Skip to contents

Unsqueeze

Usage

torch_unsqueeze(self, dim)

Arguments

self

(Tensor) the input tensor.

dim

(int) the index at which to insert the singleton dimension

unsqueeze(input, dim) -> Tensor

Returns a new tensor with a dimension of size one inserted at the specified position.

The returned tensor shares the same underlying data with this tensor.

A dim value within the range [-input.dim() - 1, input.dim() + 1) can be used. Negative dim will correspond to unsqueeze applied at dim = dim + input.dim() + 1.

Examples

if (torch_is_installed()) {

x = torch_tensor(c(1, 2, 3, 4))
torch_unsqueeze(x, 1)
torch_unsqueeze(x, 2)
}
#> torch_tensor
#>  1
#>  2
#>  3
#>  4
#> [ CPUFloatType{4,1} ]