Function dfdx::tensor_ops::broadcast_outer_add

pub fn broadcast_outer_add<Lhs, Rhs, const M: usize>(lhs: Lhs, rhs: &Rhs) -> Lhs where
    Lhs: Tensor<Array = [Rhs::Array; M], Dtype = f32>,
    Rhs: 'static + Tensor<Dtype = f32, TapeHolder = NoTape>,
    Lhs::Device: Device<Lhs::Array> + Device<Rhs::Array>,
    Cpu: ZipMapElements<Rhs::Array, Rhs::Array>, 

Broadcast the first dimension of Rhs M times, so its the same size as Lhs.

E.g If Lhs has dimension (2, 3), then Rhs could be dimension (3, ).

Examples

let a = Tensor2D::new([[1.0, 2.0, 3.0], [4.0, 5.0, 6.0]]);
let b = Tensor1D::new([-1.0, 0.0, 1.0]);
let r = broadcast_outer_add(a, &b);
assert_eq!(r.data(), &[[0.0, 2.0, 4.0], [3.0, 5.0, 7.0]]);