def test_inverse_wishart_prior(self):
prior = InverseWishartPrior(self.nu, self.K)
self.assertFalse(prior.log_transform)
self.assertEqual(prior.shape, self.K.shape)
self.assertTrue(torch.equal(prior.K, self.K))
self.assertEqual(prior.nu, self.nu)
self.assertEqual(prior.C, -2 * math.log(2) - log_mv_gamma(2, 1))
self.assertTrue(prior.is_in_support(self.K))
self.assertFalse(prior.is_in_support(self.K_bad))
self.assertAlmostEqual(prior.log_prob(self.K).item(),
-3.531024,
places=4)
def __init__(self, train_x, train_y, likelihood):
super(MultitaskGPModel, self).__init__(train_x, train_y, likelihood)
self.mean_module = ConstantMean(prior=SmoothedBoxPrior(-1, 1))
self.covar_module = RBFKernel(log_lengthscale_prior=SmoothedBoxPrior(
exp(-6), exp(6), sigma=0.1, log_transform=True))
self.task_covar_module = IndexKernel(n_tasks=2,
rank=1,
prior=InverseWishartPrior(
nu=2, K=torch.eye(2)))
def __init__(self, train_x, train_y, likelihood):
super(HadamardMultitaskGPModel, self).__init__(train_x, train_y, likelihood)
# Default bounds on mean are (-1e10, 1e10)
self.mean_module = ConstantMean()
# We use the very common RBF kernel
self.covar_module = RBFKernel()
# We learn an IndexKernel for 2 tasks
# (so we'll actually learn 2x2=4 tasks with correlations)
self.task_covar_module = IndexKernel(n_tasks=2, rank=1, prior=InverseWishartPrior(nu=2, K=torch.eye(2)))
def test_inverse_wishart_prior_invalid_params(self):
with self.assertRaises(ValueError):
InverseWishartPrior(0, self.K)
with self.assertRaises(ValueError):
InverseWishartPrior(1, self.K_bad)