rotta_rs 0.0.5

a Deep Learning library with rust language
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

use crate::rotta_rs_module::{ arrayy::Arrayy, BackwardLabel, NodeType, Tensor };

// matmul
pub fn matmul(a: &Tensor, b: &Tensor) -> Tensor {
    let tensor_a = a.node.lock().unwrap();
    let tensor_b = b.node.lock().unwrap();

    // let output = if tensor_a.multithread {
    // tensor_a.value.par_matmul(&tensor_b.value)
    // } else {
    let output = tensor_a.value.matmul(&tensor_b.value);
    // };

    let tensor = Tensor::from_arrayy(output);

    tensor.update_parent(vec![a.node.clone(), b.node.clone()]);
    tensor.node.lock().as_mut().unwrap().label = Some(
        BackwardLabel::Matmul(a.node.clone(), b.node.clone())
    );

    tensor
}

pub fn d_matmul(a: &NodeType, b: &NodeType, grad: &Arrayy) {
    let mut a = a.lock().unwrap();
    let mut b = b.lock().unwrap();

    // da = grad * b^t
    if a.requires_grad {
        let d_a = grad.matmul(&b.value.t());
        a.add_grad(d_a);
    }

    // // db = a * grad
    if b.requires_grad {
        let d_b = a.value.t().matmul(grad);
        b.add_grad(d_b);
    }
}