sum() and prod() in PyTorch

sum() can get one or more sum values as shown below:

*Memos:

• 
  sum() can be called both from torch and a tensor.


• The 2nd argument is one or more dimensions with torch.


• The 1st argument is one or more dimensions with a tensor.
  

import torch

my_tensor = torch.tensor([[5, 4, 7, 7],
                          [6, 5, 3, 5],
                          [3, 8, 9, 3]])
torch.sum(my_tensor)
my_tensor.sum()
# tensor(65)

torch.sum(my_tensor, 0)
my_tensor.sum(0)
torch.sum(my_tensor, (0,))
my_tensor.sum((0,))
torch.sum(my_tensor, -2)
my_tensor.sum(-2)
torch.sum(my_tensor, (-2,))
my_tensor.sum((-2,))
# tensor([14, 17, 19, 15]) …

dimensions function import memos prod pytorch sum tensor torch values