Hi,

The output for torch.Tensor([0])*torch.Tensor([-1]) is -0[torch.FloatTensor of size 1]

Why is it not positive 0?

Thanks.

Hi,

The output for torch.Tensor([0])*torch.Tensor([-1]) is -0[torch.FloatTensor of size 1]

Why is it not positive 0?

Thanks.

1 Like

I would guess floating point standard definition?

In any case, +0 and -0 behave exactly the same, is that a problem for you?