::Output {
19        self.map(|&i| if i > A::zero() { i } else { A::zero() })
20    }
21}
22
23impl<A, S, D> Sigmoid for ArrayBase<S, D>
24where
25    A: Float + 'static,
26    S: Data<Elem = A>,
27    D: Dimension,
28{
29    type Output = Array<A, D>;
30
31    fn sigmoid(&self) -> Self::Output {
32        let dim = self.dim();
33        let ones = Array::<A, D>::ones(dim);
34
35        (ones + self.map(|&i| i.neg().exp())).recip()
36    }
37}
38
39impl<A, S, D> Softmax for ArrayBase<S, D>
40where
41    A: Float + ScalarOperand,
42    S: Data<Elem = A>,
43    D: Dimension,
44{
45    type Output = Array<A, D>;
46
47    fn softmax(&self) -> Self::Output {
48        let e = self.exp();
49        &e / e.sum()
50    }
51}
52
53impl<A, S, D> Tanh for ArrayBase<S, D>
54where
55    A: Float + 'static,
56    S: Data<Elem = A>,
57    D: Dimension,
58{
59    type Output = Array<A, D>;
60
61    fn tanh(&self) -> Self::Output {
62        self.map(|i| i.tanh())
63    }
64}
concision_core/activate/impls/impl_nonlinear.rs

concision_core/activate/impls/
impl_nonlinear.rs
