def test_affine_shear(self, device, dtype):
torch.manual_seed(0)
shear = torch.rand(1, 2, device=device, dtype=dtype)
input = torch.rand(1, 2, 3, 4, device=device, dtype=dtype)
transform = kornia.Affine(shear=shear, device=device, dtype=dtype)
actual = transform(input)
expected = kornia.shear(input, shear)
assert_close(actual, expected, atol=1e-4, rtol=1e-4)
def test_affine_shear(self, device):
torch.manual_seed(0)
shear = torch.rand(1, 2, device=device)
input = torch.rand(1, 2, 3, 4, device=device)
transform = kornia.Affine(shear=shear).to(device)
actual = transform(input)
expected = kornia.shear(input, shear)
assert_allclose(actual, expected)
def ShearY(x, v):
batch_size = v.size(0)
shear = torch.zeros((batch_size, 2), device=x.device)
shear[:, 1] = v
return kornia.shear(x, shear)
def _transform(self, img: torch.Tensor,
param: torch.Tensor) -> torch.Tensor:
B, _, _, _ = img.shape
translation = param
return kornia.shear(img, translation.expand(B, 2))
def shear_y(img: torch.Tensor, mag: torch.Tensor) -> torch.Tensor:
mag = torch.stack([torch.zeros_like(mag), mag], dim=1)
return kornia.shear(img, mag)