torch.Tensor中的clone()方法

clone() → Tensor

Returns a copy of the self tensor. The copy has the same size and data type as self.

NOTE

Unlike copy_(), this function is recorded in the computation graph. Gradients propagating to the cloned tensor will propagate to the original tensor.

 

返回一個張量的副本，其與原張量的尺寸和數據類型相同。

與copy_()不同，這個函數記錄在計算圖中。傳遞到克隆張量的梯度將傳播到原始張量。

 

copy_(src, non_blocking=False) → Tensor

Copies the elements from src into self tensor and returns self.

The src tensor must be broadcastable with the self tensor. It may be of a different datatype or reside on a different device.

Parameters

• src (Tensor) – the source tensor to copy from

• non_blocking (bool) – if True and this copy is between CPU and GPU, the copy may occur asynchronously with respect to the host. For other cases, this argument has no effect.