sparse_GP now has a separate predict function

GP and sparse_GP used t share a predict fumction. Since we'd like to
propagate uncertainty in predictions, sparse_GP.predict needs to accept
X_new_variance.
This commit is contained in:
James Hensman 2013-05-29 13:22:25 +01:00
parent 95466d2260
commit d461f5f9bf
2 changed files with 38 additions and 1 deletions

View file

@ -447,7 +447,7 @@ class model(parameterised):
assert isinstance(self.likelihood, likelihoods.EP), "EPEM is only available for EP likelihoods"
ll_change = epsilon + 1.
iteration = 0
last_ll = -np.exp(1000)
last_ll = -np.inf
convergence = False
alpha = 0