def __init__(self, dim, variance=1.0, lengthscale=None):
super(Linear, self).__init__()
self.dim = torch.tensor([dim], requires_grad=False)
if lengthscale is None:
self.lengthscale = torch.nn.Parameter(
transform_backward(torch.ones(1, dim)))
else:
self.lengthscale = torch.nn.Parameter(
transform_backward(torch.tensor(lengthscale)))
self.variance = torch.nn.Parameter(
transform_backward(torch.tensor([variance])))
def __init__(self, dim, variance=1.0):
super(White, self).__init__()
self.dim = torch.tensor([dim], requires_grad=False)
self.variance = torch.nn.Parameter(
transform_backward(torch.tensor([variance])))
def __init__(self, variance=1.0):
super(Constant, self).__init__()
self.variance = torch.nn.Parameter(
transform_backward(torch.tensor([variance])))
def variance(self):
return transform_backward(
transform_forward(self.k1.variance) +
transform_forward(self.k2.variance))