def test_sin_(self): device = self.get_device() cpu_sin_pi_ = torch.Tensor([np.pi]) torch.sin_(cpu_sin_pi_) ort_sin_pi_ = torch.Tensor([np.pi]).to(device) torch.sin_(ort_sin_pi_) cpu_sin_pi = torch.sin(torch.Tensor([np.pi])) ort_sin_pi = torch.sin(torch.Tensor([np.pi]).to(device)) assert torch.allclose(cpu_sin_pi, ort_sin_pi.cpu()) assert torch.allclose(cpu_sin_pi_, ort_sin_pi_.cpu()) assert torch.allclose(ort_sin_pi.cpu(), ort_sin_pi_.cpu())
# version with an appended underscore (``_``) that will alter a tensor in
# place.
#
# For example:
#
a = torch.tensor([0, math.pi / 4, math.pi / 2, 3 * math.pi / 4])
print('a:')
print(a)
print(torch.sin(a)) # this operation creates a new tensor in memory
print(a) # a has not changed
b = torch.tensor([0, math.pi / 4, math.pi / 2, 3 * math.pi / 4])
print('\nb:')
print(b)
print(torch.sin_(b)) # note the underscore
print(b) # b has changed
#######################################################################
# For arithmetic operations, there are functions that behave similarly:
#
a = torch.ones(2, 2)
b = torch.rand(2, 2)
print('Before:')
print(a)
print(b)
print('\nAfter adding:')
print(a.add_(b))
print(a)
[[-1, -1, -1],
[-1, 10, -1],
[-1, -1, -1]],
device = device, dtype=torch.float)
print(fenced_ten.numel())
print(fenced_ten.size())
print(fenced_ten.cpu())
print(fenced_ten.relu().cpu())
a = torch.ones(3, 3).to(device)
b = torch.ones(3, 3)
c = a + b
d = torch.sin (c)
e = torch.tan (c)
torch.sin_(c)
print ("sin-in-place:")
print(c.cpu())
print ("sin explicit:")
print (d.cpu ())
a = torch.tensor([[10, 10]], dtype=torch.float).to(device)
b = torch.tensor([[3.3, 3.3]]).to(device)
c = torch.fmod(a, b)
print(c.cpu())
a = torch.tensor([[5, 3, -5]], dtype=torch.float).to(device)
b = torch.hardshrink(a, 3) #should be [5, 0, -5]
c = torch.nn.functional.softshrink(a, 3) #should be [2, 0, -2]
print(b.cpu())