torch.Tensor.copy_#
- Tensor.copy_(src, non_blocking=False) Tensor#
Copies the elements from
srcintoselftensor and returnsself.The
srctensor must be broadcastable with theselftensor. It may be of a different data type or reside on a different device.- Parameters:
Note
When
non_blockingisTrueand the copy is issued on a non-default CUDA stream, the caller is responsible for proper cross-stream synchronization. See CUDA streams for the required pattern.