def test_jit(self, device):
translation = torch.tensor([[1., 0.]]).to(device)
batch_size, channels, height, width = 2, 3, 64, 64
img = torch.ones(batch_size, channels, height, width).to(device)
trans = kornia.Translate(translation)
trans_traced = torch.jit.trace(kornia.Translate(translation), img)
assert_allclose(trans(img), trans_traced(img))
def test_jit(self, device, dtype):
translation = torch.tensor([[1.0, 0.0]], device=device, dtype=dtype)
batch_size, channels, height, width = 2, 3, 64, 64
img = torch.ones(batch_size, channels, height, width, device=device, dtype=dtype)
trans = kornia.Translate(translation)
trans_traced = torch.jit.trace(kornia.Translate(translation), img)
assert_close(trans(img), trans_traced(img), atol=1e-4, rtol=1e-4)
def test_dxdy_batch_broadcast(self, device, dtype):
# prepare input data
inp = torch.tensor([[
[1., 2.],
[3., 4.],
[5., 6.],
[7., 8.],
]],
device=device,
dtype=dtype).repeat(2, 1, 1, 1)
expected = torch.tensor([[[
[0., 1.],
[0., 3.],
[0., 5.],
[0., 7.],
]], [[
[0., 1.],
[0., 3.],
[0., 5.],
[0., 7.],
]]],
device=device,
dtype=dtype)
# prepare transformation
translation = torch.tensor([[1., 0.]], device=device, dtype=dtype)
transform = kornia.Translate(translation, align_corners=True)
assert_allclose(transform(inp), expected, atol=1e-4, rtol=1e-4)
def test_dxdy(self, device, dtype):
# prepare input data
inp = torch.tensor([[[1.0, 2.0], [3.0, 4.0], [5.0, 6.0], [7.0, 8.0]]], device=device, dtype=dtype)
expected = torch.tensor([[[0.0, 1.0], [0.0, 3.0], [0.0, 5.0], [0.0, 7.0]]], device=device, dtype=dtype)
# prepare transformation
translation = torch.tensor([[1.0, 0.0]], device=device, dtype=dtype)
transform = kornia.Translate(translation, align_corners=True)
assert_close(transform(inp), expected, atol=1e-4, rtol=1e-4)
def test_dxdy(self, device):
# prepare input data
inp = torch.tensor([[
[1., 2.],
[3., 4.],
[5., 6.],
[7., 8.],
]]).to(device)
expected = torch.tensor([[
[0., 1.],
[0., 3.],
[0., 5.],
[0., 7.],
]]).to(device)
# prepare transformation
translation = torch.tensor([[1., 0.]]).to(device)
transform = kornia.Translate(translation, align_corners=True)
assert_allclose(transform(inp), expected)
def test_dxdy(self):
# prepare input data
inp = torch.tensor([[
[1., 2.],
[3., 4.],
[5., 6.],
[7., 8.],
]])
expected = torch.tensor([[
[0., 1.],
[0., 3.],
[0., 5.],
[0., 7.],
]])
# prepare transformation
translation = torch.tensor([[1., 0.]])
transform = kornia.Translate(translation)
assert_allclose(transform(inp), expected)
def test_dxdy_batch_broadcast(self, device):
# prepare input data
inp = torch.tensor([[
[1., 2.],
[3., 4.],
[5., 6.],
[7., 8.],
]]).repeat(2, 1, 1, 1).to(device)
expected = torch.tensor([[[
[0., 1.],
[0., 3.],
[0., 5.],
[0., 7.],
]], [[
[0., 1.],
[0., 3.],
[0., 5.],
[0., 7.],
]]]).to(device)
# prepare transformation
translation = torch.tensor([[1., 0.]]).to(device)
transform = kornia.Translate(translation)
assert_allclose(transform(inp), expected)
