torch.
stack
( seq , dim=0 , out=None ) β Tensor
here what 's the use cases of out arguments ?
Itβs just an optional argument to store the output to.
E.g. if you already have the tensor initialized, you could pass it to the function:
out = torch.empty(2, 10)
# ...
torch.stack((torch.randn(10), torch.randn(10)), dim=0, out=out)
2 Likes