Unsqueeze

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} ]