Will a tensor still be in pinned memory which is stacked by two pinned memory

1

Here is the example, I want to ask will the tensor_res still be in pinned memory?

tensor_a = torch.zeros([1, 3]).pin_memory()
tensor_b = torch.zeros([1, 3]).pin_memory()
tensor_res = torch.cat([tensor_a, tensor_b], dim=0)
2

No, it won’t:

>>> tensor_a.is_pinned()
True
>>> tensor_b.is_pinned()
True
>>> tensor_res.is_pinned()
False