Difference between Tensor and tensor

Question

I am working with Pytorch. But when I use tensor I got an error while I don't get any error when I use Tensor .

What is the difference between Tensor and tensor?

This is my example:

tns = torch.tensor([91,21,34,56])

tns.mean()

I got this error:

RuntimeError: Can only calculate the mean of floating types. Got Long instead.

Thanks in advance.

Answer 1

If you don't specify a dtype in torch.tensor() it infers this from the data. Since your data is all integers, it used Long .

torch.mean() can only be calculated for floats.

By default, torch.Tensor is an alias for torch.FloatTensor hence it automatically has dtype float (unless you change the default behaviour).