Function dfdx::tensor_ops::vecmat_mul_transpose

pub fn vecmat_mul_transpose<const K: usize, const N: usize, TAPE: Tape>(
    lhs: Tensor1D<K, TAPE>,
    rhs_t: &Tensor2D<N, K, NoneTape>
) -> Tensor1D<N, TAPE>

vector * matrix multiplication where rhs is transposed. y * transpose(rhs)

Arguments

• lhs - a 1d tensor representing a 1xK matrix
• rhs_t - a 2d tensor representing a NxK matrix

Generics

• K: number of columns of lhs and number of rows of rhs.
• N: Number of columns of rhs.

Returns a 1d tensor representing an 1xO matrix.

Examples

let x: Tensor1D<2> = Tensor1D::zeros();
let y: Tensor2D<4, 2> = Tensor2D::zeros();
let result: Tensor1D<4> = vecmat_mul_transpose(x, &y);