Rate this Page
★★★★★

torch.Tensor.resize_#

Tensor.resize_(*sizes,memory_format=torch.contiguous_format)→Tensor#

Resizesself tensor to the specified size. If the number of elements islarger than the current storage size, then the underlying storage is resizedto fit the new number of elements. If the number of elements is smaller, theunderlying storage is not changed. Existing elements are preserved but any newmemory is uninitialized.

Warning

This is a low-level method. The storage is reinterpreted as C-contiguous,ignoring the current strides (unless the target size equals the currentsize, in which case the tensor is left unchanged). For most purposes, youwill instead want to useview(), which checks forcontiguity, orreshape(), which copies data if needed. Tochange the size in-place with custom strides, seeset_().

Note

Iftorch.use_deterministic_algorithms() andtorch.utils.deterministic.fill_uninitialized_memory are both set toTrue, new elements are initialized to prevent nondeterministic behaviorfrom using the result as an input to an operation. Floating point andcomplex values are set to NaN, and integer values are set to the maximumvalue.

Parameters

  • sizes (torch.Size orint...) – the desired size

  • memory_format (torch.memory_format, optional) – the desired memory format ofTensor. Default:torch.contiguous_format. Note that memory format ofself is going to be unaffected ifself.size() matchessizes.

Example:

>>>x=torch.tensor([[1,2],[3,4],[5,6]])>>>x.resize_(2,2)tensor([[ 1,  2],        [ 3,  4]])