The difference between torch.tensor and torch.tensor

Official explanation

torch.Tensor is an alias for the default tensor type (torch.FloatTensor).

namely torch.Tensor The default is float32 type , And cannot be changed during initialization .

and torch.tensor The data type will be automatically assigned according to the initial value , And you can specify the data type during initialization .

x = torch.Tensor([1, 2, 3])
y = torch.tensor([1, 2, 3])
z = torch.tensor([1, 2, 3], dtype=torch.float16)
print(x.dtype)  # torch.float32
print(y.dtype)  # torch.int64
print(z.dtype)  # torch.float16

Reference resources  

Difference between torch.tensor and torch.Tensor - PyTorch Forums