What is the relationship between torch.Tensor and torch.FloatTensor? In the Pytorch document, its says that torch.Tensor is an alias for torch.FloatTensor. I don’t quiet understand what this means.
For example, I created a tensor x using torch.tensor function which should return an instance of torch.Tensor class.
>>> x = torch.tensor([1., 2.])
>>> isinstance(x, torch.Tensor)
True
>>> isinstance(x, torch.FloatTensor)
True
>>> issubclass(torch.FloatTensor, torch.Tensor)
False
It turns out x is an instance of torch.FloatTensor as well as torch.Tensor. So, I initially thought that torch.FloatTensor was a subclass of torch.Tensor, which it turned out that it wasn’t!!
How is this possible that x is an instance of both torch.FloatTensor and torch.Tensor when these two classes are not in inheritence relationship?
Also, what is the relationship between torch.Tensor and torch.FloatTensor and how are they different?