Skip to contents

Performs logical AND along the specified dimensions.

Usage

nvl_reduce_all(operand, dims, drop = TRUE)

Arguments

operand

(tensorish)
Tensorish value of boolean data type.

dims

(integer())
Dimensions to reduce over.

drop

(logical(1))
Whether to drop the reduced dimensions from the output shape. If TRUE, the reduced dimensions are removed. If FALSE, the reduced dimensions are set to 1.

Value

tensorish
Boolean tensor. Never ambiguous. When drop = TRUE, the shape is that of operand with dims removed. When drop = FALSE, the shape is that of operand with dims set to 1.

Implemented Rules

  • stablehlo

  • backward

StableHLO

Lowers to stablehlo::hlo_reduce() with stablehlo::hlo_and() as the reducer.

See also

Examples

jit_eval({
  x <- nv_tensor(matrix(c(TRUE, FALSE, TRUE, TRUE), nrow = 2))
  nvl_reduce_all(x, dims = 1L)
})
#> AnvilTensor
#>  0
#>  1
#> [ CPUi1{2} ]