timecat 1.52.0

A NNUE-based chess engine that implements the Negamax algorithm and can be integrated into any project as a library. It features move generation, advanced position evaluation through NNUE, and move searching capabilities.
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

use super::*;

#[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
#[derive(Clone, Debug, BinRead)]
pub struct Layer<
    W: BinRead<Args = ()> + Debug,
    B: BinRead<Args = ()>,
    const NUM_INPUTS: usize,
    const NUM_OUTPUTS: usize,
> {
    biases: Box<MathVec<B, NUM_OUTPUTS>>,
    #[br(count = NUM_OUTPUTS, map = |v: Vec<MathVec<W, NUM_INPUTS>>| SerdeWrapper::from_boxed_value(v.try_into().unwrap()))]
    weights_transpose: Box<SerdeWrapper<[MathVec<W, NUM_INPUTS>; NUM_OUTPUTS]>>,
}

impl<
    W: BinRead<Args = ()> + Debug,
    B: BinRead<Args = ()>,
    const NUM_INPUTS: usize,
    const NUM_OUTPUTS: usize,
> Layer<W, B, NUM_INPUTS, NUM_OUTPUTS>
{
    #[inline]
    pub fn get_weights_transpose(&self) -> &[MathVec<W, NUM_INPUTS>; NUM_OUTPUTS] {
        &self.weights_transpose
    }

    #[inline]
    pub fn get_biases(&self) -> &MathVec<B, NUM_OUTPUTS> {
        &self.biases
    }
}

impl<
    W: BinRead<Args = ()> + Clone + Debug,
    B: BinRead<Args = ()> + Clone + AddAssign + From<W> + Mul + Sum<<B as Mul>::Output>,
    const NUM_INPUTS: usize,
    const NUM_OUTPUTS: usize,
> Layer<W, B, NUM_INPUTS, NUM_OUTPUTS>
{
    pub fn forward(&self, inputs: MathVec<W, NUM_INPUTS>) -> MathVec<B, NUM_OUTPUTS> {
        let mut outputs = self.get_biases().clone();
        outputs
            .iter_mut()
            .zip(self.weights_transpose.iter())
            .for_each(|(o, w)| *o += inputs.dot(w));
        outputs
    }
}