torch_unsqueeze: Unsqueeze

In torch: Tensors and Neural Networks with 'GPU' Acceleration

Unsqueeze

Description

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 documentation built on May 29, 2024, 9:54 a.m.