def obs_gp(self) -> GP:
x = torch.tensor([[0.0, 0.0], [1.0, 1.0], [2.0, 2.0]])
y = torch.tensor([[0.0], [1.0], [2.0]])
gp = GP(SquaredExponentialKernel())
gp.observe(x, y)
return gp
def test_init_ve(self) -> None:
mo_kernel = MOKernel([
SquaredExponentialKernel(),
SquaredExponentialKernel(),
SquaredExponentialKernel()
])
with pytest.raises(ValueError):
gp = GP(kernel=mo_kernel)
def test_call(self) -> None:
x = torch.tensor([[0.0], [1.0], [2.0]])
y = torch.tensor([[0.0], [1.0], [2.0]])
gp = GP(SquaredExponentialKernel()) | (x, y)
x_call = 3.0
x_post = torch.tensor([[3.0]])
p_call = gp(x_call)
p_post = gp(x_post)
assert torch.allclose(p_call.mean, p_post.mean)
assert p_call.mean.shape == p_post.mean.shape
def init_gp(self) -> GP:
return GP(SquaredExponentialKernel())
def model_nd() -> GP:
x = torch.tensor([[-3.0, -3.0], [-2.0, -2.0], [1.0, 1.0], [2.0, 2.0]])
y = -torch.sum(x**2, dim=1).view(-1, 1)
return GP(SquaredExponentialKernel()) | (x, y)
def model() -> GP:
x = torch.tensor([-5.0, -4.0, -3.0, -2.0, -1.0, 1.0])
y = torch.sin(x)
return GP(SquaredExponentialKernel()) | (x, y)