runn 0.1.1

Runn is a feature-rich, easy-to-use library for building, training, and evaluating feed-forward neural networks in Rust
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
115

use crate::{common::matrix::DMat, error::NetworkError, regression::RegressionEvaluator, MetricEvaluator, Metrics};
use serde::{Deserialize, Serialize};
use typetag;

use super::{LossFunction, LossFunctionClone};

// MeanSquaredErrorLoss is a commonly used loss function for regression tasks,
// measuring the average squared difference between the predicted values and the true target values.
// It penalizes larger errors more heavily than smaller ones, making it sensitive to outliers.
// The loss is computed as the mean of the squared differences between the predicted and target values.
// Forward pass:
// loss = (1 / N) * Σ((predicted - target) ** 2)
// where N is the number of samples, and the summation is over all elements in the matrices.
// Backward pass:
// gradient = (2 / N) * (predicted - target)
// The gradient represents the scaled difference between the predicted and target values,
// which can be used to update the model parameters during optimization.
#[derive(Serialize, Deserialize, Clone)]
struct MeanSquaredErrorLoss;

/// MeanSquared is a builder for Mean Squared Error Loss which is a commonly used loss function for regression tasks,
/// measuring the average squared difference between the predicted values and the true target values.
/// It penalizes larger errors more heavily than smaller ones, making it sensitive to outliers.
/// The loss is computed as the mean of the squared differences between the predicted and target values.
///
/// Forward pass:
/// loss = (1 / N) * Σ((predicted - target) ** 2)
/// where N is the number of samples, and the summation is over all elements in the matrices.
///
/// Backward pass:
/// gradient = (2 / N) * (predicted - target)
///
/// The gradient represents the scaled difference between the predicted and target values,
/// which can be used to update the model parameters during optimization.
pub struct MeanSquared;

impl MeanSquared {
    // Creates a new builder for CrossEntropyLoss
    fn new() -> Self {
        Self {}
    }

    pub fn build(self) -> Result<Box<dyn LossFunction>, NetworkError> {
        Ok(Box::new(MeanSquaredErrorLoss {}))
    }
}

impl Default for MeanSquared {
    /// Creates a new builder for CrossEntropyLoss
    fn default() -> Self {
        Self::new()
    }
}

impl LossFunctionClone for MeanSquaredErrorLoss {
    fn clone_box(&self) -> Box<dyn LossFunction> {
        Box::new(self.clone())
    }
}

#[typetag::serde]
impl LossFunction for MeanSquaredErrorLoss {
    fn forward(&self, predicted: &DMat, target: &DMat) -> f32 {
        let mut loss = 0.0;
        let rows = predicted.rows();

        for i in 0..rows {
            let diff = predicted.at(i, 0) - target.at(i, 0);
            loss += diff * diff;
        }

        loss / rows as f32
    }

    fn backward(&self, predicted: &DMat, target: &DMat) -> DMat {
        let (rows, cols) = (predicted.rows(), predicted.cols());
        let mut gradient = DMat::zeros(rows, cols);

        gradient.apply_with_indices(|i, j, v| {
            let diff = predicted.at(i, j) - target.at(i, j);
            *v = 2.0 * diff / rows as f32;
        });

        gradient
    }

    fn calculate_metrics(&self, targets: &DMat, predictions: &DMat) -> Metrics {
        RegressionEvaluator.evaluate(targets, predictions)
    }
}

#[cfg(test)]
mod tests {
    use super::*;
    use crate::{common::matrix::DMat, util};

    #[test]
    fn test_forward() {
        let loss = MeanSquared.build().unwrap();
        let predicted = DMat::new(2, 1, &[0.9, 0.2]);
        let target = DMat::new(2, 1, &[1.0, 0.0]);
        let result = loss.forward(&predicted, &target);
        assert!((result - 0.025).abs() < 1e-6);
    }

    #[test]
    fn test_backward() {
        let loss = MeanSquared.build().unwrap();
        let predicted = DMat::new(2, 1, &[0.9, 0.2]);
        let target = DMat::new(2, 1, &[1.0, 0.0]);
        let gradient = loss.backward(&predicted, &target);
        let expected_gradient = DMat::new(2, 1, &[-0.1, 0.2]);
        assert!(util::equal_approx(&gradient, &expected_gradient, 1e-6));
    }
}