Broadcast-multiply a vector and a matrix, GPU-aware

As add_vec_mat_gpu, but multiplying rather than adding.

Usage

mult_vec_mat_gpu(vec, mat, backend = c("auto", "cpu", "gpu"))

Arguments

vec

a numeric vector (or torch tensor).

mat

a matrix (or torch tensor).

backend

one of "auto" (default), "cpu" or "gpu".

Value

mat * vec, broadcast as per add_vec_mat_gpu.

Examples

mult_vec_mat_gpu(1:3, matrix(1, 3, 4), backend = "cpu")
#>      [,1] [,2] [,3] [,4]
#> [1,]    1    1    1    1
#> [2,]    2    2    2    2
#> [3,]    3    3    3    3