bunsen 0.0.4

burn neural network extension library
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
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164

//! # Tensor Clamping Support

use burn::{
    module::{
        Content,
        ModuleDisplay,
        ModuleDisplayDefault,
    },
    prelude::{
        Backend,
        Tensor,
    },
};
use serde::{
    Deserialize,
    Serialize,
};

/// Configuration for clamping.
#[derive(Default, Debug, Clone, PartialEq, Serialize, Deserialize)]
pub struct ClampConfig {
    /// The minimum value.
    min: Option<f64>,

    /// The maximum value.
    max: Option<f64>,
}

impl ModuleDisplay for ClampConfig {}
impl ModuleDisplayDefault for ClampConfig {
    fn content(
        &self,
        content: Content,
    ) -> Option<Content> {
        Some(content.add("min", &self.min).add("max", &self.max))
    }
}

impl ClampConfig {
    /// Create a new clamp with both minimum and maximum values.
    pub fn min_max(
        min: f64,
        max: f64,
    ) -> Self {
        Self {
            min: Some(min),
            max: Some(max),
        }
    }

    /// Extend the clamp with a minimum value.
    pub fn with_min(
        self,
        min: f64,
    ) -> Self {
        Self {
            min: Some(min),
            ..self
        }
    }

    /// Extend the clamp with a maximum value.
    pub fn with_max(
        self,
        max: f64,
    ) -> Self {
        Self {
            max: Some(max),
            ..self
        }
    }

    /// Apply the clamp.
    pub fn clamp<B: Backend, const D: usize>(
        &self,
        tensor: Tensor<B, D>,
    ) -> Tensor<B, D> {
        match (self.min, self.max) {
            (Some(min), Some(max)) => tensor.clamp(min, max),
            (Some(min), None) => tensor.clamp_min(min),
            (None, Some(max)) => tensor.clamp_max(max),
            (None, None) => tensor,
        }
    }
}

#[cfg(test)]
mod tests {
    use burn::{
        backend::NdArray,
        module::{
            DisplaySettings,
            ModuleDisplay,
        },
        tensor::TensorData,
    };
    use num_traits::clamp;

    use super::*;
    use crate::nn::layers::drop::drop_block::DropBlockOptions;

    #[test]
    fn test_clamp_config_display() {
        let config = ClampConfig::default().with_min(0.5);
        let settings = DisplaySettings::default();

        assert_eq!(
            config.format(settings),
            indoc::indoc! {
                r#"
                ClampConfig {
                  min: 0.5
                  max: None
                }"#
            }
        )
    }

    #[test]
    fn test_config_min_max() {
        let cfg = ClampConfig::min_max(-1.0, 1.0);
        assert_eq!(
            cfg,
            ClampConfig {
                min: Some(-1.0),
                max: Some(1.0),
            }
        );
    }

    #[test]
    fn test_config() {
        type B = NdArray;
        let device = Default::default();

        let cfg = ClampConfig::default();
        assert_eq!(
            cfg,
            ClampConfig {
                min: None,
                max: None,
            }
        );
        let tensor = Tensor::<B, 1>::from_data([-1.0, 0.0, 1.0], &device);
        let tensor = cfg.clamp(tensor);
        tensor
            .to_data()
            .assert_eq(&TensorData::from([-1.0, 0.0, 1.0]), false);

        let cfg = ClampConfig::default().with_min(-0.5).with_max(0.5);
        assert_eq!(
            cfg,
            ClampConfig {
                min: Some(-0.5),
                max: Some(0.5),
            }
        );
        let tensor = Tensor::<B, 1>::from_data([-1.0, 0.0, 1.0], &device);
        let tensor = cfg.clamp(tensor);
        tensor
            .to_data()
            .assert_eq(&TensorData::from([-0.5, 0.0, 0.5]), false);
    }
}