def test_norm(self):
# Test calculating notm of a quaternion
q = [0.1, 0.2, 0.3, 0.4]
x = QuaternionTensor(q)
y = pyquaternion.Quaternion(*q)
self.assertTrue(torch.allclose(x.norm(), torch.Tensor([y.norm])))
def test_exp(self):
# Test quaternion exponential
q = [0.1, 0.2, 0.3, 0.4]
x = QuaternionTensor(q)
y = pyquaternion.Quaternion(*q)
self.assertTrue(
torch.allclose(x.exp(),
torch.Tensor((pyquaternion.Quaternion.exp(y).q))))
def test_div(self):
# Test dividing two quaternions
q1 = [0.1, 0.2, 0.3, 0.4]
q2 = [1, 2, 3, 4]
x1 = QuaternionTensor(q1)
x2 = QuaternionTensor(q2)
y1 = pyquaternion.Quaternion(*q1)
y2 = pyquaternion.Quaternion(*q2)
self.assertTrue(torch.allclose(x1 / x2, torch.Tensor((y1 / y2).q)))
def test_mul(self):
# Test multiplying two quaternions
q1 = [0.1, 0.2, 0.3, 0.4]
q2 = [1, 2, 3, 4]
x1 = QuaternionTensor(q1)
x2 = QuaternionTensor(q2)
y1 = pyquaternion.Quaternion(*q1)
y2 = pyquaternion.Quaternion(*q2)
self.assertTrue(torch.allclose(x1 * x2, torch.Tensor((y1 * y2).q)))
def test_pow(self):
# Test raising quaternion to power
q = [0.1, 0.2, 0.3, 0.4]
x = QuaternionTensor(q)
y = pyquaternion.Quaternion(*q)
self.assertTrue(torch.allclose(x**2, torch.Tensor((y**2).q)))
def test_torch_casting(self):
# Test casting to standard PyTorch tensor
x = QuaternionTensor([0.1, 0.2, 0.3, 0.4])
print(x.torch())
self.assertTrue(
torch.equal(x.torch(), torch.tensor([0.1, 0.2, 0.3, 0.4])))
def test_scalar_add(self):
# Test adding a quaternion and a scalar
q = [0.1, 0.2, 0.3, 0.4]
x = QuaternionTensor(q)
y = pyquaternion.Quaternion(*q)
self.assertTrue(torch.allclose(x + 1, torch.Tensor((y + 1).q)))
def test_add(self):
# Test adding two quaternions
q = [0.1, 0.2, 0.3, 0.4]
x = QuaternionTensor(q)
y = pyquaternion.Quaternion(*q)
self.assertTrue(torch.allclose(x + x, torch.Tensor((y + y).q)))