Instead of raising an error or behaving like their equivalent gt/eq/etc functions, it simply returns a boolean.
>>> import torch
>>> a = torch.rand(5)
>>> a
0.8091
0.1895
0.0382
0.9516
0.6266
[torch.FloatTensor of size 5]
>>> a > 0.2
True
>>> a == 1
False
Instead of raising an error or behaving like their equivalent
gt/eq/etc functions, it simply returns a boolean.