Rate this Page

torch.Tensor.share_memory_#

Tensor.share_memory_()[source]#

Moves the underlying storage to shared memory.

This is a no-op if the underlying storage is already in shared memory and for CUDA tensors. Tensors in shared memory cannot be resized.

See torch.UntypedStorage.share_memory_() for more details.